ADVANCED DESIGN OF

STEEL STRUCTURES
Prof. Srinivasan Chandrasekaran
Ocean Engineering
IIT Madras
 INDEX

[Link] TOPICS [Link]

Week 1
1 Form-dominant design - I 4
2 Form-dominant design - II 20
3 Failure theories -1 38
4 Failure theories - 2 58
5 Failure theories - 3 71
Week 2
6 Material properties -1 89
7 Material properties – 2 109
8 Material properties – 3 130
9 FGM 150
10 FGM for marine application -1 169

11 FGM for marine application -2 189

Week 3
12 Design methods -1 211
13 Design methods- 2 232
14 Load combinations 244
15 Dynamic material strength 260
16 Material properties variations 273
Week 4
17 Plastic design-1 294
18 Plastic design -2 316
19 Plastic design -3 332
20 Shape factor examples 348

1
 Week 5
21 Plastic analysis-1 362
22 Plastic analysis -2 380
23 Plastic design -1 396
24 Plastic design -2 414
25 Plastic design -2 426
Week 6
26 Structural Stability 445
27 Euler’s load 465
28 Rotation coefficients for stability functions 480
29 Stability functions -1 495
30 Stability functions -2 510
31 Stability function -3 522
Week 7
32 Buckling and stability 542
33 Critical buckling load- Numerical examples 557
34 Stability problems- numerical examples 583
35 Stability of shells 605
Week 8
36 Unsymmetric bending -1 622
37 Unsymmetric bending- example problems 638
38 Shear center-1 655
39 Shear center- 2 672
40 Shear center -3 688
Week 9
41 Curved section 700
42 Shear center for curved sections 711

2
43 Shear center for unsymmetrical section 725
44 Curved beam -1 736
45 Curved beam with large curvature -1 750
Week 10
46 Curved beam with large curvature -2 766
47 Modified area factor for curved section 787
48 M factor for curved beams 804
49 Crane hook 823
Week 11
50 Thin-walled section 843
51 Open thin-walled section 861
52 Lateral torsional buckling 877
53 Design for LTB-1 897
54 Design check for LTB-2 911
55 LTB example problem 928
56 Ice loads 948
57 Ice spectrum 960
Week 12
58 Blast resistant design -1 983
59 Blast resistant design -2 996
60 Blast-resistant design -3 1008
61 Blast-resistant design -4 1023
62 Fire-resistant design -1 1040
63 Fire-resistant design -2 1060
64 Analysis under impact loads 1084

3
 Advanced Design of Steel Structures
Dr. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 01
Form - dominant design - I

(Refer Slide Time: 00:29)

Friends, welcome to this lecture on Advanced Design of Steel Structures course. In this
lecture, we are going to learn two things. We are going to understand what do we mean by
form dominant design, what is specialty about this form dominant design, where are they
applicable. And when we do a form dominant design approach what will be the basic
essential challenges which come in the design principles we will learn about that in this
lecture.

We will also try to learn more about steel as an engineering material which you must have
studied, but we will see some classical advantages in connection to the form dominant design.
So, friends this course is essentially open to multidisciplinary fraternity of civil structural
engineering, mechanical engineering, offshore engineering, naval architecture, applied
mechanics and aerospace engineering group both at the student and faculty level as well as
for the practicing engineers.

4
We will do lot of solved examples, numerical examples in this course. We will also enable
you to use matlab programs for solving the numerical examples. We will help you how to
write these matlab codes very quickly for specific application of numerical examples we will
also touch upon a variety of international codal provisions applicable to design of steel
structures. So, friends, when we talk about advanced design where does the advancement
comes from – that is a first question which one we like to know.

(Refer Slide Time: 03:30)

So, advancement essentially comes from the domains of analysis and design which are useful
for both research and industrial practice. Let us quickly look at the glimpse of the course
contents what we will be discussing under the title of this course. In unit 1, we will briefly
discuss about the form dominance and how it is influencing the design practices. We will also
talk about loads on engineering structures though we all know about the conventional loads
which are acting on the engineering structures namely dead load, live load, wind load,
earthquake load etcetera.

We will pay attention to some special loads which are unconventional and not discussed in
many of the course contents. We will talk about fire load we will explain how to compute the
blast and impact load. We will also talk about ice loads these are special loads which will be
of quite interesting to many of you will be attending these lectures because they are highly
unconventional.

5
But, they act on heavy industrial structures and very less material is available in the literature
which explains this numerical modeling of these loads and how to handle the structural
design under these loads. We will also talk about the failure analysis both in 2-dimensional
and 3-dimensional stress states. What are the different failure theories and how these theories
help to compute the failure load or the failure stress acting on a cross section under
2-dimensional and 3-dimensional stress states.

(Refer Slide Time: 06:50)

Then we will also talk about material properties of steel both under normal temperature and
high temperature because we are going to deal with the fire resistant design. Therefore, we
must know the material properties of steel at higher temperature. We will also shortly discuss
about the design codal provisions related to these codes and design methods. So, this will
comprise unit 1 friends which will have a tutorials; set of tutorials and few important
takeaway questions in the examination perspective as well as in learning and understanding
perspective.

(Refer Slide Time: 08:05)

6
The next unit will deal with the plastic behavior of structures; we will deal with plastic
analysis and design. We will do lot of numerical examples to compute the plastic load
carrying capacity of sections. Then we will also talk about the importance of shape factor of
different cross sections; then moment curvature relationships; then of course, theorems to
estimate the collapse load in structures.

In unit 3 we are going to focus on stability analysis; we will start with beams, column
connections under axial tension and axial compression. We will also discuss about beam
column design under elastic support.

(Refer Slide Time: 10:15)

7
Then we will extend this to do stability analysis of frames. We will use a classical stiffness
approach. We will derive the stability functions from the first principles. We will also use
matlab codes to estimate the buckling load or what we call otherwise as critical load. In the
next unit we are going to talk about unsymmetric bending. We will also deal with curved
beams with small and large initial curvatures.

In industrial structures design of crane hooks in fact, choice of crane hook for a specific load
carrying capacity plays a very important role. So, we will do analysis of crane hooks. We will
also use matlab codes to solve the problems of estimating load capacity of crane hooks or
estimating critical stresses in the cross-section of the crane hooks.

(Refer Slide Time: 12:28)

In the next unit, we will talk about column design phenomena where we will place emphasis
on lateral buckling, torsional buckling. Then we will also talk about the design of stiffeners to
take care of lateral torsional buckling. We will discuss about lateral torsional buckling in open
sections. We will also solve examples of LTB from different codes we will take Euro code
example we will also take Australian code example. Then we will discuss using Indian code.

8
(Refer Slide Time: 13:49)

In the next unit we are going to talk about estimate of blast loads, impact loads, ice infested
loads on arctic structures. We will also talk about fire loads and we will touch upon the
concept of fire resistant design. We will also deal with design examples of large industrial
structures, offshore platforms and other special elements. When we talk about material other
than steel we are also going to introduce the advantage of functionally graded materials and
their recent applications in offshore structures and other structural systems.

So, what is an FGM, how FGM is manufactured and what are the essential advantages of
using FGM and their applications and limitations of using FGM, we will talk about this. So,
friends, this course content has got many novel and advanced topics which will be very
helpful for research as well as industrial practice purposes. So, let us join together to learn
this course, understand the contents and pay attention to also get ourself, self-examined at the
end of the course to earn a decent credit and a certificate from NPTEL.

So, having said this let us ask a question why steel is considered as a very important
construction material you are coming to form dominance now .

9
(Refer Slide Time: 16:34)

We know framed structures are very common for many industrial applications. The moment I
say framed structures they are usually statically indeterminate. So, the degree of
indeterminacy is imposed on the structural system by two ways: 1 – from the supports or
support conditions; 2 – arrangement of members. Friends, the moment I say arrangement of
members there is always a concept of geometric form playing a role here.

I am sure all of you will know what is called as a statically indeterminate structure. The
degree of indeterminacy otherwise referred as redundancy enables higher load carrying
capacity of the system which we have studied in conventional steel design of structures. So,
the arrangement of members the geometric layout of members what we call as a geometric
form helps us to get this redundancy.

So, the form dominant design is not a new concept. It is embedded in the design philosophy
for the past many years which we have been learning in the classrooms and practicing in the
industrial applications. We will talk about couple of simple examples to realize what is form
dominance and where does it come from.

10
(Refer Slide Time: 18:53)

The moment I say form I am referring to geometric form . If I take a simply supported beam
subjected uniform distributed load and try to draw the bending moment diagram and we all
know the maximum bending moment will occur at the mid span which will be wl2/8 where
this is the load acting on the system and this is the span of the beam.

So, to make it form dominant let us replace this support system by a fixed support and apply
the same intensity of load w over a span of l. If I now draw the bending moment diagram this
will have two components I am super imposing both of them. One is of course, to simply
support a bending momentum diagram whose ordinate is at the center will be wl2/8, but this
moment will be now compromised by a fixed moment at both the supports which is given by
wl2/12.

So, the net effective bending moment at the mid span is now reduced because you know this
is negative and this is positive. So, just by converting or changing the support from a simply
supported system to a fixed support system the net effective bending moment at the critical
section is reduced. One can easily find out this reduction which is nothing but wl2/8 - wl2/12
which will be wl2/24 of (3 – 2).

So, I get a great reduction in the net bending moment at the critical section just by changing
the support system. So, this is a very simple illustrative example to understand how a form
dominance plays a role. Please note a great understanding in this example. 1 – In both the

11
beams if the cross-section is same, then the moment capacity of the cross-section is higher in
the fixed case.

On the other hand, the maximum bending moment at the critical section is lesser in the fixed
case. Material is same; it can be concrete, it can be steel, cross-section is same. It can be
rectangular, it can be I – section, it can be a channel section. By keeping the cross-section
same only we change the support condition which is one of the characteristic of form
dominance help us to decrease the effective bending stress at the cross-section at the critical
section or there is a reduction in bending moment, this concept is again not new.

(Refer Slide Time: 23:19)

If I have a beam which is placed to carry a load which is uniform distributed, in olden days
people replace this beam by an arch. And people understood it was self explained in the
literature that the configuration (b) carries more load or capable of carrying more load
compared to the configuration (a). When we talk about large span trusses the bottom chord
members are not horizontal, but they are given a slight camber.

This camber is given to counteract the deflection of the tie member under self weight. So, the
bottom members are lifted by a specific value usually the camber is suggested as per the
design provisions of different types of trusses which are very common practice in industrial
structures. So, form dominance which is used either to increase the load capacity or to control
the deflection has been a practice in structural design which is not new.

12
We are not going to talk about this form dominance. we are going to talk about form
dominance on industrial structures which are applicable for oil and gas exploration. I will
give a very simple example again.

(Refer Slide Time: 25:56)

Let us say I have an oil platform which is essentially like a cube box. The box is not solid, it
is hollow inside, but it is having a large size. When I place this box in water body, the
immersed volume of this object will improvise buoyancy force. So, I have a design concept
which is a form dominant design where the weight of the structural system is much lesser
than the buoyancy.

So, the buoyancy exceeds the weight. So, what will happen when you have a system where
the buoyancy exceeds the weight? The system will have a flotation capacity, the system will
have a free flotation capacity. Now, the difference between this buoyancy force and the
weight has got to be nullified. This comes from an axial tensile cable which is connected to
the system which is used to pull this body down.

And, these are axially loaded cables which will be in pre-tension that is why I have marked
tensile here. If I consider this as To which is acting downward so, w acting downward plus To
acting downward should be now made equal to the buoyancy. Now, friends please note, the
load acting on the structure which will be a wave load which will again be a wind load
etcetera are not counteracted by the strength of the member.

13
So, the load is not counteracted by the strength of the member, it is not counteracted with the
material strength, but counteracted by the geometric form. So, the geometric form helps the
system to counteract the horizontal loads acting on the system. So, for such structural
systems, steel is the most commonly preferred material being used. In industrial structures the
example what I am citing here is referred as tension leg platform which is one of the classical
offshore platform used for oil and gas exploration.

(Refer Slide Time: 30:12)

So, friends, framed structures which have high degree of redundancy help to improve the load
carrying capacity and decrease the stress concentration on critical members. While they are in
service these structures are rarely subjected to accidents such as fire, blast, explosion etcetera.
But, if they are subjected then the consequence is very severe it will have a catastrophic effect
and hence one should know how to estimate and how to design the structural system under
these kinds of accidental loads.

So, they will cause damage to the personnel, to the asset and of course, to the environment.
So, this should be avoided. This is not a desirable feature.

(Refer Slide Time: 31:53)

14
Further, when they are subjected to accidental loads under extreme conditions we need to
understand the elastic plastic design methods which we will discuss in this course. The
analysis and design of steel structures under extreme conditions and accidents is challenging
not because they have non-linear aspects. It is because it invokes multiple physical process
and multiple design criteria, that is important. The non-linear structural response of such
systems arise from two sources, the non-linear response comes from two sources.

It can come from geometry which we call a geometric non-linearity which refers to buckling
large deformation etcetera; can also come from material non-linearity which we call as
plasticity. So, we must know how to handle the design of steel structural systems under both
non-linear response categories for buckling and plastic loads which we will discuss in this
course.

(Refer Slide Time: 34:16)

15
So, friends, having said this when we talk about strategic structures for example, nuclear
power plants offshore structures etcetera they are capital intensive industry. So, the option
given today to design a capital intensive structural systems is only steel. There is no other
altered material available and people never tried any material for strategic structures.

There are many advances in design. The moment they say advances in design it means
efficiency and effectiveness and cost saving. Furthermore, the capital intensive industries like
strategic structural systems are also vulnerable to economic slowdown because of the many
reasons. One is they have a fixed overhead cost; the cost of plants, equipment, machinery and
appurtenances are fixed and there is a significant depreciation of the assets.

(Refer Slide Time: 36:21)

16
Steel as a construction material justifies certain advantages. It has got a very high strength to
weight ratio which justifies the choice of compact sections to carry high stress concentration
loads. The second advantage could be the sturdy shape of cross-section. This justifies
effective use of the cross-section. So, I am talking about a full utilization of material.

Furthermore, it has a very low life cycle cost because steel construction is a rapid
construction process. Considering this as a rapid process it can yield return on investment
very fast. Therefore, the life cycle cost will be low. Furthermore, steel has got high toughness
and ductility. This helps steel to withstand accidental loads and cyclic loads. Steel structural
components become non replaceable for certain applications.

(Refer Slide Time: 38:39)

17
So, large span structures, offshore structures which are used for oil and gas exploration, cable
stayed bridges where you cannot think of any alternate material except steel. However, when
we talk about form dominant design, then the foremost challenge comes in the form dominant
design is stability of the structural system. It is one of the essential feature which should be
ensured in form dominant design.

Under extreme cases such as accidental loads and fire, plastic design plays a very important
role, but when you use a plastic design philosophy though you say there is an effective use of
material. But, we have to also ensure the geometric stability of the system.

(Refer Slide Time: 40:27)

18
So, friends in this lecture we learnt few things. We understood the importance of form
dominant design from simple examples to offshore structural examples. We have understood
why steel is an important engineering material. We have also understood and we are curious
to learn about the accidental loads that can occur on cross-sections such as fire blast loads
explosion loads etcetera. We understood that steel structures have many advantages the
foremost advantages return on investment is faster.

Low life cycle cost, high strength to weight ratio and quick rapid mode of construction. So,
having said and realized the importance of steel as a engineering material and realize that
form dominant designs are generally facilitated with steel as a construction material, it is very
essential for us to learn the advanced analysis design procedures that govern this kind of
structural systems in detail. So, friends, we come to this end of the lecture.

We have explained essentially the basic concepts of form dominance in this lecture. We have
given you an overview about what we will be covering in the entire course.

(Refer Slide Time: 42:47)

Look into the NPTEL website you will have references which are very useful for this course.
of course, matlab codes available to solve the numerical examples will also be given to you
and we will do lot of numerical examples in the class to explain the concepts. And, we will
touch upon application of various international codes for design of steel structures.

Thank you very much. Stay safe. Have a good day. Bye.

19
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 02
Form-dominant design - II

(Refer Slide Time: 00:22)

Friends, this is the second lecture in which we are going to discuss and learn some extended
properties or characteristics of Form Dominant Design. Then, we will also see special design
considerations for industrial structures.

20
(Refer Slide Time: 01:25)

So, in the last lecture, we said steel structural components or non-replaceable, there is no
replacement, there is no alternative on certain applications. We can give some classical
examples, where steel structures are essentially important. We can see oil and gas exploration
platforms, large-span structures, cable-stayed bridges and many more to add.

We said under these kind of special structural systems, we generally prefer to have
FORM-dominant design. As explained earlier, form dominance is load resistance not by
strength, but by the geometric form of the structural system.

(Refer Slide Time: 03:37)

21
The moment we say geometric form, it can come from modified support conditions. Modified
cross sectional properties and to some extent novelty in the material which come from recent
innovations on materials. One classical and recent advantage which is applied to structural
systems undergo large and extensive corrosion is functionally graded material.

So, in this course, we are also going to talk about what are FGMs; how are they
manufactured; what are the combinations that make a functionally graded material and if it is
manufactured on a lab scale, what are the explicit advantages this material possess in
comparison to the classical structural steel as an engineering material.

So, we will compare this with steel and see what are explicit merits that prefer engineers to
use FGM in the future in comparison to the so called non-replaceable engineering material
for large industrial structures, which is currently steel. So, friends, when we draw our
attention towards form dominant design, we are talking about the structural arrangement of
members that is a structural geometry.

(Refer Slide Time: 05:58)

So, when we talk about an innovative geometric form, under the conditions of innovative
geometric form. The foremost requirement for a structural system to function properly is its
stability. So, stability becomes the foremost requirement. This is much more important when
you talk about structural systems, where form dominant design is invoked.

22
So, friends, let us recollect form dominant design is resistance by the geometric form and not
by the strength of the material. So, when the geometric form has got to resist the applied
loads, apart from being innovative, it should satisfy the stability requirements first.

So, in this course, we have to also discuss about the stability of structural systems much in
detail. In particular, under accidental loads and fire loads, plastic design plays a very
important role. As we all know plastic design procedure effectively uses the cross sectional
property of any material, of course using plastic design for composite materials like
reinforced concrete is complicated.

Because the behavior of concrete and steel in a close agreement under the given load pattern
is difficult to establish; whereas, steel as a structural material can be comfortably handled in
plastic design which was done for the past 20-30 years. So, when steel structures are
subjected to accidental loads and fire loads, plastic design plays a very important role in the
design phenomena of these kind of structures.

Because plastic design apart from using effective cross section of the material, it also ensures
effective and safe use of the material. So, friends, please note very carefully form dominant
design is parallelly getting along with two objectives.

(Refer Slide Time: 09:44)

The first objective is innovative structural geometry and ensuring load is resisted not by
material strength, but by the geometric form that being the first objective. The second

23
objective is the chosen geometric form should be stable; it should ensure effective and safe
use of the material. So, you see in form dominant design, these two factors are circumscribed
to each other. One is the load resistance which is the requirement of any design procedure;
second is the economy which is return on investment.

We have already saw in the last lecture how because of rapid construction process steel
enables a lot of advantages over other standard construction materials. Therefore, it
empowers a quick return on investment. So, form dominant design is not only a design
procedure, it also ensures economic use and safe use of the material. So, in form dominant
design, we have to get along with the plastic design procedure which we will be discussing in
this course in detail.

(Refer Slide Time: 12:14)

Having said this, we will then review the design principles in the perspectives of plastic
design, design for fire, design under impact loads, while we ensure the geometric stability of
the structural form.

24
(Refer Slide Time: 13:14)

Friends, to make it very simple, when the material strength exceeds the yield value or the
yield limit, then it intuits plastic design. Though in plastic design, the permissible stress
cannot exceed the yield value. So, we can stress the material only up to the yield; but looking
at the modified elastoplastic curve is an idealized elastoplastic curve of steel. Beyond yield,
we allow large inelastic displacement.

So, friends, please note a very careful derived demerit of form dominant design. The derived
demerit could be as it employs plastic design, one should be prepared a large displacement
systems. I am not talking about large deformation. It is not related to any member
individually, it is related to the structural system in total. So, we are not talking about any
local deformations, we are talking about all the time global displacements. So, you would
have got a clue at this time that how a form dominant design resist the load by geometry.

25
(Refer Slide Time: 15:57)

The answer could be the form dominant design or form dominant system primarily resist the
load by undergoing large displacements. Please note these displacements are related to the
rigid body motion of the system. It is on the global perspective. It is not related to any
specific member and it is not certainly related to plastic deformation of any material.

So, the material is permitted to undergo inelastic deformation. With that permission available
on board, the structure undergoes large global displacements to resist the loads. So, if this
concept is clearly understood, then one can readily say plastic design procedure becomes an
inherent requirement of form dominant design which is one of the advanced design
procedures in steel structural systems.

26
(Refer Slide Time: 17:45)

Having said this, one should then ensure even under plastic design, even under large
displacements at the global level of the system, not at the member level ok; let us make it
very clear. Even though the system is permitted to undergo large displacements; however,
there should be a check that such displacements should be limited to permissible values. As
usual being a designer, you could have guessed by this time who would be recommending
these permissible values.

The permissible values generally come from the Codal guidelines. So, the plastic design
which is chosen as one of the design procedures for form dominant design should ensure that
the deformations should be within the permissive limits even under excessive or accidental
loads. So, by ensuring this statement, one can very happily note that this design procedure is
safe because it checks the displacement levels. Friends, when we talk about application
problems or difficulties of these kind of form dominant design procedures to large industrial
structures.

27
(Refer Slide Time: 20:33)

Let us talk about an application problem or an application issue of this design process to large
industrial structures. You know in large industrial structures like factories, paper industry, oil
and gas exploration platforms, submarines etcetera you will notice that crane hooks which are
used for on-load and off-load process of the manufacturing and fabrication in industry is a
special element is one of the special element.

Further, structural members will be subjected to special loads; as I said special loads, I am not
covering the dead load, live load, impose load etcetera, even wind load is also considered as a
special load. I talk about only ice infested loads, fire loads and impact loads as special loads.

When structural members will be encountering special loads, they induce unsymmetric
bending. Further crane hooks or special elements because they are curved beams and curved
beams are quite complex in design; but they are quite common in form dominant systems. I
have given an example in the last lecture about arches. In form dominant systems, curved
beams are quite common; but in design, they are quite complex.

So, in this course, we will try to discuss about the procedure for analyzing systems or
members under unsymmetric bending. We will also discuss about the analysis of crane hooks
or curved beams in general, its analysis procedure and the design procedures.

28
(Refer Slide Time: 23:50)

When we talk about structural systems under special loads, we need to pay attention to
special loads like blast loads, fire loads and impact loads. In addition, we will also talk about
ice infested loads because they are very useful for arctic structures and they are not
commonly available in the literature.

So, we will talk about and give a overall comprehensive understanding of all these loads in
terms of estimating them, in terms of designing a system under these loads and so and so
forth, we will also give you the MATLAB codes for doing the analysis of systems subjected
to these loads.

29
(Refer Slide Time: 25:16)

Having said this friends, when we talk about advanced design; when we talk about new or
novel geometric form, we emphasize on the form dominant design. In form dominant design,
we are repeatedly insisting that the structural system will encounter the loads and disperse
these loads not using the material strength; but by its arrangement of structural members. So,
the novelty comes from the geometric form .

So, there is a total deviation in the design philosophy. Now, materials though they possess
strength; but not used to resist loads. So, therefore, material strength does not control or
govern the design. In fact, I should say material strength alone does not govern the design.
However, when we talk about steel for such complicated or advanced application, steel has
got a very major problem of corrosion.

30
(Refer Slide Time: 27:14)

Steel has got an inherent problem of corrosion. So, corrosion as per the material science is
material loss and we know a corroded element will have strength degradation and corrosion is
a time variant process and it is severe in marine environment.

So, on one hand, we say steel has got many advantages; it is very good for form dominance,
it is very comfortable to use plastic design, it has got very good return on investment, it can
handle unsymmetric bending loads, it can also get fabricated for curved beams. But it has got
a major demerit, it undergoes corrosion.

(Refer Slide Time: 28:48)

31
Now, to address this issue, the recent innovation happened in structural engineering is
functionally graded materials, very commonly known as FGMs. So, they are very good
substitutes for steel. They possess strength similar to steel. In fact, I will show you one
application what we did in our lab to fabricate an FGM, it possess strength in fact, higher
than steel. In addition to that, it has got a very high corrosion resistance compared to steel.

So, now, we have got an innovative material which is suitable to form dominant applications
because one of the primary constituent of FGM is steel. While satisfying the form dominant
requirements of the members, it also addresses the common problems what steel faces in
special environments like marine environment, where corrosion is a major challenge.

So, we will talk about FGM in detail in this course and give you a very clear example
application of an FGM for a marine riser, where corrosion is one of the serious problem, we
still remains unaddressed in the modern engineering era.

(Refer Slide Time: 31:02)

We have already said that form dominant design is advantageous, but they should possess a
very important ability called recentering. Let us quickly see what is recentering. Before we
understand what is recentering, let us talk about what is elasticity. Elasticity is the property of
the material which enables it to regain its original shape and size upon removal of load.

A simple example could be a bar; pull the bar to some load P. If this load P as long as lesser
than P elastic limit of the material upon removal of load, the member will regain its original

32
shape and size . That is what we are studied as elasticity. Let us apply this concept for
systems undergoing large displacements which is a form dominant system.

So, let us take a body which is enabled to float. If this is my water level, the submerged
volume of the members will impose buoyancy to the system. The weight of the system will
any way act downward, the system is designed in such a manner that W and FB will be
co-linear, co-planar and act at the same point.

However, for form dominant design as I discussed in the last lecture, buoyancy will exceed
the weight. So, that if buoyancy exceeds the weight, this body which is our structural system
in terms of a rigid body model will have a free floatation characteristic. It will keep on freely
floating.

To control this free motion of floatation, this has got to be held down by some cables which
we call them as tethers which are nothing but axially stressed wires. If the tension in the cable
is To, then the difference between buoyancy and W should be taken care of by this To.

Therefore, equation of static equilibrium for the system will be now W + To = FB ok because
the left hand side acts downward, right hand side acts upward. Let us imagine a system of this
order and try to explain what is recentering.

(Refer Slide Time: 35:00)

33
So, I have a body which is anchored to the seabed using cables. Let us say this is my water
line. So, this portion is submerged; this portion is in air . The body has some weight. The
submerged portion of the body will offer an upward force which is buoyancy force.

As per design, if the system has to be positive buoyant, then FB will exceed W to compromise
this. To adjust this difference, we will have tension in the cables. So, we now said the
equation of static equilibrium for this problem could be FB is equal to W+To because this is
acting downward; this will act upward.

Now, under this condition, let us say the system is imposed by a wave load which is
horizontal. As the system is free to float, the system now assumes a new position. Under the
new position, the system has undergone large displacement ok in the presence of waves; but
please understand, the system will try to regain its original position automatically. See how?

Now, these cables will be in again tension, they will have a vertical and horizontal
component. These horizontal components will try to bring this body back to the normal
position. So, the additional components of the new axial force helps the body to regain its
original position. I am saying it help; there is no guarantee, it will come back to the original
position. But it tries. Now, in simple terms when such body, when such body tries or enables
to return to its equilibrium position, let me very clear to its static equilibrium position because
the equation of static equilibrium is this.

If the body possess a capability to return to its static equilibrium position in the presence of
load, then we call that as recentering. So, recentering is the geometric ability of a structural
system to return back to its static equilibrium position in the presence of external loads;
whereas, elasticity is a material property. So, for form dominant design systems, we must
ensure recentering ability.

34
(Refer Slide Time: 39:28)

So, the form dominant design systems should ensure a good recentering ability, then only the
system can sustain the loads acting from the environment. While doing so, the system should
remain stable. Therefore, we must understand the stability of the system under such complex
behavior. So, stability plays an important role in understanding advanced design of steel
structures under form dominant design.

So, in this course, we will also help you to derive the stability equations using matrix
approach. We will also give you the MATLAB codes to solve numerical examples of stability
problems which is quite novel and very scarce in the present literature friends. It is not
commonly available. In addition, we will also talk about the design procedures of beam
column connections under Lateral Torsional Buckling. LTB is a big menace in structural
systems of large span and cylinder systems.

35
(Refer Slide Time: 41:31)

When we talk about a large span and slender geometry, then we have to also consider lateral
torsional buckling. So, this course will help you to understand the lateral torsional buckling
analysis and design. We will also do a design example using different international codes
including Indian code for members under lateral torsional buckling in this course.

Furthermore, when we talk about structures under eccentric loads. These structural systems
will undergo unsymmetric bending and one should have an idea how to do analysis under
unsymmetric bending. So, we will talk about the derivations of design equations for
estimating stresses under unsymmetric bending and also, under curved beams.

36
(Refer Slide Time: 43:02)

So, friends, in this lecture, what we have learnt? We have learnt more insight about form
dominant design. We have understood the importance of plastic design. In this context, we
clearly understood what will be the innovation requirement of material in the modern era.
One classical application what we just mentioned is functionally graded materials and we are
also interested to learn more about unsymmetric bending, curved beams, lateral torsional
buckling; in addition to learning more on fire load, impact load, blast loads and ice loads.

So, this is what the promise what we made which will go through this course and the course
will support you through and through using MATLAB programs solving lot of numerical. On
the screen using MATLAB program and this programs will also available to you from the
NPTEL link free to download and use them and the course has got a very strong set of
reference books authored by me, on the same topics which are also published by publishers of
international repute.

So, friends, you will I believe you will join hands with me in making this learning experience
quite interesting, quite robust and of course, quite knowledgeable and learning.

Thank you. We will see in the next lecture, bye.

37
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 03
Failure theories -1

(Refer Slide Time: 00:21)

Friends, welcome to the 3rd lecture in which we are going to learn more on Failure theories.
So, I will put this as Failure theories – I, this will have about series of lectures, where we are
going to talk about the failure theories in detail. We know all engineering materials undergo
failure, this failure can be in 2 ways, it can be either by yielding or by fracture. When we talk
about failure, we are very closely associating this as a material property; it is not a member
property.

Then one may ask me a question, if I wish to associate failure as a member property, what
can I call that as in the present context of form dominant design, the answer could be if you
want to associate failure, with respect to the geometric form, but not the material then I can
say this can be related to stability or I can say it is a unstable configuration.

So, when we talk about failure of a geometric form, then stability is checked. When we talk
about material failure related to the material form, it can occur broadly in two ways, one is

38
yielding other is fracture. So, friends let us quickly see the difference between these two, they
are definitely not same and as we all know that.

(Refer Slide Time: 02:53)

Yielding is a permanent deformation, which occurs due to significant sliding on planes

through the crystalline structure of the material. I am defining this in the material perspective
friends. When does it take place? It takes place without actual rupture of the material. What
will be the consequence of yielding? One of the serious consequences of yielding is the
member will lose its functional value under excessive yield.

Instead of saying we will lose we will say the functional value will get reduced ok, let us put
it like this. Since the functional value of the member is deteriorated we can always say
yielding is a criteria of failure though yielding actually is not actual rupture of the material,
the material does not get ruptured, but still it is a failure criteria. Because it has a very severe
consequence of degrading the functional value of the member therefore, yielding is
considered as a criteria failure.

39
(Refer Slide Time: 05:50)

Alternatively what is fracture? Fracture is a failure in which separation occurs on a cross

section, this separation is perpendicular to the direction of tensile stresses. Friends please
note: fracture is also associated to the nature the direction of the stress. This is a very
common failure criteria on brittle materials.

In practice a limit of about 5 percent elongation is the partition line between the ductile and
the brittle materials. What does it mean is, if the elongation exceeds 5 percent beyond the
elastic limit elongation, then the material can be said as a ductile material. Otherwise, the
material is designated as a brittle material; this 5 percent elongation is just a thumb rule.

Friends, all material do not obey this partition line and I am not talking about only steel here,
I am talking about materials in general. So, fracture is also a failure, yielding is also a failure,
though yielding does not show any pronounced rupture of the material, but it degrades the
functional value of the member therefore, it is considered as a failure criteria, what are
theories saying about this ok.

40
(Refer Slide Time: 08:42)

Now, when we talk about estimating this failure low so, what is our objective? Under the
given load or to be very precise load combinations, one is interested to know the failure. In
fact, to be a very precise one is center to know the failure stress, but it is addressed as critical
stress.

So, we all agree that it is a material related property, generally in engineering perspective you
have learnt about estimating this failure loads in your undergraduate level itself. We conduct
something called uniaxial tensile test, is it not we are conducted this in an UTM, we used to
conduct this test. So now, let us pay attention more to this to explain the failure theory.

41
(Refer Slide Time: 10:16)

When we talk about an uniaxial stress, system the material will develop, limiting allowable
stress for design. All codes look into the stress strain curve of the material, fix certain limits
and say that is my design stress, is it not. Take for example, IS 800, IS 456 the steel and
concrete design codes, they look into the stress strain curve of the material and fix up the
value as a design stress.

And they are all based exactly on uniaxial stress system ok, which is essentially up time
please note this is essentially up time only under uniaxial tensile test ok. It is always assume
that the material obeys the same property both in tension and compression, with an ideology
many material do not obey this ideology, please understand that ok.

So, in reality, structures or members are subjected to complex stress system. So, this uniaxial
concept is actually ideal. In reality this does not work members are under complex stress
systems. Therefore, many factors will influence their failure. What are those factors? State of
stress that is it uniaxial, bi-axial or triaxial is it a 1-dimensional stress state, 2-dimensional or
3-dimensional stress state.

So, state of stress will govern the failure criteria, the second could be type and nature of load.
If the load is tensile compressive, if it is axial or if it is eccentric will it cause bending and
buckling?

42
(Refer Slide Time: 13:33)

The third could be heat treatment procedure of the material, which is used for fabricating that
structure. Therefore friends, since many factors govern the failure criteria in reality it is
important to establish the criteria for behavior of the members in general and behavior of the
materials in particular under the combined stress states; is it not?

But so far, we always estimate the permissible loads based on uniaxial tensile test, then one
may ask me a question, what are those challenges which are over sighted when we use the
test results from uniaxial tensile test?

(Refer Slide Time: 15:09)

43
So, if you use an uniaxial tensile test, to estimate the design stress value and if this design
stress value is exceeded in the actual case then we call this as a structural failure that is right.
So, all the time we try to fix up the threshold limit of the stress of the material based upon
uniaxial tensile test.

And we compare the actual stresses coming on the cross section with this value obtained
from the uniaxial tensile test and then we declare whether structure is failed or satisfactory
performed. That is the design process and procedure what we have been following for eras,
but there are some difficulties associated with uniaxial tensile test. So, what is an uniaxial
tensile test? It is a simple tensile test where the specimen is subjected to axial tension.
Usually the specimen is circular in cross section, is usual practice ok circular in cross section.

(Refer Slide Time: 17:04)

So, for example, when you take a circular bar, subject it to axial pull p the bar elongates and
the diameter gets shortened and there is a neck formation and the member fails. This is a
usual observation what we have seen and learned, in simple tension test in undergrad levels.
So, if this material happens to be steel and try to plot the stress strain curve of this a typical
stress strain curve of steel will look like this.

So, what we first observe is the material yields that is the first observation what we make,
unfortunately friends that is the only observation we make. And based on this observation we
fix everything elastic limit, proportionality, upper and lower yield point’s, ultimate strength,

44
breaking strength, everything we fix looking only at this value. But there are many things
which are happening in parallel to this, what are they?

(Refer Slide Time: 19:02)

When the material is yielded, the principal stress which is addressed as σ max reaches yield
point, stress which is σy, let us put it as σyp of the material. Secondly, the maximum shear
stress which is τ max is usually taken as σ max by 2 also reaches the yield point stress.
Mathematically, the yield point stress in shear is yield points in axial tension by 2. So, this
happens simultaneously.

(Refer Slide Time: 20:43)

45
Thirdly, the tensile strain ε reaches yield point strain which is called as εyp. 4th, the total strain
energy which is termed as u absorbed by the material per unit volume of the material also
reaches the value corresponding to yield point; mathematically, uyp is equal to σyp2 by 2E. 5th,
the strain energy distribution ud absorbed per unit volume of the material also reaches the
corresponding value to yield value; mathematically, ud at yield point reaches 1 + µ by 3E of
σyp2.

(Refer Slide Time: 22:48)

6, the octahedral shear stress reaches a value corresponding to

yield mathematically octahedral stress at yield point is equal to
√2 by 3 of σyp, which is about 0.47 σyp.

So, friends many quantities like shear stress, strain energy at distortion, tensile strain
maximum shear stress, principal shear stress also reach the corresponding values, but they are
all not considered when we try to fix up the threshold value of comparing the actual stress
with the design stress limit for any distinct procedure. One may ask me a question, how these
parallel values obtained during uniaxial tensile test govern the failure.

46
(Refer Slide Time: 24:50)

Now, let us redefine the failure again. We have agreed upon a concept that failure of a
member is related to 2, 1 is by its geometry 2 by its material. The geometric failure is related
to stability and let us say we ensure that the geometric failure does not happen. So, let us now
try to define the failure only through the material perspective and this can have two things,
one is yielding, other is fracture which we just now discussed.

To be very clear friends let us have slightly an open minded understanding, yielding is related
to ductile material and fracture can be related to brittle material. And there is a very fine mark
between the ductile and brittle metal material in terms of elongation which is about 5 percent,
though all material do not obey this fine patch. So, when the actual load or when the actual
stress exceeds the permissible stress in the material then we say it is a failure. So, all the time
failure is correlated to the permissible stress in the material and this permissible stress in
general is obtained from a uniaxial tensile test, whatever may be the material.

But in uniaxial tensile test, when the material undergoes yielding 6 more parallel engineering
values are obtained corresponding to this yielding value. Those engineering values are
becoming the governing criteria for defining the failure, friends.

47
(Refer Slide Time: 27:19)

So, the failure is not only defined based on the yield value, there are many theories which
classify the failure based on different criteria. Again, friends let me re-insist for our learning.
We are talking only about the material failure.

We are not talking about the geometric failure which can be addressed otherwise by checking
the stability conditions of the geometry. We are not discussing this now at this moment, we
are focusing completely towards a failure criteria that arise only from the material. So,
friends, let us see what all these theories addressing failure as, what are the governing factors
which these theories talk about.

48
(Refer Slide Time: 28:38)

So, therefore, in a multi axial stress state the engineering values, like maximum shear stress, a
principal stress etcetera will not occur simultaneously. So, they occur under a sequence and
that sequence governs the failure criteria. So, therefore, one need to choose the conditions for
failure based on the sequence or the governing criteria as decided by the engineer.

So, the decision cannot be purely based only on the yield stress value which is obtained from
the uniaxial tensile test. So, engineers can always have a different governing criteria for
design perspectives, based on this criteria based on the sequence of this failure an order of
this criteria various theories define failure.

49
(Refer Slide Time: 30:53)

So, let us see the first one is a maximum principal stress theory, this theory is based on
Rankine’s failure theory, it assumes that failure of a material occurs when the maximum
principal stress at any cross section attains its critical value regardless of other stresses.

Then what is the critical value? the critical value with which it is controlled is the ultimate
stress, which I refer as σultimate, this ultimate stress is determined using simple tensile test, that
is what this theory says.

(Refer Slide Time: 33:02)

50
Therefore, according to this theory failure is defined as either due to large or excessive
elongation, or it can be a fracture. Hence, for a system under complex loading, the maximum
principal stress σ1 is given by the following equation.

σ𝑥+σ𝑦
σ1 = 2
±
1
2 (σ𝑥 − σ𝑦)2 − 4τ2 = σ𝑢𝑙𝑡 --------- (1)

it is a classical equation, which we all know and we equate this to σultimate which is obtained
from the simple tensile test.

(Refer Slide Time: 35:00)

Let us plot this stress graphically, let us say this is my x and y axis. So, if this is my origin
and this indicates σ1/σultimate and this axis indicates σ2 /σultimate, I like to draw the proportion.

So, in these points will be (1,1) and this will be (-1,-1). Let us name these points as A, B, C
and D, this my origin O, this my failure envelope. So, any point lying outside the envelope is
a failure point; any point lying inside the envelope is a safe point. So, try to find out the stress
value, plot this ratio if it falls within this envelope then the member or the material is safe, it
is beyond this envelope then it is a failure.

To be very clear, mathematically the maximum principal stress is represented by a square A,

B, C, D as you see in figure 1. So, σ1/σultimate is plus or minus 1, σ2 /σultimate is plus or minus 1.
We call equation number 2 and equation number 3. So, friends based on this we can write
some observations. Failure occurs if the point falls on the periphery A, B, C, D or outside.

51
So, even if it falls on the periphery, it is a failure. Interestingly experiments are conducted and
those works show that this theory is good for brittle materials in all ranges of stresses
provided both σ1 and σ2 are tensile. So, very clearly this theory says a failure is by fracture.
So, in the 1st quadrant and in the 3rd quadrant the theory is in very good agreement for brittle
materials.

(Refer Slide Time: 39:29)

Let us talk about the next theory which is maximum shear stress theory. This theory is also
called as Tresca’s theory. This theory was suggested by Gust and Tresca based on the
experimental observations conducted on ductile materials, they observe that slipping between
the crystalline structure of the material occurs by yielding along the critically oriented planes.
This theory is based on certain following observations;

1. The maximum stress alone produces in-elastic deformation

2. If σ1 is equal to σ2 then they will have no influence on the inelastic behaviour.

52
(Refer Slide Time: 42:02)

3. Failure will occur when the maximum shear stress, see the deviation. In a complex system
reaches the maximum shear stress value, which is obtained from a simple tensile test in yield
point. Mathematically, let us assume a bi-axial system or bi-axial stress state, the bi axial
stress state is plotted. Let us hatch the surfaces, we call this stress σ1, we call this stress as σ2,
let us say figure 2. Let us name the plane A, B, C, D, E, F, G, H.

Now, let us pick up the plane separately as seen here. Let us say I pick up a plane which is A,
B, E, F, the green one let us pick up the green one. The green one says this is A, this B, this E
and this is F and this is now subjected to let us say σ1. That is then draw the next plane which
is A, E, H, D and this is now subjected to σ2.

We also assume that σ1 is greater than σ2 is greater than σ3. In such condition τmax is actually
equal to σ1 minus σ2 by 2. I equate this to the shear stress obtained from the yield point,
usually this value is equal to the yield stress by 2, in simple tension. I call this equation
number 4.

53
(Refer Slide Time: 46:49)

So, therefore,

--------- (5)

So, that gives me failure envelope based on this particular theory.

54
(Refer Slide Time: 48:23)

Interestingly friends, when σ2 is 0, σ1 will be plus or minus σyp when σ1 is 0, σ2 will be plus
or minus σyp. So, this can be expressed graphically as below, say this is my σ1 and σ2 axis.
Let us plot plus or minus 1 so, this is σyp, then plus or minus 1 in the third quadrant and the
failure envelope simply joins this way.

So, this becomes my failure envelope. So, in this figure consider plane A, E, H, D, for
yielding to occur in this plane plus or minus σ2 should be equal to σyp. similarly for the plane
A, B, F, E failure occurs by yielding when plus or minus σ1 is σyp.

So, this will represent the first and third quadrants of the failure envelope. If there is a point
which is lying in between here inside this envelope no failure, if it lies on the envelope or
outside the envelope this indicate failure.

55
(Refer Slide Time: 51:15)

So, the failure envelope in this case is a hexagon. If you consider a triaxial stress state, then in
that case the shear stress is very marginal in magnitude, in that case failure occurs by fracture
and not by yielding. So, in that case you must apply maximum principal stress theory, to
estimate the failure load. Friends, there are some observations and literature based on this
theory. It says that Tresca’s theory is good for ductile materials corresponding to state of
stress which can develop large shear stress. So, in short it is very good for 2-dimensional
stress states.

(Refer Slide Time: 53:20)

56
For a tensile test, in case of pure shear, where the maximum shear is developed the shear
elastic limit of the ductile material is taken as about 0.57 of the tensile elastic limit. This is by
experimental observations. Hence in such cases maximum shear stress theory gives
conservative results.

(Refer Slide Time: 54:34)

So, in this lecture friends, we have learnt what is the difference between yielding and
fracture. Why simple tensile test cannot be a governing criteria for design, what are the
problems associated in parallel with this simple tensile test and how the failure theories argue
based on the failure criteria.

Then we have learnt the maximum principal stress theory, then maximum shear stress theory
and we understood the failure envelope of both these theories in this lecture. I believe you
will be able to follow this and have a good reading back on this with additional support
material and any doubts accumulated you will post back to me in the discussion forum.

Thank you very much, have a good day, bye.

57
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 04
Failure theories - 2

Friends, let us continue to discuss about the Failure theories. In the last lecture, we discussed
about two failure theories. We gradually started understanding and learning the importance of
assessing the failure criteria.

(Refer Slide Time: 00:38)

So, what we are trying to look here is to fix or to assess the failure criteria. Why is it
necessary? I will explain this and interconnect this to the plastic design in the upcoming
lectures. So, when the structural member is subjected to uniaxial, biaxial or triaxial stress
states different theories predict the failure stresses at different perspectives for unlike and
alike stresses in a biaxial or a triaxial stress states.

So, in this context we discussed about two theories in the last lecture. Now, we will talk about
the third theory in this lecture which is the maxiµm strain theory. This theory is also known
as [Link]’s theory. The statement of the theory is like this, in a given complex stress state
yielding at any section, we will also say this as critical section begins in a material when the

58
maxiµm strain exceeds the strain corresponding to the yield point of the material. So, this
theory focuses essentially on the strain criteria that is why it is called maxiµm strain theory.

(Refer Slide Time: 03:44)

If σ1 and σ2 are the principal stresses under the condition that σ1 is larger than σ2 in
magnitude. Strain in the direction of σ1 is given by the following equation;

We will call this equation number 1. Or that is what the theory states. Now,
if both the stresses σ1 and σ2 are of same nature, let us say maybe tensile then substituting
this condition in equation 2, we get σ1 - µσ2 = σyp which means that σ1 will be σyp + µσ2
which exceeds even σyp , is it not. So, that is a very interesting condition; of course, this
condition does not occur when σ1 and σ2 are dissimilar in nature.

59
(Refer Slide Time: 06:47)

For σ1 tensile and σ2 compressive equation 2 will become σ1 - µ(-σ2) = σyp which says that σ1
will be σyp minus µσ2. So, in this case σ1 does not exceed σyp is it not, but whereas, in the
earlier case if both stresses are same in nature; let us say both are tensile, there is a
probability that one of the principal stresses can even exceed the stress at the yield point.

So, this is an alarming condition which says that when both the stresses are tensile, there exist
a possibility of σ1 greater than the stress at yield point that is σ1 even exceeds σyp, even
without causing yielding in the material that is very interesting. So, one can then say that the
maxiµm strain theory is an improvement upon the principal stress theory.

60
(Refer Slide Time: 09:37)

This theory, as observed in the literature, this theory does not hold good for ductile materials,
it is better for brittle materials. So, let us try to plot this theory graphically. So, you µst
understand one important statement for σ1 and σ2 both being tensile σ1 can even go higher
than σyp. Same story for σ1, σ2 both being compressive, still σ1 can exceed σyp. It means in
the first and third quadrants σ1 exceeds σyp, I should say plus or minus σyp.

(Refer Slide Time: 11:17)

61
Let us say the x axis plots σ1 and y axis plots σ2, this is positive and this is negative σ1 and
this is negative σ2. So, let us try to get a control line to plot this failure envelope which used
to do for the remaining theories as well. Let us say my σyp is here, these are my σyp.

So, ideally speaking this value should be σyp, but for σ1 and σ2 both being tensile that is the
first quadrant, I am getting σ1 more than σyp. So, let us take that point as somewhere here.
Similarly, even in the third quadrant for σ1 and σ2 both remaining compressive even this is
true it exceeds whereas, in the off-diagonal sides these values are lesser. I will come to that
point how we work it out.

So, the envelope says it should cross σyp and on the dissimilar states; so, let us try to write
down this equations for this particular line. So, in this case it will be σ2 minus µ σ1 which
will be YP. So, σ2 exceeds σyp, this particular line the equation will be σ1 minus µ σ2 will be
σyp. So, σ1 will exceed σyp.

Similarly, this particular line σ2 minus µσ1 will be minus σyp. This particular line it will be σ1
minus µ σ2 will be minus σyp. Let us talk about this particular line, let us say we will divide
this line as a b c d e f g and h. So, this equation is for a h, this equation is for a b, this
equation is for e f, this equation is for e d. Let us write down the equation for c d, you know
here it is dissimilar state, σ1 is compressive, σ2 is tensile.

So, it should be σ1 plus µ σ2 will be minus σyp, this is for c d whereas, g f will have σ2 plus µ
σ1 as minus σyp. Now, let us come to the line, this is for g f; let us come to the line equation b
c. So, this equation will be σ2 plus µ σ1 will be σyp ok. This is for line b c and this line will
have an equation σ1 plus µ σ2 will be σyp, this is for the line g h.

So, this is my failure envelope, as we understand any point lying on the periphery or inside or
outside. So, if I say A, B and C, A indicates failure, B also indicates failure, C no failure. So,
we can write a very interesting information here, in the second and fourth quadrants that is a
value at c and g are lesser than σyp; it means when the stresses are dissimilar in nature the
prediction of failure is poor.

Because, it says the system fails even before yielding is it not because, this value is not σyp,
σyp is somewhere here; its somewhere here. But, it fails at g which is much lower than σyp
whereas, in quadrants I and IV, when both stresses are either tensile or compressive that is
similar in nature, it says one of the stress even exceeds σyp, then only the failure will occur.

62
Without yielding value starts whereas, when σyp is reached yielding should initiate in the
material.

So, one can see very well here that this theory has lot of controversies in a biaxial stress state,
when the principal stresses are same in nature or different in nature that is represented in all
the four quadrants as you see in the figure on the screen. I think it is very easy and convenient
for all of you to really know how do we generate these equations. For example, take this
equation, this is same equation as equation 2 whereas, σ1 being positive σ2 will exceed σyp, is
it not.

So, this line for example, I am marking in black dotted, this line is the rate of growth of σ2,
the rate of growth of σ2 beyond σyp as an influence from σ1.

So, that is how this equation is written. Similarly, if you want to find the rate of growth of σ1
beyond σyp as an influence from σ2, this equation is written. So, by this algorithm you should
be able to realize the equations of all other segments in various coordinates and various
quadrants represented as shown on the screen. So, let us move on to the next theory which is
total strain energy theory.

(Refer Slide Time: 21:11)

So, now we are looking at the IVth theory which is total strain energy theory. This theory was
proposed by Haigh, it is also called as Haigh’s theory in the literature. The statement of the
theory is as follows: inelastic action or yielding at a section begins only when energy per unit

63
volume absorbed at a point. So, energy absorption is actually the strain energy, is equal to the
energy under uniaxial tensile test or uniaxial stress state of the material which is obtained
from the simple tension test.

This theory has got a very interesting and major contradiction. This theory says in this case,
failure does not depend on the stress state. That is a very interesting deviation this theory
gives in comparison to the remaining three theories. But, failure is governed by the energy
stored in the material per unit volume. So, the focus is on strain energy. Let us get more
details of this theory and derive the controlling equation for the failure envelope.

(Refer Slide Time: 24:54)

Let us assume a triaxial stress state, where the following condition is satisfied. The condition
is σ1 exceeds σ2 exceeds σ 3. For this stress state ε1, ε2 and ε3 are given by the following
equations, it is very easy. Please look at the screen carefully and then watch how I am writing
this equation. When you talk about strain, it is stress by Young’s modulus.

So, σ1 minus Poisson’s effect on the remaining two axis. So, when you talk about ε 2, it is σ2
Poisson’s effect on the remaining two axis. When you write please see the order, 1, 2 and 3,
2, 3 and 1. So, I think you have to follow this order. So, by this logic now you can easily
write ε 3 which is σ 3 minus µ of 1 plus 2.

64
I call this equation as 1 a. Once I know the strain values, I can always find the strain energy
per unit volume which is usually expressed as u, which is given as

. I call this equation as 2. Now, I have the equation for the

strains along three principal axis 1, 2 and 3 which is given by equation 1 a, let us substitute 1
a in equation 2 and see what happens.

(Refer Slide Time: 27:31)

So, substituting 1 a in equation 2, we get the total strain energy per unit volume u, let’s say

𝑈=
1
2𝐸 [σ1(σ1 − μ(σ2 + σ3)) + σ2(σ2 − μ(σ3 + σ1)) + σ3(σ3 − μ(σ1 + σ2))].
I think it is very easy to substitute, you can do it yourself. Let us simplify this. So, 1 by 2 E of
one can see here this product will give me squares. So, let us say σ1 square σ2 square σ 3
square, then you will also see σ1 σ2 once and σ1 σ2 twice. So, you get two products of both
repeatedly. Let us do that, minus µ times of σ1 σ2 plus σ1 σ 3 plus let us say σ2 σ 3 plus σ2
σ1 plus σ 3 σ1 plus σ 3 σ2.

65
𝑈=
1
⎡⎢σ 2 + σ 2 + σ 2 − μ σ σ + σ σ + σ σ + σ σ + σ σ + σ σ ⎤⎥.
( )
2𝐸 ⎣ 1 2 3 1 2 1 3 2 3 2 1 3 1 3 2 ⎦

Let us put it in green. Now, you see one σ1 σ2 here, one is here. Similarly, σ1 σ 3 σ1 σ 3,
similarly σ2 σ 3 and σ2 σ 3. So, there are two values here. So, therefore, let us simplify this
further and say it is going to be 1 to 1 by 2 E of σ1 square 2 square 3 square minus 2 µ of σ1
σ2 2 3 and 3 1 agreed.

𝑈=
1
⎡⎢σ 2 + σ 2 + σ 2 − 2μ σ σ + σ σ + σ σ ⎤⎥.
( ) ----------------(3)
2𝐸 ⎣ 1 2 3 1 2 2 3 1 3 ⎦

We will call this as equation number 3, I hope there is no confusion in this derivation. Please
follow is very interesting and very important and I want you to derive as you do it on the
screen. So, that you have no confusion later and you do not have to memorize anything.
Learning is a very important skill friends. We are emphasizing on learning, not on
memorizing. Once you learn, keep on doing it you will understand, once you understand
there is no question of memorizing it, there is no requirement for that.

So, let us focus on that very clearly. So, equation 3 gives me the total strain energy per unit
volume of the material.

(Refer Slide Time: 31:47)

66
Now, if the stress state is biaxial that is σ 3 is 0, then let us see what happens. Then equation
3 will become

is it not, that is very obvious. Let us call this equation number 4. And, more simple for an
uniaxial stress state put σ2 and σ 3 both as 0, then we get u is going to be σ12 /2E.

Now, the theory says this is where the yield value will be reached. Hence, I can say strain
energy at yield point will be similarly the stress at that point divided by 2 E, can I say this
equation number 5. I think we all agree to land up in equation 5 without any confusion.

See equation 3 is for triaxial stress state, equation 4 is for biaxial stress state. The theory says
equation 3 or equation 4 should be identical to equation 5. So, let us equate that.

(Refer Slide Time: 34:01)

So, for a triaxial stress state should be equal to σyp square right, for a biaxial stress state we
call this equation number 6 and this equation number 7. Take a simple case, please look at
equation 7. Equation 7 actually represents an ellipse with the condition that the major and
minor axis of the ellipse are placed at 45 degrees to each other, look at this factor.

67
(Refer Slide Time: 35:40)

So, let us plot this envelope, let us call this as σ1 and this as σ2. Let us say this is positive
axis, let us draw the major and minor axis of the ellipse at 45 degrees. Now, in this equation if
any one stress value is 0 in equation 7, you will automatically see the corresponding axis will
be representing the σyp. So, σyp should be marked and the curve should pass through σyp, let us
mark.

So, let us draw the ellipse, that it cannot cross σyp. So, when you draw this ellipse, let us be
very careful about this. So, I have drawn the ellipse. This angle is 45 degrees. Now, let us do
some calculations for σ2 equals 0, we know σ1 is plus or minus σyp which is indicated here,
for σ1 equals 0, σ2 will be plus or minus σyp is indicated here.

So, it has to pass through these four points which is passing through, the curve has to pass
through this point. Let us say we will draw the curve passing through this point, For σ1
equals σ2 equals σ that is the stresses are same magnitude, then in that case σ will become
plus or minus σyp /√2(1-µ), please do this calculation and check; same
intensity, same magnitude this equation is true.

For a typical value of µ as 1/3, you will find this stress will become plus or minus 0.87 σyp.
So, this value when the stress of the same intensity and nature may be tensile, may be
compressive. This is 0.87 σyp, it is not σyp please understand that. So, the curve should not

68
pass through this point. Further, for σ1 equals σ and σ2 equals minus σ, that is we are talking
about now, the second and the fourth quadrant.

In that case, you will see σ will be plus or minus σyp /√2(1+µ). Please see this
equation, which gives me for µ equals 1 by 3, this stress value will
become 0.613 σyp. So, these values are 0.613 σyp. So, this is my failure region, I mean
safe region and anything exceeding is a failure region. So, this theory predicts that at failure,
yield stress is not reached. It is only about 87 percent the material fails or the failure happens
in the member.

Even though the material has not touched the yield stress, when the stresses are similar in
nature and same magnitude. When they are dissimilar in nature, failure occurs even much
lower than σyp. So, that is a very critical and dangerous composition and interpretation this
theory gives me.

(Refer Slide Time: 41:36)

So, friends let us quickly see the summary what we learnt in this lecture. The summary says
we have learnt two more additional theories. We have also saw that these theories show good
disagreement when the stresses are dissimilar in nature. One of the theory says when the
stresses are tensile or compressive, I mean both of them that is in first or third quadrant, the
stress even exceeds σyp without yielding, which is not correct. Because, when the stress
exceeds σ reaches σyp yield has to start in the material.

69
So, this theory one of the theories amongst these two I leave it to you, which theory is that
which gives you this and the failure envelope is elliptical or a rhombus depending upon the
pattern what you assume. So, it is very interesting and critical to really guess what theory will
be used for assessing the failure condition. because, at failure condition the engineer desires
to know what is the stress.

It is not always that when the stress reaches σyp failure starts, you know the theory says even
at 0.87 the failure can happen, even at 0.613 the failure can happen and so on. So, there are
serious discrepancies suggested by these theories when we talk about failure.

Thank you very much, have a good day.

70
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 05
Failure theories - 3

(Refer Slide Time: 00:30)

Friends, let us continue to discuss about the Failure Theories more in detail in this lecture. So,
we have discussed about 4 failure theories so far. We will now discuss the 5th theory which is
the maximum distortion theory, this theory was proposed by Von Mises and Henckey and
famously also known as Von Mises theory.

The statement of the theory is as follows. Inelastic action or yielding at any point or cross
section in a member occurs under combination of stresses. This begins only when the strain
energy per unit volume observed absorbed at a section is equal to the strain energy per unit
volume corresponding to the yield point stress obtained from the stress strain curve of simple
tension test.

The statement looks similar to what we have in the strain energy theory, but there is a
marginal variation in this theory, which we will explain as we proceed.

71
(Refer Slide Time: 03:34)

So, for a triaxial stress state we already wrote this equation total strain energy per unit
volume which is given by U is expressed as follows, we call the equation number 1.

Neglecting the higher powers one can also write the volumetric strain as follows: the
volumetric strain is given as a sum of the strains in all the 3 axis neglecting the higher powers
where the strain in all the 3 axis are known to us; we call this equation number 3. I can
substitute 3 in 2 and let us see what do we get.

72
(Refer Slide Time: 05:33)

So, we get the volumetric strain as

which can be said as . We call this equation number 4, look at this

equation carefully this equation 4 states that volumetric strain is proportional to the sum of 3
principal stresses.

If the summation is 0, then volumetric strain vanishes. So, if the volumetric strain vanishes
what happens to the member or the body? Only the body or the member will be subjected to
distortion.

73
(Refer Slide Time: 07:51)

So, friends the total strain energy of a member or a body consists of 2 parts; one the one
which is associated with volumetric change of the body, the one which is due to change in the
shape of the body which we call as distortion.

Therefore, for the condition of zero-change in volume we get 𝛔1+ 𝛔2+ 𝛔3 as 0. Now, if these
stresses are equal to magnitude P, then there will be no distortion in the body because the
stresses are equal in magnitude and direction. So, no volumetric change and no distortion.

(Refer Slide Time: 09:53)

74
Hence, P should be 𝛔1, 2 and 3 divided by 3. Let us explain this graphically, if I have a body
subjected to triaxial stress state, let us say this figure 1, this can be expressed as plus. So, 𝛔1
is equal to, the following relationship holds good, I call this equation number 7.

(Refer Slide Time: 11:49)

Now, adding 𝛔1 plus 𝛔2 plus 𝛔3 is 3 P plus these variables. So, for no distortion case P
should be equal to this relationship, hence 𝛔1, 2 and 3 dash should be 0; For no distortion
case, hence the condition given in figure 3 that is this condition given in figure 3 causes only
distortion and no change in volume.

(Refer Slide Time: 13:17)

75
Then further let us say the stress intensities are equal. Substitute this condition in the total
strain energy equation. Which says u will be now

, which now becomes

. We will call this as equation 9.

Since, P is sum of these expressed by this equation u volumetric, will be now

(Refer Slide Time: 14:59)

Distortion energy can be obtained by subtracting equation 10 from the total strain energy
because total strain is summation of volume plus distortion; is it not, which now given the
equation 9.

So, I should say Udistortion should be total u minus u volume, let us do that. So, let us write
down that u distortion will be 1 by 2 E of 1 square 2 square 3 square minus 2 mu of the
products. Minus 1 minus 2 mu by 6 E, let see 2 mu by 6 E of 𝛔1, 2 and 3 the whole square.

So, let us do a simple mathematics here,

2 2 2 2
(𝑎 + 𝑏 + 𝑐) = 𝑎 + 𝑏 + 𝑐 + 2(𝑎𝑏 + 𝑏𝑐 + 𝑐𝑎). Hence, 𝛔1, 2 and 3 the whole square

76
will become 1 square 2 and 3 square plus twice of 1 2, 2 3 and 3 1. So, let us substitute this in
the previous expression.

(Refer Slide Time: 17:26)

So, now Udistortion will be 1 by 2 E minus 1 by 6 E. Let us simplify this which will be 1 by 2 E
of minus 1 by 6 E of plus 2 mu by 6 E of.

(Refer Slide Time: 19:24)

77
Which will become 𝛔1 square 2 square and 3 square. Let us take it common and see what
happens. If I take this terms common then I can say which will be 1 by 2 E, which is this term
minus 1 by 6 E which is this term plus 2 mu by 6 E which is this term.

Then we can also the remaining terms minus 1 by 3 E of 𝛔1 𝛔2 3 and 3 1; This term comes
from this, because twice 6 E. So, 1 by 3 E correct plus mu by 6 E of I can say here I get only
mu by 6 E, where I also have mu by 2 E here because mu by E because this 2, this 2 and this
2 goes away so mu by E. So, I can now say it is going to be 6 times of 𝛔1 𝛔2 𝛔2 𝛔3 𝛔3 𝛔1,
that is these terms all are negative ok.

Because you see they are all negative plus I have this term also. So, there is 2 and 2 = 4 here.
So, I should say plus 4, let us simplify this. So, which I do that I will get 2 plus 6 mu by 6 E
of squares minus 1 by 3 E of products minus 2 mu by 6 E of products.

(Refer Slide Time: 22:40)

Let us simplify this further which will become minus 𝛔1 𝛔2, 2 3 and 3 1 of 1 by 3 E plus 2
mu by 6 E; See here. 1 by 3 E 2 mu by 6 both are negative. So, I take a negative here I further
simplify this, I get 1 plus mu by six6 E of the squares plus 2 into 1 plus mu by 6 E of minus
products, which says 1 plus mu by 6 E of plus 1 plus mu by 3.

78
(Refer Slide Time: 24:24)

(Refer Slide Time: 24:45)

I can further simplify, this which will now become 1 plus mu by 3 E of the squares minus I
think I will get here let me make this correction I will get here 3 Es right; please make that
correction. So, I get here 3 E because 2 times of 1 plus mu by 6 E becomes 1 plus mu by 3 E
ok; so 3 E.

So, I take 1 plus mu by 3 E as common. So, 𝛔1 square minus the products, my u distortion is
actually equal to 1 plus mu by 3 E of 1 square 2 square and 3 square minus the products ok.

79
So, let us do little bit more mathematics in this, you know 𝛔1 minus 𝛔2 the whole square is
this way, the whole square is this way ok.

Let us sum them up, which will be equal to twice of 𝛔1 square 2 square and 3 square ok,
minus the products. What I can say now is, 𝛔1 minus 𝛔2 the whole square plus 𝛔2 minus 𝛔3
the whole square plus 3 minus 1 the whole square by 2 should be equal to 𝛔1 square 2 square
3 square minus 𝛔1 𝛔2 2 3 and 3 1.

So, what I have it here? So, I am going to replace this bracket term with this. So, there is 1
plus mu by 3 E, there is 2 here.

(Refer Slide Time: 27:42)

So, I can now write 1 plus mu by 3 E of, this will now become 1 plus mu by 6 E of 𝛔1 minus
𝛔2 the whole square, am I right. So, therefore friends u distortion is given by this equation
now. The equation number will be we can say 11 ok, let us say 11 will be given by 1 plus mu
by 6 E of this equation.

80
(Refer Slide Time: 28:45)

Now, having said this distortion energy for material under uniaxial stress state in a simple
tension test is given by the following expression.

So, in uniaxial stress state 𝛔2 and 𝛔3 will be said to 0 and 𝛔1 will reach 𝛔yp, Hence, u
distortion at yield point will be 1 plus mu by 6 E of 𝛔yp square, Simply substituting this
expression and make 𝛔2 and 𝛔3 is 0, you get this; We call this equation number 12.

For yielding condition according to this theory in a biaxial stress state, where 𝛔3 will be said
to 0 𝛔1 minus 𝛔2 the whole square plus 𝛔2 minus 𝛔3 the whole square plus 3 minus 𝛔1 the
whole square should be equal to 2 𝛔yp square;. How do you get this? It is very simple, this is
1 plus mu by 6 E here and this is 1 plus mu by 6 E, this is going to be 3 E.

So, I can now straight away say equating this with the previous case. So, I can write this
equation know, I can write this equation is it not; 𝛔YP2.

81
(Refer Slide Time: 31:22)

So, 𝛔1 minus 𝛔2 the whole square is the whole square 𝛔3 is the whole square is 2 𝛔yp
square, for 𝛔3 equals 0 𝛔1 minus 𝛔2 the whole square plus 𝛔2 the square plus 𝛔1 square
should be equal to 2 𝛔yp square.

So, for a plane stress problem 𝛔3 is 0. So, 𝛔1 square plus 𝛔2 square minus 𝛔1 𝛔2, should be
equal to 𝛔yp square. Again, if you look at this equation, equation 14 also represents an
ellipse.

(Refer Slide Time: 32:44)

82
Let us try to plot this envelope, let us say we get 𝛔yp in all the cases, I will mark that as an
envelope for our control line.

So, these are all 𝛔yp and minus 𝛔yp ok. So, for mu equals 1 by 3 and 𝛔1 equals sigma, 𝛔2
equals minus 𝛔you will find the 𝛔value will become 𝛔yp by root 3 when you substitute in
this equation. Now, let us draw the envelope which is going to be these are the points, it must
touch here ok, this is envelope.

So, this value 𝛔yp by root 3. Whereas, this values are 𝛔yP. So, this theory has a good
agreement with the experimental results, for ductile material. Further, if one of the principal
stresses at a point is very large compared to others, then all theories lead to same results that
is very interesting.

(Refer Slide Time: 35:46)

This is a graphical comparison of all the theories, as we just now derived and plotted in this
previous 3 lectures. So, in this plot 𝛔3 is taken as 0. So, we are plotting for a biaxial stress
state. So, all theories are the maximum principal stress theory, maximum strain theory, total
strain energy theory, Von Mises and maximum shear stress theory ok. This is the plot what
you have obtained please turn back your notes and compare them is what we are getting. We
can write down some important observations for this.

When we compare St. Venant’s theory with maximum principal stress theory, it is seen that
since tension is in one direction, reduce a strain in the other one in the perpendicular

83
direction, 2 equal magnitude tension will cause yielding at much higher values. We can
compare this point 5 and point 1, point 5 is here point 1 is here when you have got both
stresses as tensile, you will see that yielding is caused at much higher stress exceeding 𝛔yp,
is it not.

So, stress exceed 𝛔yp to cause failure, if I say failure is a yielding then this stress is
exceeding 𝛔yp to cause yielding.

(Refer Slide Time: 38:25)

In case of unlike stresses maximum strain theory indicates yielding at point 7 and 8, the
values are equal to 𝛔yp of 1 plus mu these values are much lesser than 0.2 and 0.4.

So, it means in quadrants II and IV, there is a very strong disagreement. What are these
quadrants II and IV? When the stresses are dissimilar in nature of the same magnitude the
disagreement between the theories are very large, maximum shear stress theory that is given
by Tresca, which is indicated by this curve. Indicated by an irregular hexagon coincides with
the principal stress theory in case of similar stresses.

In case of dissimilar stresses, it shows disagreement. Maximum shear stress theory has only a
very small deviation with maximum distortion theory. Maximum distortion theory is this, the
purple line. Let us look into quadrants II and IV, the maximum shear stress theory is in cyan
line, is got a very marginal disagreement.

84
So, friends, if I really wanted to find out what would be my stress at failure and if you believe
yielding is one of the failure modes and I want to estimate the stress at yield under a different
stress state conditions and different stress magnitudes different theories give me
disagreemented and non-proportional results.

So, it is very difficult for an engineer to predict what would be my deciding stress value
because certain theory say stress can be even more than 𝛔yp, to cause failure. Certain theories
and certain quadrant says stress is definitely lesser than 𝛔yp to cause failure. It means even at
lesser stresses than 𝛔yp failure can be caused.

So, for causing a failure stress need not be equal to yield value. So, that is what we have
learnt by comparing these theories and assessing failure.

(Refer Slide Time: 42:48)

Let us quickly do a problem and understand this more in detail. The major principal stress let
us say 𝛔1 of a member is 200 N/mm2 tensile.

The minor principal stress seems to be compressive. If the yield strength of the material is
300 N/mm2, find the minor principal stress using the following theories. Maximum strain
theory, which is St. Venant’s theory, maximum shear stress theory which is called a Tresca
theory, using total strain energy theory, also using maximum distortion energy theory. We can
take 𝜇 as 0.25 for the problem.

85
(Refer Slide Time: 44:21)

Let us solve this problem, for maximum strain theory, the governing equation is 𝛔1 minus mu
𝛔2 is 𝛔yp, please check the derivation what we did. In this case 𝛔2 is compressive, so we
taken as negative ok. So, 𝛔1 minus mu of minus 𝛔2 is 𝛔yp. So, 𝛔1 plus mu 𝛔2 is 𝛔yp.

So, 𝛔1 is known and mu is known, 𝛔2 is not known and 𝛔yp is known, I can solve this to get
𝛔2 as 400 compressive, when I use this first theory.

(Refer Slide Time: 45:26)

86
For maximum shear stress theory, the governing equation is 𝛔1 minus 𝛔2 is 𝛔yp. So, for 𝛔2
compressive, 𝛔1 minus of minus 𝛔2 is 𝛔yp. So, 200 plus 𝛔2 is 300. So, 𝛔2 is 100 N/mm2
compressive.

For total strain energy theory this is 2, the governing equation is 𝛔1 square 2 square minus 2
mu 𝛔1 𝛔2 is 𝛔yp square. Please check the governing equation for 𝛔2 compressive.

(Refer Slide Time: 46:57)

This equation now becomes 𝛔1 square minus 𝛔2 the whole square minus 2 mu of 𝛔1 minus
𝛔2 is 𝛔yp square, which is 𝛔1 square 𝛔2 square ok plus 2 mu 𝛔1 𝛔2 is 𝛔yp2.

So, 2002 plus 𝛔22 plus 2 of 0.25 of 200 of 𝛔2 is 3002. So, this becomes a quadratic in 𝛔2, you
can solve this quadratic and you find 𝛔2 as 179.13 N/mm2 compressive.

Now, for maximum distortion theory, the equation is given by for 𝛔2 compressive. So, let us
substitute, that again it is a quadratic in 𝛔2 solve we get 𝛔2 as 144.95 newton per mm square
compressive. So, friends you see different theories estimate different principal stresses; is it
not. So, that is the whole analogy what we learned from this theory.

87
(Refer Slide Time: 48:46)

So, in this lecture we learnt more about failure theories, we also understood a comparison
between the failure theories and noticed a major discrepancy in II and IV quadrants, correct.
We have also learnt the stress need to be more than 𝛔yp to cause failure by yielding, that is
what one of the theory says.

All the theory says stress even at less than 𝛔yp can cause failure in quadrants II and IV, is it
not, which is not acceptable because if I take yielding as a failure criteria no stress lesser than
yielding can cause me failure. So, it is very interesting that we have to fix up the stress at
which the failure is initiated, then only we can start proceeding that as a landmark for design.

So, that is a very interesting argument which we had in this couple of lectures and learnt how
different failure theories give us different understanding of failure modes between fracture
and yielding, that is for ductile and brittle materials.

Thank you very much, have a good day, bye.

88
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 06
Material Properties - 1

Let us see in the 6th lecture, what we are going to learn in the course of Advanced Steel
Design of Structures.

(Refer Slide Time: 00:28)

So, in this lecture, we are going to learn more about material characteristics, paying attention
to steel as a very friendly, common, and intuitive construction material. To be very precise we
will focus only on steel. We will also talk about functionally graded materials in the
subsequent lecture. So, this will be material characteristic, lecture 1. We will have two
lectures on this, we will focus on this.

So, friends let us recall that what you have learnt the last set of lectures is that failure theories
will help us to quantify the stress and the failure state. It also helps us to declare the type of
failure, essentially, is it a brittle or a ductile failure. We will also have some interesting link
towards the stress state.

89
So, we have learnt the theories of failure. We understood the complexities and disagreements
of various theories, in different quadrants of plots of these theories. We will quickly see one
design example to highlight the importance of the failure theories.

(Refer Slide Time: 02:32)

So, we will see quickly a design example to understand the influence of failure theories on
the design. We will say example 1. The example is like this, design or compare, the diameter
of a steel shaft, subjected to torsion take Poisson’s ratio as 0.3. Compare this using various
theories that is maximum principal stress theory, maximum strain theory which is otherwise
called as Saint Venant’s theory, maximum shear stress theory which is otherwise called as
Tresca’s theory and using maximum strain energy theory.

Let us try to answer this question and understand the influence of choosing an appropriate
failure theory on design of the circular shaft.

90
(Refer Slide Time: 05:07)

Let us make certain assumptions. The first assumption what we make is let the yield strength
of the material in tension be identically equal to that in compression. Two, for a pure shear
case we can always say the principal stresses are equal to the shear value.

Now, the conditions for yielding because yielding is a type of failure, we already said that.
The conditions for yielding according to the above 4 theories are as follows. σ1 = σ𝑦𝑝, this is

as per maximum principal stress theory. σ1 - µσ2 = σ𝑦𝑝, this is according to maximum strain

theory.

2 2 2
σ1 - σ2 = σ𝑦𝑝, this is going to tresca’s. σ1 + σ2 − 2µσ1σ2 = σ𝑦𝑝 as per maximum strain

energy theory for a biaxial stress state. We have already seen that in the last lecture.

91
(Refer Slide Time: 07:31)

For a pure shear case, as we have assumed σ1= σ2 = τ, and hence let us call this equations as
equation 1. Equation 1 can now become σ1= σ𝑦𝑝= τ𝑦𝑝, this is for maximum principal stress
σ𝑦𝑝
theory. σ1 - µσ2 = σ𝑦𝑝. So, τ𝑦𝑝- µ (-τ𝑦𝑝) = σ𝑦𝑝, which means τ𝑦𝑝 = (1+µ)
. This is according

to maximum strain theory.

(Refer Slide Time: 08:54)

92
 σ𝑦𝑝
σ1 - σ2 = σ𝑦𝑝. So, τ𝑦𝑝 − (-τ𝑦𝑝) = σ𝑦𝑝. So, τ𝑦𝑝= 2
. This is according to Tresca’s theory.
2 2 2
σ1 + σ2 − 2µσ1σ2 = σ𝑦𝑝.

2 2 2 2 2
( )
So, τ𝑦𝑝 + τ𝑦𝑝 − 2µτ𝑦𝑝 − τ𝑦𝑝 = σ𝑦𝑝, which is (2 + 2µ)τ𝑦𝑝 = σ𝑦𝑝,

σ𝑦𝑝
which say τ𝑦𝑝 = . This is as per maximum strain energy theory.
2(1+µ)

(Refer Slide Time: 10:23)

Having said this for µ = 0.3, let us substitute these values in equation 2. Let us call these
equations as 2 a, 2 b, 2 c and 2 d. Let us substitute them. So, 2 a will yield τ𝑦𝑝 = σ𝑦𝑝, 2 b will

yield τ𝑦𝑝 = 0. 76σ𝑦𝑝, 2 c will yield τ𝑦𝑝 = 0. 5σ𝑦𝑝, and 2 d will yield τ𝑦𝑝 = 0.62σ𝑦𝑝.

Let me call this as equation 3. So, this is 3 a, this is 3 b, this is 3 c, this is 3 d. Let us apply
this for the design and see what happens.

93
(Refer Slide Time: 11:40)

Now, for the design of circular shaft in tension, let us say the allowable working stress in
τ𝑦𝑝 16𝑀𝑡
shear be τ𝑤 = 𝐹𝑎𝑐𝑡𝑜𝑟 𝑜𝑓 𝑠𝑎𝑓𝑒𝑡𝑦
. For torsional moment, acting on the shaft this τ𝑤 = 3 for a
π𝑑

circular cross section, the standard equation called equation number 4. Let us say we call this
as 4, we call this equation number 5.

Now, for the maximum principal stress theory,

τ𝑦𝑝 16𝑀𝑡 σ𝑦𝑝
τ𝑤 = 𝐹𝑎𝑐𝑡𝑜𝑟 𝑜𝑓 𝑠𝑎𝑓𝑒𝑡𝑦
= 3 = 𝐹𝑎𝑐𝑡𝑜𝑟 𝑜𝑓 𝑠𝑎𝑓𝑒𝑡𝑦
calls this equation number 6 a.
π𝑑1

16𝑀𝑡
Now, for maximum strain theory 3 , let us say the diameter from maximum principal
π𝑑2

theory is d 1 and for the maximum strain theory is d 2, and so on, will be now equal to
16𝑀𝑡 0.76σ𝑦𝑝
3 = 𝐹𝑂𝑆
. Why 0.76? Please see this equation, 0.76 σ𝑦𝑝, Let us call the equation 6 b.
π𝑑2

94
(Refer Slide Time: 14:35)

σ𝑦𝑝 16𝑀𝑡
So, one can rewrite this equation 6 b as 𝐹𝑂𝑆
= 3 , I call this as equations 6 c. Now, let
0.76π𝑑2

16𝑀𝑡
me compare equation 6 a and 6 c. So, what is 6 a? This is 6 a which is 3 . And 6 c is also;
π𝑑1

the left hand side are equal. So, let us compare them.

16𝑀𝑡 16𝑀𝑡
So, we can say that 3 = 3 . So, when you do this; so, let us say we can simplify
π𝑑1 0.76π𝑑2

1
3 3 𝑑1
this as 0. 76 𝑑2 = 𝑑1. So, (0. 76) 3 = 𝑑2
= 0. 913. so, this tells me that d 1 to d 2 is at the

ratio of 1 is to 1.095.

So, now, let us take condition c, this is for the maximum shear stress theory.

95
(Refer Slide Time: 16:24)

16𝑀𝑡 0.5σ𝑦𝑝
So, that is 3 = 𝐹𝑂𝑆
. Why 0.5? For 2 c, the condition is τ𝑦𝑝 = 0. 5σ𝑦𝑝. So, I call this as 6
π𝑑3

d.

16𝑀𝑡 16𝑀𝑡
Now, compare 6 a and 6 d. So, that tells me 3 = 3 . Am I right? So, simplifying 16
π𝑑3(0.5) π𝑑1

3 3
Mt goes away, π goes away, so 𝑑1 = 0. 5 𝑑3. So, now, I can say 𝑑1 and 𝑑3 will be at the ratio

of 1 is to 1.26.

(Refer Slide Time: 18:11)

96
 16𝑀𝑡 0.62σ𝑦𝑝
Let us do this for the 4th case, that is using strain energy theory 3 = 𝐹𝑂𝑆
. You can see
π𝑑4

16𝑀𝑡 0.62σ𝑦𝑝
it is 0.62. So, I can now say 3 = 𝐹𝑂𝑆
I call this as 6 e.
π𝑑4

16𝑀𝑡 16𝑀𝑡
Now, I compare 6 a and 6 e, so which will be 3 = 3 . This tells me 𝑑1 and 𝑑4 will
π𝑑4(0.62) π𝑑1

be in the ratio of 1 is to 1.17. So, now, friends I have the ratios of 𝑑1 and 𝑑2, 𝑑1 and 𝑑3, and

𝑑1 and 𝑑4 .

(Refer Slide Time: 19:47)

Let me quickly say 𝑑1 with respect to 𝑑2, 𝑑3 and 𝑑4 will be given by this relationship as 1,

1.095, 1.26, 1.17. Am I right? 1.17, 1.26, 1.095. So, that is the influence of choosing the
diameter of the circular, shaft subjected to torsion from different failure theories. So, friends
you see how the dimension or the cross sectional properties of the member vary significantly
when you choose a failure theory.

So, failure theories affect the design, design of sections, significantly. It affects the plastic
design of sections which I will show you later. So, this is a very simple design example to
illustrate the influence of the failure theories on design of steel sections. Having said this, let
us move on to the material properties or I should say material characteristics. Here we are
going to talk exclusively about steel as a construction material.

97
(Refer Slide Time: 21:23)

Now, friends we all know that different types of materials are used in construction. They are
used for a variety of purpose. For example, installation and commissioning, you can also use
material for repair and rehabilitation, you can also use material for corrosion protection
etcetera. So, the application of material in construction industry is widely different. So, you
cannot have a common material which can suffice the requirements of all types of needs of
the construction industry.

(Refer Slide Time: 23:02)

98
So, a wide range of materials like composites, concrete, steel, copper, aluminium, titanium,
fiber-reinforced, plastics and other buoyancy materials are used in construction. Now, the
question is when you have got such a wider choice of material then which governs the
selection.

(Refer Slide Time: 23:59)

So, the choice of the material for construction depends on the application requirements. So, it
is a new dimension we are talking about now. We do not take material first and then design
the member. We first choose the functional characteristic of the member, then choose
appropriate material.

One may ask me a question, if the member has multifunctional characteristics, then how do I
choose the material. The answer is very simple. Look for the dominant functional
characteristic. So, we say the material choice depends on application requirements. That is a
very important statement we want to make here.

99
(Refer Slide Time: 25:32)

Now, when we talk about the environment, then material choice becomes complex. I can give
an example. Say for marine environment, no single material or no single material
characteristic will help us to choose the material. Therefore, a detailed knowledge of the
structural characteristics and the relevant code compliancy are important.

So, now, two factors that govern the material choice. One is the functional application, that is
the broad objective, within that the two factors could be, code compliant, I mean the material
should be advised by the code, international codes, advice materials. If the material chosen by
you does not reflect any international guidelines for the design you cannot use the material
for construction. So, code compliance is very important.

Next would be the structural characteristics of the material. The moment we say structural
characteristics, then the question comes what are those structural characteristics which are
important.

100
(Refer Slide Time: 27:39)

Before we look into them, let us say, in addition, when we talk about form dominant design,
complexities increase. I will tell you why. Because form dominant design are subjected to
severe or extreme events. They undergo large displacements. So, let me reiterate, form
dominant structures resist loads by geometric form or arrangement of members. They do not
resist the loads by the material strength. That is a very important property we learnt in couple
of lectures earlier.

So, they do undergo large displacements, in fact, they are permitted to undergo large
displacements and their large displacement only help them to counteract the loads provided
they ensure recentering. Furthermore, they are also subjected to combination of loads such as
wind, wave, ice loads, fire loads, blast loads etcetera.

101
(Refer Slide Time: 29:43)

When we talk about explosion and fire, then one is interested to know the material
characteristic at elevated temperatures. So, we are looking for a high level of reliability in the
design. That is the foremost objective we have in the design. Therefore, we can say an
interesting statement, design of form-dominant structures is complicated due to the materials
in use. You may ask me a question why it is so. Please recollect the statement.

Materials are chosen based on application requirements, based on environmental conditions,

based on code compliancy, and this should now tally with the design. So, earlier the
conventional design is the other way. You choose a design procedure and material is selected
from a limited option and design is carried out.

So, it is a reverse process whereas, in the present scenario it is entirely different. That is why
we say when we talk about advanced steel design of structural systems, the design of
structural system is no more conventional. Because structural system itself is getting
innovative and novel because of form dominant aspect of the design which is a very very
common and very recent innovation in structural system design, in structural engineering.

102
(Refer Slide Time: 32:30)

Having said this, we now agree that materials used in such systems or such structural systems
require unique characteristics. They should be compatible with extreme weather conditions.
They should sustain environmental challenges and apart from that, they should be code
compliant and they should be cost-effective.

That is the two inherent requirement running parallel in a designer’s mind, when we talk
about application material for any structural systems.

(Refer Slide Time: 33:44)

103
Having said this, we now agree that structural design of such members or structures is not a
conventional design. The conventional design used to be strength dominant or strength
governed design. So, in such strength governed designs, the design procedure or the design
process is dominated by material characteristics.

So, in such designs people used to choose alter, research, material characteristics; improve
material properties, enhance them to apply to the conventional design. Whereas, in the latest
form dominant systems the process is reversed. The material should possess certain special
characteristics. We should shoot the form dominance of the whole design.

Having said this, let us come to the selection of materials.

(Refer Slide Time: 35:46)

Now, we agree that there is a wide choice of materials for construction. We already listed
them and we all agree on the statement. But few parameters govern the choice. Namely, type
of application you are looking for. Two, structural properties of the material under regular,
cyclic, high and low temperature loads.

We are also interested to look at the recycling characteristics of the material. We wanted to
ensure that the material remains non-toxic and bio-ecofriendly. The material should also be
sustainable for the entire service life of the structure.

104
(Refer Slide Time: 37:32)

It should be environment compatible, especially when you talk about marine involvement. To
be very specific corrosion is a very severe problem in marine environment. And above all
whatever material is used, it should be code compliant codes should advise recommend this
material for construction. Water may be the code, Indian international. There should be a
code confirmation of using this material for that kind of application.

You may wonder how codes specifically say this material is useful for this kind of
construction. I will cite certain examples and codes, where there are international codes
which specifically say use this material of this characteristic for this kind of members. They
are very very specific in recommending them for design.

Therefore friends, in the structural design we all agree that code requirements govern the
design, but use of advanced materials are also welcome in the code. So, advanced materials
are also inserted as an optional replacement of the existing conventional material.

So, when you talk about advanced, please think that, steel is not the only non-replaceable
material for construction. I am going to give an alternative in this course. We will see how
other materials can also replace steel with better structural characteristics. We are going to
talk about that.

105
(Refer Slide Time: 40:01)

Functionally graded materials, abbreviated as FGM are the recent advancements in structural
engineering, in material perspective. Now, the question is, is FGM recommended by any
code? As of now the answer is a big no. But, however, when the material research is
supported with enough data and validation, codes do consider them for design. That is the
recent advancement we have.

(Refer Slide Time: 41:33)

I can give a very simple example of FGM. We will talk about this in detail in the coming
lectures. Simple example of FGM is that FGM under cyclic loads is not available in the

106
literature. The structural properties of FGM and the cyclic loads is not enough it is very
scarce in the literature.

Therefore, use of FGM in marine applications where cyclic load or reversal of loads is
predominantly present is restricted as of now. But still, to assess the fatigue life, a
conservative approach can be used, to compare the results and present it to the research board
of material innovations. Material characteristics influence the design and hence choice of
material for construction is a major engineering decision. So, that is a very interesting
declaration we want to emphasize upon for the learners.

So, far in the conventional design, people were thinking that we have been given about
choice of 1 or 2 materials. Let us say reinforced concrete, structural steel, sometimes
composites. Use the relevant core, find out the critical stresses, check them for bending shear,
torsion etcetera and let us say we declare the design is safe. That is all we have been doing for
ages.

But under the present era friends, it is not sufficient. We have to look for advancements in the
design, in the structural geometry, and application oriented structural systems, where the
material choice governs the design and that becomes one of the major engineering decision in
the whole project.

(Refer Slide Time: 44:34)

107
So, in this lecture friends, we learnt one design example of influence of failure theories on the
design. We have also agreed that material choice for construction is a major decision.

Thank you very much. Have a good day.

108
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 07
Material properties - 2

(Refer Slide Time: 00:22)

Friends, welcome to the 7th lecture on advanced field design course. This lecture is again
going to focus on Material characteristics. We are going to again emphasize on steel as a
construction material as per our engineering choice. So, we learnt about something off form
dominant design, which is quite effective via innovative and recently popular in construction
systems.

We already said that form dominant design or form dominant approach is not a new one, but
arches, cable stage bridges, large span structural systems using thin shell members were all in
practice for years and years. We are now formally trying to understand how this can be
helpful in alleviating the applied loads on the system and we are now going away from a
conventional approach of strength-based design to geometric based design.

So, the material chosen for this purpose should support the challenges what the structural
system shall encounter and what are those characteristics which makes steel as special that is

109
the focus. So, we understand now environmental loads and their combinations create huge
complexity in design.

For example, wave load, wind load, earthquake load, ice load, impact load, fire load, etcetera,
are few examples which can get combined as the worst and rare events to cause huge
complexities on the structural members. So, when we say this that form dominant design is
able to sustain this combination successfully in parallel materials used in such systems also
face challenges.

(Refer Slide Time: 03:59)

So, one of the most important requirement is that material should not fail or should not
undergo catastrophic failure. Please re-emphasize the statement, I am talking about the
material failure not the structure. Material should not undergo a catastrophic failure.
Catastrophic in sense to make it very simple it should not initiate a sudden failure that is the
foremost.

The other requirements could be it should possess a good fatigue resistance, then it should be
it should be largely available in desired shape, size, modular sections, apart from being cost
effective. We are looking for these kind of objectives when we choose materials for
construction.

110
(Refer Slide Time: 05:33)

So, let us try to pay attention on the mechanical properties of the material, in general and
steel in specific. We must agree on a common statement that, material choice has a very
strong relationship with type of structure. You cannot apply a common material to all types of
structures because the functional requirement of varieties of structures are different.

Therefore, the material choice has got a very strong relationship, one to one correspondence
to the type of structure you are going to design, we can give some example. According to
American Bureau of Shipping, which will popularly called as ADS which is one of the
important guideline of governing design in United states, recommend prefer to recommend.

The following characteristics for design of structures in marine environment recycling,

sustainability and nontoxic in nature in addition to the classical mechanical properties which
are essentially required for a material to use as a structural material. We cannot overcome
though those properties. They anyway exist. They are in addition to that, these are in addition
to that.

111
(Refer Slide Time: 08:05)

Further they also emphasize on physical and chemical properties, cost effective, fabrication
facilities and their viability of fabrication and of course, most importantly the maintenance
cost. In addition, the chosen material let us emphasize this again, should not possess or
should not undergo catastrophic failure. That is an important statement. Then what does it
do?

It should give enough warning before failure. Friends, one may wonder that how material can
give warning. Structural system gives warning as per the design we know we can recollect
them.

112
(Refer Slide Time: 09:39)

When you talk about reinforced concrete design, we hear something called under reinforced
sections, we also have something called over reinforced sections and we have an hypothetical
case which is called a balanced section. So, the principle in this hall the three is if C stands
for the total compressive force acting on the cross section and T stands for the total tensile
force acting on the cross section, for a balance section they should be identically equal, for a
balance section identically equal.

But we prefer to design an under reinforced section considering that under reinforced sections
where tension governs the design gives sufficient warning before it fails. This was the design
philosophy what we had in Indian Code 456 right from 1974 till even 2000 year of revision.

So, I agree that the design can emphasize that it should give sufficient warning, it does not
talk about the material is the choice of the section understand that. But what I am
emphasizing here is the material should give warning that is what we are looking at. So, it is
we are looking at different philosophy.

113
(Refer Slide Time: 11:47)

Furthermore, materials should withstand hazards that R is during installation and operations.
Under this context what are those mechanical properties which are important? These are the
requirements. Let us say which mechanical properties reflect this requirements, let us see
them. Yield strength, modulus of elasticity, Poisson’s ratio.

Because this is important for multi axial loading systems which is very common, fatigue
performance or fatigue strength, because this is important for cyclic loads which is also
necessary in the present case. We look for strength of the material, we look for hardness of
the material, we look for toughness of the material, we also look for elasticity property of the
material.

114
(Refer Slide Time: 13:24)

We look for plasticity property of the material, we also check the brittleness nature of the
material and we also be sure the material should have enough malleability and of course,
ductility. So, these properties listed will help us to choose a variety of materials available in
the literature so that they can be sufficiently placed in the structure to encounter the forces
acting on the member.

So now, if you ask me a general question, what should be the strong focus in material choice?
The strong focus in material choice is the functional and structural degradation. So, the
material what we choose should not functionally and structurally degrade as ages of the
structure.

So, this can be checked with respect to the environment. So, we are now looking for a system
which can sustain the forces with the help of geometric complacency. In addition to that the
material should also support the system to sustain them in the given environment, without
undergoing functional and structural degradation.

At any point of time the material should not expose and initiate a catastrophic failure, it
should give sufficient warning before it fails.

115
(Refer Slide Time: 15:56)

Having said this when we talk about the interaction of environment on the material, let us see
now, what are the effects of environment on the materials. We have already agreed upon one
fact that form dominant design requires special characteristic materials, we already said that.
We already agreed that form dominant designs need special characteristics of the materials.
We already agreed this.

So, now large displacements which are a common feature in this kind of compliant structures
or form dominant structural systems, demand, high ductility of the material. Cyclic loads
acting on the structure which arise from the environment for example, wave load is a cyclic
load. There is a possibility of reversal of wave loads in terms of its direction amplitude and
both. They demand fatigue strength of the material high fatigue strength of the material.

So, friends please see we are trying to relate the effect of environment and material this is
from the environment, this is from the form dominant compliancy of the structural system
design.

116
(Refer Slide Time: 18:24)

So, if you ask me a question what is one of the major concerns in materials? One of the major
concerns in the material choice in environmental approach or environmental influence, let us
put it like this. Under environmental influence, materials undergo strength degradation.

So, let us quickly see what are those degradations. They undergo chemical degradation,
strength degradation, loss in fatigue resistance or fatigue strength, poor performance under
high stress concentration, corrosion and of course, bio fouling. Is one of the major concern
the recent trend of Environmental Protection Act that the material used in marine
environment specifically should be bio friendly.

So, bio fouling effects are a major concern when you want to choose a material for
construction or repair or even painting of let us say ships, offshore, structures etcetera, the
material choice is to be done very carefully.

117
(Refer Slide Time: 20:52)

In addition to that, what are other general concerns that govern the material choice? The other
concerns are material should possess survivability under accidental loads. Even under
extreme cases like hurricane they should be able to withstand very high hydrostatic pressure.
They should not initiate any catastrophic failure. It should give enough warning before
failure. The failure should not be sudden even under accidental loads.

(Refer Slide Time: 22:47)

Now, let us focus on the design considerations with respect to materials. We have looked
upon the environmental considerations on materials, now we look into the design

118
consideration on materials. Following properties of the material or following characteristics
are important in design perspective.

1 the yield strength of the material, 2 Modulus of elasticity, 3 Poisson’s ratio, 4 fatigue
strength of the material and 5 fracture resistance of the material. In general, if you ask me a
question, how this is handled in design.

(Refer Slide Time: 24:07)

How these are handled in design? Codes handle them slightly in a different manner. Codes
recommend material allowance, which can encounter or which can counteract the
environmental issues on materials. Let us quickly ask a question what is material relevance.
Material relevance is actually recommendation of extra thickness of the member, required
more than the design.

This is one way of doing it. One way of doing it is material relevance. The second way of
doing it is of course, we all know that you can recollect them that is in procedure. We use
what is called partial safety factor to take care of these effects in the design. This is not new
we have been doing this. Friends, let us come to an important segment of understanding now.

119
(Refer Slide Time: 25:53)

If steel is taken as the most preferred material for engineering construction for engineering
construction and it is also accepted to be environmentally influenced in a lesser manner then
interestingly steel is classified according to its application.

(Refer Slide Time: 26:48)

So, now let us see classification of steel. Friends, all codes do not do this. Major codes major
design codes around the world do this classification. I will give you the name of the codes
slightly the end of this lecture, but major international codes and Indian Code also do this
classification. Let us first understand how this classification is done.

120
First let us try to learn that many international codes design codes, classify steel according to
application of the members. To be very interesting this is more common practice in offshore
engineering. And I am sure in land-based structures this application or classification is not
common, but this is very common in offshore engineering.

For example, API-RP-2A American Petroleum Institute-Recommended Practice-2A the year

of revision 2000. Classify steel for various applications of offshore structural members. For
example, what should be the material or classification of steel for piles, what should be the
classification of steel for bracing and patterns, what should be the classification of steel for
the top deck for the hull of the vessels, propellers of the vessel and ships, etcetera.

It is very very interesting and they do in depth classification depending upon the application
perspective of the material. Probably for many of you this may be the first time to know
really the steel is classified based on application. So, far we have been learning only on
conventional steel Fe 250, Fe 300, Fe 450, Fe 550 etcetera, where we learn only the focus of
classification based on strength or yield strength.

There are many varieties of classification friends. international codes are very advanced in
this perspective. So, this course will help you to get an insight of this which is one of the
important objectives of this course actually. So, let us see how this classification is done.

(Refer Slide Time: 30:08)

121
So, now, let us ask a question, how steel is classified. Steel is classified based on its
composition, its manufacturing methods, finishing methods, microstructure, strength, heat
treatment, product form. So, one can say that classification of steel is one of the advanced
ways to initiate appropriate material. That is appropriate steel grade for application and this is
a very common feature in offshore engineering.

Let us quickly elaborate them and see how they are really classified, let us take one by one.

(Refer Slide Time: 32:10)

Let us say based on composition how they are classified. Based on composition it is classified
as, carbon steel. The moment I say carbon steel I have again a variety; low, medium, high and
ultra-high. Low alloy steel and stainless steel, these are based on their composition. Now
based on the manufacturing methods, they are classified as electric furnace steel and
open-hearth process steel.

122
(Refer Slide Time: 33:24)

Based on the finishing methods, they are classified as hot rolling, cold rolling. Now based on
micro structure, they are classified as ferritic, pearlitic and martensitic.

(Refer Slide Time: 34:26)

Based on strength, they are classified as different types and they vary with different codes. I
will come to that slightly later. They are also classified based on heat treatment, as annealing,
quenching, tempering. Based on product forms, they are classified as bars, plates, sheets,
strips, tubes, L section, T section, channel section, etcetera. Friends, let us pay more attention
towards how the classification of steel is based on strength more in detail.

123
(Refer Slide Time: 36:01)

Let us say classification of steel based on strength. Now, before we look into the
classification directly let us see what would be those factors which will help the code to
classify steel based on strength, because strength is actually a load dependent phenomena.
Strength is the capacity to withstand the applied load that is how the basic definition goes in
mechanics.

So, we have to look into the factors which will help us to classify steel based on strength. The
strength is actually not a single independent parameter or characteristic of material. It
depends on the application of load on the material stress state and so on and so forth. So,
based on the component and type of combination of loads codes classify steel according to
yield strength.

2, carbon content present in steel composition plays a very important role in classifying steel.
You may ask you a question, carbon being a chemical composition how does it relate to
strength? Carbon content governs the strength of steel my friends. So, therefore, carbon
content in the chemical composition helps to classify steel based on strength. For example,
low carbon steel which has carbon lesser than 0.3% and it does not contain other elements
like chromium, cobalt, nickel etcetera. It classifies this way.

124
(Refer Slide Time: 39:00)

Second: Medium carbon steel, where the carbon content varies from 0.3 to 0.6 % carbon
content, high carbon steel, where the carbon content varies from 0.6 to 1 % and ultra-high
carbon steel, where the carbon content exceeds 1.25 to 2%. So, we have learnt an important
statement here that, carbon content influences strength of steel because under the
classification of strength carbon content is being identified. So, let us see in detail about this.

(Refer Slide Time: 40:17)

Low carbon steel is also referred as low strength steel, which has got yield strength is about
415 Mpa which is as same as 415 newton per mm square. Now it has got specific

125
applications. We already said that steel is classified based on applications, the applications
recommended by the code or hull of ships and platforms offshore. Platforms, fittings, tanks,
instrument ancillaries, and of course, buoys.

Medium carbon steel has got yield strength. It is about 1035 Mpa. This is recommended for
fabrication of ice breakers and buoys in arctic region.

(Refer Slide Time: 42:16)

High carbon steel has an yield strength about 1100 Mpa. A classical example of this steel is
maraging steel is an example. Maraging steel has σ𝑦 about 1 to 2 Gpa. There is an issue with

this kind of steel. High carbon steel is relatively less ductile compared to other steel.

It means, increase in carbon content decreases the ductility. So, what they do? To overcome
this problem, they do heat treatment during manufacturing to improve its ductility.

126
(Refer Slide Time: 43:42)

So, as I said a varieties of codes help in classifying steel. The list of codes which does
classification of steel as I discussed just then are as follows. One is Euro Code 10225, Euro
code for design of structures. British Standard 7191, then NORSOK which is a Norwegian
material data sheet used for design of ships and offshore structure.

And of course, American Petroleum Institute used in America and Asian regions and of
course, IS 1762 Part-I; the Indian Code 1974 which is the code for designation of steel.

(Refer Slide Time: 45:03)

127
Friends it is very important that we learn to know all these codes have cross
recommendations. What does it mean? It means that they have mutual inter-reference. So, the
final summary is they have a close agreement in classification with each other. So, when a
specific code classifies steel based on strength, the other code does not deviate much from
this classification. They mutually agree closely on this.

And we are very proud that we are also advanced in this front and our Indian Code also helps
to identify classification of steel based on very explicit parameters. I am sure how many of
you would have gone through this course in the design classes. So, I think it is now high time
for us to really understand update our advanced knowledge in design of steel structures and
this course is certainly going to help you in that front.

I am sure that you are enjoying the new contents discussed in these lectures the way in which
they have been discussed I am following a very simple classroom model, I am using a white
board and I am writing and speaking and explaining and deriving all equations line by line,
which I will practice in the entire lecturing of this particular course.

Whenever there is an example, we will also help you to use MATLAB codes. So, I will intuit
a good level of confidence and a very strong understanding and therefore, love for this
particular course of NPTEL.

(Refer Slide Time: 47:37)

128
So, having said this we will also now talk about how steel is grouped. So, steel is also
grouped. This is different from classification. Please understand grouping of steel is different
from classification. How they are different, how they are grouped, we will see in the next
lecture.

(Refer Slide Time: 48:15)

So, in this lecture as a summary we have learnt various material characteristics, which are
application oriented which help us to choose steel for construction purposes. Two, we have
also learnt how steel is classified. We have also learned what are those important mechanical
properties that are useful to choose steel in the design perspective. We have also learnt how
carbon influences strength of steel.

So, friends we will continue this in the next lecture and talk more about their material
characteristics at normal and elevated temperatures and we will learn more about the material
selection as recommended by Indian and international codes.

Thank you very much. Have a good day, bye.

129
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 08
Material properties - 3

(Refer Slide Time: 00:17)

Let us continue to discuss something about the Material properties which are very special to
select materials like steel which are applicable to form dominant design principles. Steel is
the most preferred material for construction of conventional and special types of structures.
Steel has got lot of advantages which we have been discussing in the previous lectures and
we all agree on that the steel has gained an importance in the construction industry.

When it comes to special applications like industrial structures, strategic structures like
nuclear power plants, large in type of industries cement industries tall buildings as well as
offshore oil and gas exploration and production platforms, shipbuilding structures etcetera.
So, we have also seen in the previous lectures that the how steel is classified what are the
different international codes including Indian code that help us to classify steel based on
certain special characteristics and properties that steel possess.

So, steel is not identified as an entity it has got lot of special characteristics, lot of subgroups,
classifications, groups, etcetera. So, when you say I am using steel as a construction material

130
in the present scenario it has no value. In fact, it does not make any sense you got to
specifically say what grade of steel are you using, what chemical compositions that steel has
and what are its mechanical and structural characteristics that make it as a desirable material
for the construction process you have chosen.

So, steel is a large set within which there are lot of classifications which we discussed in the
last lecture. We will now continue to do that and try to understand the properties of steel at
elevated temperatures in this lecture. So, we will talk about special properties of steel at
elevated temperatures. We will also talk about use of new novel construction material for
special functional use which we are going to say as FGM.

(Refer Slide Time: 04:06)

So, with the last note we had in the previous lecture we said there are many international
codes that classify steel based on certain properties. In fact, to make a special emphasis
Indian code 1762 (Part 1)-1974; also helps us to designate steel for construction purposes. It
is important to note that these codes have lot of cross references and they are in very close
agreement to each other. So, the classification of steel it is a globally accepted phenomena
and many countries do follow this. Having said this apart from classification of steel; steel is
also grouped.

131
(Refer Slide Time: 05:52)

Now how are they grouped? Based on what parameters steel is grouped? Steel is grouped
according to its strength and welding characteristics. Let us say how are they grouped? Group
I states those steel which has got a minimum yield value as 280. So, friends it is very
important that steel whose σ𝑦𝑝 is less than 280 is not recommended for form dominant design

structures please note. Say Fe-250, the classical mild steel is not usable for special and
strategic structures. The first requirement is σ𝑦𝑝 should be minimum of 280 and the carbon

equivalency should be less than 0.4% by weight.

(Refer Slide Time: 08:08)

132
Group II is that steel whose yield value is up to 360 Mpa and carbon equivalency is up to
0.45% by weight. Group III steel it is also referred as high strength steel in the literature; this
has yield value exceeding 360. Now this kind of steel has got special welding requirements
which are emphasized during fabrication. We should also make a note that members design
with group III should be investigated for fatigue resistance that is a condition as a routine
design check.

(Refer Slide Time: 10:22)

Another important property what steel possess which is also useful for us is steel possesses,
low temperature toughness. This is helpful to avoid brittle failure at the joints. So, now I talk
about toughness and we talk about carbon equivalency. We spoke about two things now
toughness, carbon equivalency.

Carbon equivalency is a criteria which governs weldability requirements I will come to that.
Toughness is a criteria which helps us to check the capability of the material under cyclic
loads. So, now, most of the international codes specify impact test properties for steel using
Charpy impact test. So, charpy impact test is considered as a standard procedure to qualify
steel as a construction material when we talk about strategic and other important structures in
the advanced scenario of design.

Furthermore, they also emphasize on a fact that steel should possess good crack tip open
displacement properties. This test is briefly known as CTOD test. So, please be very careful
in specifying steel or recommending steel as a construction material. The mechanical strength

133
alone does not qualify this material for construction purposes in the present scenario of
advanced steel design.

You have got to ensure that the material has good toughness the material possesses a carbon
equivalency that makes it weldable comfortably and it should also have desirable CTOD
properties as specified in the international codes. I will come to the specific details in the later
part of this lecture.

(Refer Slide Time: 14:01)

So, if you ask me a quick question what is CTOD test? Crack Tip Open Displacement test is a
family member of test conducted in the domain of fracture mechanics. What does this test
do? This test is used to measure resistance of the material to crack growth. Friends initiation
of crack is a different scheme once the crack is initiated it propagates it grows. CTOD test is
an index which will help you to qualify steel based on its resistance against growth of crack.
So, you can very well see here we are imposing certain serviceability conditions not in the
design scheme.

But in the material scheme itself like the material should be non-corrosive, the material
should sustain cyclic loads, the material should sustain or should oppose a crack growth and
the material should be weldable and ductile which are all earlier considered to be the design
properties, which are earlier considered to be nomenclature recommended in the structural
design.

134
But in the latest trend the recent phenomena of structural design they have all moved to
qualify the material for construction itself. So, one can say very well here we are slowly
moving the design procedures from a conventional strength base to form base and to a larger
extent material specific.

(Refer Slide Time: 16:42)

The next interesting property which steel is expected to possess is the toughness. Toughness
is an important property which is considered to be necessary when you use under marine
environment. It is a desirable property otherwise when used in marine environment this is
compulsory.

One may then quickly recollect what is toughness? Toughness is described as resistance to
failure in the presence of a notch a crack or any similar stress concentrator. So, in general
what do we prefer? We prefer high toughness for the material.

135
(Refer Slide Time: 18:34)

So, how toughness is expressed? Toughness is expressed as two ways impact toughness, 2,
fracture toughness. Let us quickly also try to recollect the meaning of these two toughness.
Impact toughness is the energy measured in joules and determined using charpy’s impact test.

I am not anyway describing the procedure of charpy impact test most of you may be knowing
it or I would always request you to go to the website of leading institutions where there are
live videos uploaded the institutes website about charpy’s impact test procedures. So, you
could learn that separately I am not talking a believe that you know what is charpy’s impact
test. Fracture toughness is computed based on the CTOD test or J integral test.

So, mechanical engineers may know both of the test procedures very clearly. However, does
not mean that others may not be interested in knowing them can go through the open domain
websites of these days available to explain these test procedures in a nice video for about a
couple of minutes. How a charpy test is conducted and how the results of the charpy test are
interpreted please learn both. How it is conducted? How the results are interpreted? Both are
important for us.

136
(Refer Slide Time: 21:17)

Having said this we have learnt the steel is classified based on certain properties by
international codes. Steel is grouped based on σ𝑦 and carbon equivalency. Steel has got

special properties like toughness which need to be investigated either using CTOD or using
charpy test.

These are all mandatory before you select steel as structural material for structural systems.
In addition to that friends to add to more complication to your knowledge steel is also group
based on notch toughness.

137
(Refer Slide Time: 22:14)

Based on notch toughness steel is grouped as class C, class B and class A. Now one may ask
me a question how notch toughness is estimated? Notch toughness is also computed based on
impact test by providing a notch in the specimen of a standard size using ASTM controls.
You conduct the impact test and estimate the notch toughness. Based on the notch toughness
obtained from the impact test can again group steel as class A B C.

(Refer Slide Time: 23:45)

Let us see what is this class refers to? Class C refers to a specific group of steel where no
impact test results are specified explicitly by the codes. Now where do we apply this? This

138
group of steel is applicable to primary members, primary structural members with limited
thickness that is very important here.

They can be also used in structural systems with moderate forming not large curved
structures where you require low restraints you can use class C steel. When you have got
moderate stress concentration, you can use this group of steel when the structure is subjected
to quasi static loading, you can use this group of steel. So, friends please see here very
carefully steel is grouped based upon the type of load, type of stress developed and also on
the shape and cross-sectional dimensions.

So, far I think I am sure most of you would not know and would have not agreed that codes
specify steel based on these conditions. We have been using steel as a simple material and
find out the stresses and check whether this material and the cross section can sustain the
bending stress, the shear stress and the torsion coming on to the cross section at the maximum
sections and the declare the design as safe and check for its serviceability requirements.
Those days of design have been gone friends.

We are looking for novelty in design advanced design procedures. So, we have to be very
specific and we must know how steel is grouped, so intricately even for a specific choice of
application.

(Refer Slide Time: 27:08)

139
Continuing this argument class B group of steel is recommended for members of large
thickness. You can also use this group of steel for members having high stress concentrations.

Now, getting back to this group class C examples could be piles, bracings, battens etcetera.
Class B group of steel can be used on sections or members which are likely to attract large
and higher stress concentrations. You can also use them where impact loads are expected
example could be boat landings, top side of offshore platforms, deck of cable stayed bridge
etcetera.

(Refer Slide Time: 28:52)

Now, let us talk about class A steel; Class A refers to a group of steel which is used in
sub-freezing conditions. These are applicable to structures built in arctic regions. Friends we
are slowly discussing the important mechanical and additional properties of steel which help
to qualify it as a construction material.

140
(Refer Slide Time: 29:34)

In that line we have one more important characteristic which is weldability as I said steel is
grouped based on carbon equivalency and yield strength. Carbon equivalency helps to qualify
steel for its welding characteristics. So, let us quickly see how this is approached in
international design codes. friend’s weldability is an important requirement of any
construction material steel it is very important because steel need to be welded or bolted.

They need to be connected the members need to be joined usually the joining is a welding,
but people can go also for bolted connection specifically required if you do not want a
welding connection at that particular location. So, weldability is essentially required.

Now how can you compute weldability? How can you check whether the steel is weldable or
not? This can be checked using cold cracking susceptibility test. You can also do this by
estimating the carbon equivalency from the chemical composition of steel. If you know the
components present in steel each component by weight. If you know that percentage, based
on that composition by weight one can compute carbon equivalency.

If you know the carbon equivalency value you already know how to group them. We
discussed that in the previous slides. Now we want to see how based on carbon equivalency
one can check its weldability. Let us be very specific about this objective at this moment. So,
now we can say very well clear here that carbon equivalency is a measure of the tendency of
the weld to form martensite on cooling and to suffer a brittle fracture. So, it is a measure to
assess this property.

141
(Refer Slide Time: 33:06)

Now, in general if you ask me a thumb rule. Sir when can we say that steel is weldable?
When the carbon equivalency or the carbon equivalent is between 0.4 to 0.6 by weight, then
one can say a weld preheat is necessary. When the carbon equivalency is more than 0.6 by
weight then preheat and post heat may be necessary.

I am talking about preheat and post heat of the member. So, we have a very interesting table
recommended by the literature. You can refer to me list of references in detail to know more
about this.

(Refer Slide Time: 34:20)

142
Let us open up a table carbon equivalency and weldability. If it is up to 0.35% by weight
these all by weight. Then we declare this the weldability is excellent. If it is between 0.36 to
0.4% by weight we call the weldability as very good. If it is between 0.41 to 0.45% by
weight, we call it is having a good weldability characteristic.

If it is between 0.46 to 0.5% by weight, we call the weldability as fair. If it exceeds 0.5 then
we say it is poor and not usable or not recommended. So, before you check the material for
construction especially steel look into the chemical composition of that product work out the
equivalency of carbon by weight and see whether the code recommends it or not. Blindly you
cannot use all grade of steel for construction.

(Refer Slide Time: 35:56)

Now, the question interestingly comes how do we compute the carbon equivalency? Carbon
equivalency as per API RP 2 A can be computed using the following relationship. Carbon
equivalency is equal to carbon plus manganese by 6 plus nickel and copper by 15 plus
chromium, molybdenum and vanadium by 5% by weight.

𝑀𝑛 𝑁𝑖+𝐶𝑢 𝐶𝑟+𝑀𝑜+𝑉
𝐶𝑒𝑞 = 𝐶 + 6
+ 15
+ 5
% 𝑏𝑦 𝑤𝑒𝑖𝑔ℎ𝑡.

The other parameter which is also used to check the weldability is a weld cracking parameter
which is called as 𝑃𝐶𝑀. 𝑃𝐶𝑀 can be obtained from the chemical composition relationship as

you see here carbon plus silicon by 30 plus manganese by 20 plus copper by 20 plus nickel

143
by 60 plus chromium by 20 plus molybdenum by 15 plus vanadium by 10 plus 5 times of B;
where B is the Boron is all again in percentage. For your curiosity if you really want to know
what the chemical composition of mild steel has for our knowledge sake.
𝑆𝑖 𝑀𝑛 𝐶𝑢 𝑁𝑖 𝐶𝑟 𝑀𝑜 𝑉
𝑃𝐶𝑀 = 𝐶 + 30
+ 20
+ 20
+ 60
+ 20
+ 15
+ 10
+ 5𝐵 %

(Refer Slide Time: 38:01)

Let us say ask this question what a chemical composition mild steel has? Let us take mild
steel mild steel should be designated by grade. We will not talk about steel as it is as mild,
hard, beautiful, no adjectives.

So, let us say mild steel is specify as per BS 970-1983 or as per BS 970-1955 grades. So, this
has the following composition. It has carbon as 0.16 to 0.18% by weight; it has manganese
0.7 to 0.9% by weight, it has silicon 0.4% by weight. Let us quickly compute the carbon
equivalency for this which will be C plus manganese by 6.

This is what I know because other compositions are not involved in contributing to the
carbon equivalency. So, we can quickly find out this as 0.6 plus let us say 0.9 by 6 which is
0.31% looking into this table up to 0.35 it is considered to be excellent. So, I can declare that
mild steel is having excellent weldability. For our additional information mild steel has
ultimate stress an ultimate strength let us be like specific is 400 to 560 Mpa.

144
Friends Mpa is as same as newton per mm square because mega is 10 power 6 and it is
newton by meter square I divide by 10 power 6. So, is as same as Newton per mm square. So,
400 to 560 Mpa yield strength of this is 300 to 440 there is a range, 0.2 % of proof stress.

I believe most of you would know what is the proof stress and what is 0.2% and this stress is
280 to 420 Mpa and percentage elongation. Let me write it here of mile steel is 10 to 14%
minimum.

(Refer Slide Time: 41:05)

Let us take one more grade of steel which specified by ASTM, A 36. They call this as low
carbon steel it is also referred as mild steel. This has carbon 0.25 to 0.29% by weight, it has
copper 0.2% by weight, it has iron 98% by weight, it has manganese 1.03% by weight.

When I try to work out the carbon equivalency of this I get this as 0.25 plus 1.03 by 6 plus
0.2 by 15 like what the equation. You will get this as 0.435. when you look back the table I
fall here. So, I can say this has got a good weldability. So, based on the chemical composition
of steel one will be able to declare easily about the weldability characteristics.

I think it is clear enough for you to understand how this is done. I am not getting into more
metallurgical part of it because I am not a metallurgist and I want you to know structural
oriented characteristics only. This course is focused thoroughly only on structure engineering
perspectives though we have to touch upon the material properties, but still I urge you to
focus only on those things which are essentially required to know not beyond that.

145
(Refer Slide Time: 43:07)

Now, friends we have learnt that for special application structures which are environmentally
governed, which are strategic which are novel in terms of form dominant. They are usually
subjected to a 3-dimensional stress state is it not? We have already said in a 2 dimensional or
3-dimensional stress state how complex it is to estimate the failure load. Because we have
seen the failure theories and failure theories disagree on different stress states, on different
nature of stresses etcetera is it not?

We have very clearly saw that and in the four quadrants the second and fourth quadrant the
disagreement is very severe. And in this one of the theories says even beyond σ𝑦𝑝. The failure

does not happen is it not? Please recollect those the lectures what we had and you will really
appreciate that how we are correlating the growth of understanding gradually to build your
knowledge level on advanced steel design.

146
(Refer Slide Time: 44:40)

(Refer Slide Time: 44:55)

So, now interestingly friends, steel has got desirable properties in different perspectives. So,
we talk about construction materials in general most of them are equally strong in
compression, tension and shear showing a marginal difference. Most of them are strong in all
these aspects. Now by carefully allowing them the metals we can achieve a good or higher
plastic strength or plasticity and a higher ultimate strength.

We can be able to do that you have to work on the alloying part. Metallurgists will strongly
agree and mechanical engineers to a larger extent will agree on my statement by changing the

147
chemical composition and manufacturing processes. Fabrication procedures of making steel
or manufacturing steel you will be able to achieve the better strength and plasticity of steel.

(Refer Slide Time: 46:37)

With this idea in mind in the literature in the very recent past about 2, 3 years back. People
have introduced what is called functionally graded materials for strategic structures. FGM is
not a new concept because this is being used in aerospace industry and special application
systems very long back. But when it comes to structural applications when it comes to
functional applications like anti corrosive etcetera it is a very very recent application friends.

So, now, FGM we will discuss in detail in the next lecture. One of the important fabrication
procedure to manufacture FGM, which is widely used is WAAM which is Wire Arc Additive
Manufacturing. We have a very strong and rich laboratory expertise in WAAM in the
department of metallurgy in IIT Madras.

So, those who are interested those who want to inspect and see how WAAM actually is
powerful in manufacturing FGM. You can email me we can arrange a short visit to the lab
when the experiments are conducted and fabrication is going on we will help you how FGM
can be fabricated using a WAAM process.

I will give you some more videos and some more lecture connectivity’s in the coming
lectures for you to know more about it, but keep your knowledge growth horizontal. So, that
you keep on learning more and more as you go ahead. So, WAAM is one important

148
manufacturing method which is very useful in manufacturing FGM. Now we will talk about
WAAM much in detail later.

(Refer Slide Time: 48:45)

Let us quickly go through specific properties of steel at elevated temperatures. We will pay
attention to this in the next lecture. But now we will write quickly the summary of this
lecture. We have learnt a few mechanical properties of steel which are important, we have
learnt how steel is grouped based on toughness, based on σ𝑦 and carbon equivalency, also

based on notch toughness.

We also learnt how to work out the carbon equivalency and how to check the weldability of
steel. And we have also seen what are those special requirements for steel as a construction
material. So, we close the lecture here have a good day bye.

149
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 09
FGM

(Refer Slide Time: 00:23)

Friends welcome to the 9th lecture on Advanced Steel Design course. We have been talking
about the material characteristics. In this lecture we will talk about functionally graded
material and also we will discuss about steel at elevated temperatures.

150
(Refer Slide Time: 01:01)

So, far friends a very good reference for the previous context of discussions we had can be
seen from these textbooks as shown on the screen. Advanced Steel Design of Structures
authored by me for CRC Press, Ocean Structures and Materials authored by me and Professor
Arvind Kumar Jain for CRC press and a recent book published by CRC on Design Aids for
Offshore Topside Platforms under Special Loads, which has got foreign authors as well in the
textbook.

So, these books contain lot of information for additional and enhanced learning on the topics
whatever we have. So, far discussed and I urge you to kindly go through them and improve
your readability in a wide horizon apart from what we discussed in the classrooms during
these lectures.

151
(Refer Slide Time: 01:57)

Having said this let us talk about steel property at a higher temperature. The figure what you
see on the screen shows the reduction factor on different parameters of steel when the
temperature is increased from about 200 degree centigrade to about 900. So, there is a
reduction factor applicable to yield strength, proportionality limit and slope of the elastic
range which will help you to compute the Young’s modulus.

See effective yield is given by the crayon color, proportional limit is given by the black color
and the slope of the linear elastic range of stress strain curve which is helpful to compute the
modulus of elasticity is given by the blue color. One can see here from this figure from the
plot the effective yield modulus of elasticity and yield strength decreases with increase in
temperature. So, all of these properties decrease with increase in temperature.

However, ductility increases with the temperature to some extent you can see here to some
extent. It is constant it is not influenced by the temperature which indicates a positive sign of
strength development.

152
(Refer Slide Time: 04:29)

So, let us see the other properties of steel. This curve shows the variation of thermal
conductivity which drops the thermal conductivity drops you can see here it drops with the
increase in temperature. It is also see that up to 800 degree Celsius it drops then at higher
temperature range it remains constant. So, temperature more than 800 degree centigrade
approximately does not influence the thermal conductivity of steel.

(Refer Slide Time: 05:44)

Similarly, if you look at the specific heat requirement, it gradually increases and the specific
heat increases rapidly in the close range of temperature varying from 650 to about 750 degree

153
Celsius after which it remains more or less constant. That is temperature range between 650
to 750 it rises very steeply and the range below and after they have no influence on specific
heat.

(Refer Slide Time: 06:47)

We can also see the curve which talks about thermal strain in steel which increases up to
around 750, then remains constant for a short plateau of 50 degrees, then again increase up to
1200 degree Celsius. So, these are all very interesting behavior what we see as far as steel a
construction material is concerned.

(Refer Slide Time: 08:05)

154
So, friends now we will talk about the new material which is recent and which is important
for construction purposes which we say as functionally graded materials. The functionally
graded materials or special kind of applications which are used or special kind of applications
which are used for type of functional requirements. One of the desired functional requirement
of steel in marine environment.

Let us say we will talk about marine environment because that is one of the vulnerable
environment for corrosion. So, one of the desired functional requirements of steel to be used
in marine environment should be corrosion resistant. Is it not? So, what actually people have
been using so far is a corrosion resistant alloy of duplex stainless steel which has been used
people also alternatively using carbon manganese steel.

So, now these have been used as individual materials for construction in marine environment.
Let us quickly see how we can try to combine the best properties of these two and form a new
material which is functionally graded material that is our objective now. Let us see that.

(Refer Slide Time: 10:38)

Before we understand to learn it further, let us ask a question what would be the fundamental
objective of a design requirement of members in marine environment? The fundamental
objective of design requirement could be that it should resist or structures in marine
environment could be. It should resist harsh or extreme weather conditions it should posses
both structural and functional advantages that is the idea. And most importantly as we
recollect it should not lead to a catastrophic failure.

155
(Refer Slide Time: 12:05)

So, therefore, friends importantly when we talk about application of such conditions in
marine environment. One classical example could be marine risers. Marine risers are actually
used for various purposes in offshore structures for drilling, for completion, for production
and export and for injection. These are various structural actions and functional requirements
of elements of marine risers in offshore environment.

Now, what is the specific problem these risers have? These risers are subjected to severe
corrosion. Why? Because due to the presence of chlorides hydrogen sulphide and carbon
dioxide. This presence is felt as well as is encountered in the marine environment. Therefore,
the marine risers are susceptible to this kind of environmental influence in addition to
combination of loads that arise from wave, current, ice, shock and impact.

So, here I have taken a classical example where the loads are also very severe, the
environment influence also very severe. Therefore, the design should focus on the functional
requirement which is anti corrosive I do not want the metal to corrode forget about the design
strength etcetera. Forget about giving corrosion allowance for the member.

Can you look into a material where the corrosion resistance is inhibited as a material
property. Instead of using an existing steel or existing material and making it anti corrosive
can I have a new material which is anti corrosive by itself that is innovation; that is advanced
steel design we are talking about.

156
So, a recent research has been conducted in IIT Madras as a joint research between
metallurgist, Professor and myself in Ocean Engineering Department. I would like to share
some important information about that particular context of recent research development
happened in IIT Madras for the benefit of the listeners of this course that is the advantage.

Friends in NPTEL courses where recent developments happening globally are presented for
the knowledge sake for learning purposes an academic contributions through NPTEL
platform. So, friends we are now looking forward for a new material which can have this
combination.

(Refer Slide Time: 15:52)

So, what is existing in the literature? What is now existing? Let us see that first before we
develop a new material. The existing ones are focused on composites we have composite
materials which are used as an alternate material for marine risers. But they have severe
problems. One of the foremost severe issue is, delamination under the combined action of
mechanical and thermal loads.

It also results in reduction of strength both in compressive and flexural perspectives and these
are undesirable for marine risers because we need them in a larger perspective. People also
investigated something related to the bonding which is done using metal and ceramic bonding
Aslan and Daricik in 2016. Further Kawasaki and Watanabe into in 1987 gave this idea of
improving the superiority of the material in terms of heat resistance.

157
Toughness because toughness is required to control the crack propagation or crack growth.
And we are looking for this under high temperature, high pressure, conditions because marine
risers are subjected to high temperature and high pressure conditions. So, under high pressure
high temperature conditions composites are likely to have delamination therefore, we cannot
use them. So, the alternate which is suggested in the literature could be the functionally
graded materials.

(Refer Slide Time: 18:48)

Let us see what are these functionally graded materials about. If you ask me the novelty of
this material this has been used in various industries earlier it is not new. People have used
this in defence, aerospace and in medical field. But you will; obviously, see the functional
requirements of this material for different applications are different. In medical field people
do not look for corrosion.

People look for durability in defence application, people look for strength in aerospace,
people look for strength versus low weight strength ratio that is largest strength and lowest
weight we are looking for the buoyant materials. So, the functional requirements of this
applications are different, but people have been using this in these industries much earlier.

In addition, they have been also used as cutting tools in cemented carbide applications. They
have been used as thermal barriers and they have been also used as anti-oxidation coatings.
And if you look at the manufacturing process of these they have been manufactured mostly

158
using thermoelectric materials by grading the carrier concentrations as clearly pointed out in
Koizumi 1997.

(Refer Slide Time: 21:41)

So, what we are now looking for is a material which has strength and durability. People have
been using FGM for various applications and one of the primary requirement what we are
now going to see is corrosion resistance as an important functional requirement. Of course,
strength is a basic requirement for the member which is to be needed this cannot be
compromised. In addition, we are looking for this is an additional requirement we have.

(Refer Slide Time: 22:55)

159
Now, interestingly the strength of FGM broadly depends on the manufacturing process.
Because, you should be able to establish a reliable interface between the constituent
materials. So, that you get a single material. The first application of FGM was tried in the
year 1987; it was used as a thermal barrier not as a construction material. This was done by
Kawasaki and Watanabe in the year 1987.

However, the cold metal transfer based WAAM Wire Arc Additive Manufacturing was
brought to commercial use only in the year 2011. So, this is called CMT; Cold Metal Transfer
this becomes the basis for wire arc additive manufacturing. This has been in commercial
application only about just 10 years back. So, therefore, it is a recent development and we are
investigating it for its suitability for special structures like marine risers.

(Refer Slide Time: 25:13)

What do we do in FGM manufacturing? In FGM manufacturing materials of desired

characteristics are chosen. What are the desired characteristics we need? We already said that
we need the constant material should have strength and should have corrosion resistance. We
are looking for essentially these two properties as functional requirements we choose this
material then we decide their geometric composition.

Please note we are not talking about the metallurgical composition. We are talking about here
the geometric composition friends please very careful about it. What do you understand by
geometry composition? It means we have to decide about the cross sectional shape, the
number of layers and of course, the thickness of each layer and each material.

160
So, in functionally graded material concept we pick up different materials of our choice
having desired characteristics. And then we are going to unify join them fabricate them
together using CMT based wire arc additive manufacturing technique. So, we decide during
the time what should be the cross sectional shape, thickness and the number of layers to
achieve the desired characteristics.

What are the desired characteristics we fixed earlier? What is the requirement of strength we
need? What should be the corrosion resistance we want? So, we have to examine them right.
So, we have to prefix them and examine them.

(Refer Slide Time: 28:10)

So, now if you look at a conventional FGM manufacturing process. They are manufactured
by two ways; one is a continuous variation, the other one could be a stepwise variation. Now,
what is the continuous variation? I look at a material this is my material boundary. So, the
materials are varied continuously and there are no distinct boundaries. So, it varies
continuously. So, the material varies continuously, no distinct layers.

The second could be a stepwise variation where what we do here is we keep on putting
material layer by layer. So, we are doing layer by layer, but please note friends these layers,
these boundaries will be merged. There will be no distinct layers visible after fabrication that
is very important. It will look like a single material where the manufacturing is done in
layers. So, that is the method of manufacturing FGM. One is called stepwise grading; other is
called continuous grading.

161
(Refer Slide Time: 30:38)

So, now FGM utilizes the following. What is the advantage if you manufacture FGM? It
utilizes the following; it has the mechanical, the metallurgical and structural, advantages of
the constituent material. Let us see what are those possible constant materials which can suit
our functional requirement? What is our functional requirement? Our functional requirements
are strength and corrosion resistance that is what we are looking at this moment as an
example.

For this the material available are duplex stainless steel, titanium and carbon manganese steel
where they show high resistance to corrosion an acidic environment. They also have high
resistance to stress corrosion cracking. They also sustain chlorides in marine environment;
those are present in marine environments and so on.

162
(Refer Slide Time: 33:16)

Now, let us ask a question what are the conventional manufacturing methods available in the
literature for manufacturing FGM? There are many methods available in the literature
centrifugal casting, spark plasma sintering, laser deposition and powder metallurgy. These are
all of course, referred in literature very widely I am not discussing each one of them in detail.

For example, you can look at Yuan et al in 2012; we also look at Chen et al in 2017. These
are for some cross references, but there are many things available please look into the list of
references in the website of this course you will get this links.

(Refer Slide Time: 34:45)

163
However, additive manufacturing technique is recent. There are many additive manufacturing
methods, there are many additive manufacturing methods, selective laser melting, direct laser
melting, electron beam melting and wire arc additive manufacturing. However, WAAM is
very comfortable to use in lab scale and it has got one of the advanced techniques for
manufacturing metal components for engineering applications. So, it is suitable to
manufacture metal components for engineering applications.

(Refer Slide Time: 36:38)

Let us now talk about WAAM. Let us talk about WAAM a bit, WAAM is a compatible
manufacturing process. It enables use of composition of metals which can be user defined
that is very interesting. You can define the composition to manufacture that combined
product in WAAM process. So, WAAM practices stepwise manufacturing method. We have
seen there are two methods of FGM continuous in stepwise WAAM adopts stepwise
manufacturing, and the product development is 3 dimensional.

So, how does it do this manufacturing? The manufacturing process is digitally controlled
wherein the chosen metals or I should say materials or functionally graded as per the user in
the form of wires that is why it is called wire arc additive manufacturing. And these wires are
interchanged these wires are interchanged during the deposition process.

164
(Refer Slide Time: 39:07)

This method actually overcomes the limitations of powder bed process of manufacturing. The
power bed process has got certain limitations; the limitations in powder bed manufacturing
are deposition rate, equipment cost and cost of raw material. So, these factors are overcome
in WAAM. So, in WAAM the powder material is replaced in the form of wires ok that is one
physical difference you can see here. So, the heat source used for manufacturing for
manufacturing is an electric arc that is the heat source we use for manufacturing.

(Refer Slide Time: 41:21)

165
WAAM is considered to be economical because it can manufacture large size, it can
manufacture any geometric shape, it can manufacture faster so shorter time with minimum
wastage. So, it is economical these are of course, stated in many studies one classical
example could be Williams et al 2016 and Martina et al 2011.

(Refer Slide Time: 42:25)

So, WAAM can be also expedited you can make this manufacturing faster if the wire feed is
replaced with twin wire feed system. So, this can control and improve the deposition rate and
the composition of materials. As I told you electric arc is used as a heat source it supplies
high pulse current in the form of an arc between the electrode wires and the substrate and
makes the deposition.

166
(Refer Slide Time: 44:14)

So, let us talk about the application of FGM to marine risers. So, what are the objectives in
this case? Objective means what are the functional requirements for the marine riser? The
requirements are it should possess the basic strength; it should have corrosion resistance. It
should sustain marine environment high temperature, high pressure conditions, accidental
loads thermal loads and it should be cost effective.

(Refer Slide Time: 45:28)

167
(Refer Slide Time: 45:36)

We will see this in detail in the coming lecture friends. So, as a summary in this lecture let see
what we have learnt. As a summary we learnt, what is a functionally graded material, what do
you understand by functionally grading, what are various manufacturing methods of FGM. In
that we also saw what is wire arc additive manufacturing, how this method is advantages in
comparison to other methods used to manufacture FGM correct.

So, in the next lecture we are going to talk about evolution of a new material for marine riser
which is an FGM. We will see what are the constituents of this material, how are they
composed, how it is manufactured and then once we attain this material as a single material
without any stepwise layers.

Let us see the assessment of mechanical properties of this FGM and compare those properties
with the constituent materials and see is any improvement on FGM or not. So, we will do that
study and learn them in the next lecture.

Thank you very much, have a good day bye.

168
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 10
FGM for marine application - 1

(Refer Slide Time: 00:21)

Welcome to the 10th lecture in which we are going to learn more about the Functionally
Graded Materials let us quickly highlight what we discussed in the set of last lectures. That
we are looking for advanced structural systems, which are designed to be form dominant. We
are also looking for materials which support the system for its functional and structural
requirements.

Structural requirements refer to the mechanical properties such as yield strength, Young’s
modulus Poisson’s ratio of a t strength, fracture resistance, which are generally and possibly
very good as far as different steel grades are concerned, there are no issues about it. When
you come to the functional aspects of steel especially under the influence of environment to
be very particular in marine environment steel undergoes corrosion.

And that is one of the disastrous requirement which is not agreed upon by any engineer in
practice for marine systems. So, looking forward for a material, which is having similar
strength to that of steel, but possesses better corrosion resistance so, were looking for a
functional requirement of the material. So, advanced steel design is dealing with two aspects

169
of the design procedures - one is a geometric centric design, where the structural assembly
takes care of the load acting upon the structure apart from the advanced design procedures
such as plastic design etcetera.

In addition, we are also looking for a material which is compatible with this large
displacement structural systems. What does it mean? The material should possess enough
ductility; to adopt plastic design the elastoplastic domain of the material should be very
dominant. To have good toughness the area of the stress strain curve should be large enough,
the material should exhibit ductility etcetera.

Apart from possessing these advantages properties we will also choose a specific requirement
let us say corrosion resistance which is indeedly required in marine environment. So, we did
a recent research at IIT Madras with assistance with Professor Murugayan in Department of
Metallurgy and Material Sciences, we came out with an fabrication of a new novel material
which is functionally graded of different constituents to have corrosion resistance and
strength at par with steel.

So, in this lecture we are going to talk about that particular recent innovation on FGM which
is yet to be applied to marine systems, but still it is a research stage and I believe you love to
know more information about this kind of recent advancements in structural design as far as
steel is concerned. So, in FGM again steel is an important ingredient, but the corrosion
resistance is improved by combining or alloying this with another metal.

That is what we have tried to do it. Let us do this and see how do we proceed with this
further.

170
(Refer Slide Time: 04:51)

As we said in the last lecture FGM is actually grading 2 or more metal components or metal
to fulfill a particular functional requirement. So, friends at this moment we are taking the
functional requirement as corrosion resistance. There can be any such functional
requirements, let us say though low thermal conductivity, high temperature performance,
excellent fatigue characteristics.

You can dictate, you can choose any specific functional requirement and modify the
characteristics of steel by altering the ingredient metals to form a functionally graded
material. So, as I said there are many manufacturing processes which are available in the
literature to fabricate a functionally graded material out of which, wire arc additive
manufacturing is one of the prominent method which is compatible manufacturing procedure
which enables the metallurgical components of the materials to be user defined.

So, it enables, it permits user defined metallurgical composition of materials to be graded. So,
is user friendly, so it does it. So, in WAAM, stepwise addition is done; we have already said
continuous and step wise are 2 different methods of FGM manufacturing in WAAM step wise
addition is generally carried out.

171
(Refer Slide Time: 07:29)

The manufacturing process is controlled in a digital mode. So, the manufacturing process is
digitally controlled whereas, materials are chosen as per the functional requirements.

Friends usually 2 or 3 materials are graded, not more than that. You can have more, but the
complexity of manufacturing will be very large. So, usually 2 or 3 materials are graded to get
a FGM product. So, in WAAM what we do, this chosen materials are interchanged or let us
say deposited in the form of wires. So, WAAM has got a great advantage over the powder
bed process in terms of deposition rate, equipment cost, cost of the raw material etcetera.

So, this has got an edge over this particular process of manufacturing and WAAM is
considered to be advantages compared to this.

172
(Refer Slide Time: 09:46)

In the WAAM process, the raw material or I should say the ingredient metals are added in the
form of wires. So, I should say the metallic powder is replaced as wires in metallic powder
bed process, the metal powder were are actually added in this case the raw material is added
in the form of wires and for manufacturing electric arc is used as a heat source.

Therefore, friends WAAM is considered to be economically viable and it is capable of

manufacturing large size members, it is also capable of manufacturing any shape including
circular, because circular is considered to be one of the most difficult fabrication process.
Furthermore, it has got a minimum wastage of materials. So, I should say this process is
highly economical compared to the conventional processes used in manufacturing alloys or
composites.

173
(Refer Slide Time: 12:17)

High pulse current rate is supplied to form an electric arc and that is remaining as a heat
source for the problem. Now, let us talk about a specific application of FGM for marine riser.
So, let us say we take a member for our consideration, the member of our choice is the
marine riser. Now, the question is for the knowledge of other people listening to this course
who are not offshore engineers.

Let us quickly explain what is a riser. Let us see if this is my offshore platform is having
multi tire deck, which has got the crane facility, which has got the top soil details in terms of
drilling platform, drilling derrick etcetera and a helipad. Or let us say the platform is
constructed and resting on the sea bed by a template arrangement structural system and this
becomes my water level let us say.

From the drilling, head a pipeline is driven to the sea bed which is now going to explode the
oil and it is going to be carried up through this pipeline and then it is processed or stored on
the top side. Since the oil is traveling against gravity this component is called as a riser and
since it is used for marine application so it is a marine riser. So, marine riser is actually a
circular pipe or a cylindrical tube, which is essentially used to explore, suck oil or gas from
the seabed to the top side.

So, our application member is a marine riser, you can see here the marine riser is completely
embedded in sea water which is corrosive in nature and in addition to that the oil and gas
which is passing through this riser is of a very high pressure and high temperature. So now,

174
we are looking for a member which is located in the corrosive environment, the member
should sustain high pressure, the member should also sustain high temperature. So, looking
for a material which caters to these functional requirements.

If you ask me a question what is the general material being used currently in the industry?
Because oil and gas exploration is not today’s job.

(Refer Slide Time: 16:15)

The first oil well was drilled in United States in the year 1912, 100 years back. So, this
process is there for the past 100 years. So, people must have been using say material for
marine risers, we state that they are corrosive in nature, but people have been using them.

So, what is that material being conventionally used so far is X-52 grade steel, which is the
marine steel. I will show you the mechanical properties of this steel slightly in a minute later
in the next slide. It is a conventional steel being used, now we want to replace this steel with
a new material which is functionally graded. So, to improve the performance of corrosion
resistance, we are looking for alloying the duplex stainless steel with carbon-manganese steel.

So, friends these ingredients, these metallurgical constituents are chosen based on specific
functional requirements. The functional requirement what I have is of course, the strength,
there is no compromise on strength we are also looking for additionally a corrosion resistance
right. So, which is 2, we can choose these 2 material. If you ask me a question, do only these

175
two-material available which are corrosion resistant, there are many titanium, there are many
material available.

For example, we have demonstrating an FGM fabricated in our lab with these two material,
you may ask me why duplex stainless steel that is a common material being currently used as
a single material for risers. So, we want to add that to carbon manganese steel and see
whether corrosion distance can be improved, that is the question here that is the argument.

(Refer Slide Time: 18:56)

Having said this, if you look at the properties of duplex stainless steel, it has very high
corrosion resistance towards hydrogen sulfide gas, carbon dioxide gas and acidic
environment. It is also capable of resisting cracks arising from stress concentration. Now, we
want to add this material with carbon manganese steel which X-52. So, X-52 is a carbon
manganese steel which is currently used as marine steel in rises. So, we want to add this to
duplex stainless steel and see what happens.

176
(Refer Slide Time: 20:28)

Let us quickly see the stress strain curve of materials what we are discussing now.

(Refer Slide Time: 20:41)

Before we go that, let us see the material what we are choosing and what is the code
compliance of this. Because we already said the choice of the material also depends on
whether the code accepts it or not. So, duplex stainless steel is anyway recommended as 22
Cr sorry, ER 2209, which is recommended by DNVGL-RP-F112.

177
Carbon manganese steel is also recommended for marine application, which is ER 70S-6
which is recommended by DNV-OS-F101. So, friends one can easily get convinced that these
two ingredients steel are not new recommended for marine environment, already existing
codes permit, use of this material for marine environment. So, code compliance is satisfied.

So, one can really go for alloying these materials and see what happens. Now, let us look into
the stress strain curve of carbon manganese steel. The one what you see on the screen is the
stress strain curve of carbon manganese steel. So, the carbon manganese steel does not have a
pronounced yield point. So, to obtain this, one can easily find out the 0.2% proof stress value
and get the yield strength, one can also see it has got enough ductility.

Carbon manganese steel possess enough ductility right, because the ductility ratio is
significantly larger for this steel. So, this is the stress strain curve for carbon manganese steel,
which is going to be the ingredient for my FGM. So, what is that element we are looking at
how we are going to composite, let us see that here.

(Refer Slide Time: 23:23)

Now, I am looking for a marine riser which is circular in cross section, which should be now
functionally graded correct, with two materials, one is duplex stainless steel another is carbon
manganese steel. We are going to do this.

178
(Refer Slide Time: 23:56)

Let us see how we are going to show the cross section? A typical cross section of the
proposed riser is now shown on the screen. So, the cost of duplex stainless steel is
phenomenally high compared to carbon manganese steel. Friends it is because of this reason
duplex stainless steel as such as a single material is not being used as a riser. The cost will be
phenomenally high.

Carbon manganese steel is cheaper compared to duplex stainless steel. Therefore, what we
propose is only a small portion will be duplex stainless steel and the majority will be carbon
manganese steel. So, economically the proposed new material will not be very expensive in
comparison to the conventional carbon manganese steel, because you are adding only one
layer of duplex stainless steel.

But remember it is not a coating, it is not a composite, it is not a layering material. It is

fabricated as a single material that is the beauty here. So, the functionally graded riser now
has got 2 material where the inner portion will be susceptible to corrosion because of HPHT
conditions, and the carbon manganese steel is located outside as you see in the screen here.

Having said this, that we now have the cross section, we now have the material let us see
what will be the thickness we are proposing. Let us say what is the material and what is the
thickness we are proposing. So, the material we are saying is duplex stainless steel, which is
22 Cr 2209 which is only for 3 millimeters. Whereas, the carbon manganese steel which is
qualifying Er 70 S-6 gray is about 14.47 millimeters.

179
So, the total thickness of the riser is 17.47 millimeters out of which 3 millimeter is duplex
stainless steel and remaining is carbon manganese. So, now, one will be curious to know how
do we get this thickness of the riser, that is a very interesting question friends. We have taken
a live example of a riser present in Auger TLP, tension like platform in the Gulf of Mexico
and we have adopted the same cross-sectional dimension of the riser for the analysis.

The marine raiser present in the Auger TLP is of thickness 17.47, please note the existing
riser present in Auger TLP does not contain duplex stainless steel, it is a single layer material
it is not FGM. So, we are now improving the corrosion resistance of that material by adding
stainless steel of a marginal thickness, without compromising the string let us see what
happens to this combination right. So, we are looking for a thickness of this order.

(Refer Slide Time: 27:42)

Now, we have the cross-sectional dimensions ready, we have the material composition ready,
we have the thickness to be done ready. We also have the manufacturing process which is
also finalized I am going to use wire or additive manufacturing, where I am going to feed the
carbon manganese steel and duplex stainless steel in the form of wires. I am going to
manufacture this in layers.

180
(Refer Slide Time: 28:35)

Having said this, let us see what the code advises for the desired wall thickness of risers. The
desired wall thickness of marine risers should take care of the corrosion allowance. So, the
minimum thickness is the internal pressure, the nominal diameter divided by twice of eta h σ𝑦

plus thickness meant for corrosion, equation 1.

Whereas 𝑃𝑖 is internal pressure, D nominal is a nominal diameter of the member of the riser

eta h is a usage factor which is a factor used in the design usually this value is taken as 0.6
and the corrosion allowance given is about 6.265 millimeters. As per international practice
the minimum corrosion allowance given is 3 millimeters; if you provide carbon manganese
steel.

So, this is not t steel this is t minimum. So, you provide t minimum for corrosion as 3
millimeter, then work out for the given pressure and diameter and this work out t minimum
that comes to be 6.265 millimeters. So, friends please note along the inner surface of the riser,
there is a corrosion resistant layer of 3 millimeter, which is going to handle the scouring
effects of the carbon dioxide.

You see here along the inner side, along the inner side there is going to be 3 millimeter of
corrosion distance material, which is the duplex stainless steel, which is now going to handle
the corrosion problems right. So, that is what we are planning now.

181
(Refer Slide Time: 31:46)

So, furthermore, a stainless-steel substrate is used to deposit the material during WAAM.
Now, we are looking at the manufacturing process.

The power source used for deposition is Fronius, Cold Metal Transfer which is CMT trans
pulse synergic 4000, a welding torch is used for depositing materials in layers. The whole
process is now controlled using tri-axial computer numeric control machine which we call as
CNC which I am going to show the photograph here.

(Refer Slide Time: 33:37)

182
So, you have the CMT torch, you have the substrate and you have the 3 axis triaxial CNC
machine which is now going to be used to manufacture or fabricate this FGM. Of course, the
whole process is also measured and monitored using an high speed camera and also an
oscilloscope for varying the pressure and also measuring the response of the wire feed or the
rate of wire feed with respect to the manufacturing process. So, G codes are used, G codes are
programmed to drive the metal deposit along x-y-z axes.

(Refer Slide Time: 34:57)

So, a typical G code which is used for a bidirectional strategy is shown here. So, this is the
bidirectional weaving pattern for successive depositions and this figure shows the scan
weaving path of metal deposit. So, one can say this is scan pattern what you see here is being
mapped here and that is the substrate, which is actually stainless steel, below that there is a
base and CMT torch and electrode wire are being fed.

And the wire is fed in the same pattern, which is actually obtained depending upon what is
the wire feed rate and how the pattern will be established to get the normal thickness of the
desired value.

183
(Refer Slide Time: 36:25)

Having shown this, the functionally graded material now is manufactured or I should say is
build in 2 stages.

So, essentially, we must realize that this FGM contains 2 materials; one is carbon manganese
steel other is duplex stainless steel. The thickness, shape, everything is fixed. So, they are fed
in wires in the form of wire and we are going to deposit in 2 stages. So, to bind these two we
also require a filler material. The filler material is actually used in this case ER 70 S-6 which
is compatible with carbon manganese steel and ER 2209 grade, which is a duplex stainless
steel. So, now we can show the chemical composition of the filler material.

184
(Refer Slide Time: 38:03)

The chemical composition of the filler material is shown percentage by weight. It is got
carbon, so ER 70 S-6 and ER 2209. In fact, there are two filler materials I am giving the
value of both the filler material being used in our research. Suppose 0.06, here it is 0.008.

We have also used manganese, the component sorry the content which has got 1.45 and this
contains 1.62; silicon, 0.9 and 0.49; phosphor, 0.02 and 0.016; Sulphur 0.03 and less than
about 0.005, I mean 0005, very very negligible it contains nickel of 0.12, 8.59 very high
content.

It contains chromium of 0.12 and this contains 22.93, it has copper of 0.4 and this contains
0.083, it also has nitrogen which is very insignificant whereas, in this case it is 0.152. It also
has vanadium, cobalt and Ferrum. Vanadium is 0.03, cobalt is negligible and ferrum is the
balance. Whereas, in this case vanadium is negligible, cobalt is 0.041 remaining is the
balance.

So, we have got the chemical composition of two different filler material, which has been
tried in this current research to fabricate the functionally graded material, constituting two
ingredients, duplex stainless steel and carbon manganese steel.

185
(Refer Slide Time: 40:41)

The arc energy being used for manufacturing or otherwise the heat input required per unit
length of the wire in the deposition process is given by this equation which is IP, the heat
input which is 1 by n of summation of i equals 1 to n.

𝑢𝑖 and 𝐼𝑖 where 𝑢𝑖 is the instantaneous voltage and 𝐼𝑖 is instantaneous current. Arc energy is

therefore, given by the efficiency multiplied by the input, heat input divided by nu usually the
efficiency of vamp wire deposition process is taken as 80% and the variable parameter in this
case is going to be the nu which is called as a deposition speed.

The deposition speed depends upon the wire feed rate and also depends on the ingredients of
the metal.

186
(Refer Slide Time: 42:29)

So, the deposition parameters is an optimized and these are the following data being used in
the current study.

(Refer Slide Time: 42:37)

So, the deposition parameters for FGM composition, these are the welding parameters for
duplex stainless steel and for carbon manganese steel. The welding process being used is
CMT in both cases.

187
The filler material being used is ER 2209 and, in this case, it is ER 70 S-6, the filler diameter
is used for fabrication is 1.2 millimeter in both the cases. The mean current being used for
depositing is 100 and time 19 amperes whereas, in this case 107 amperes. And the mean
voltage developed during the deposition process is 12.86 volts whereas, in this case 16 volts.

The shielding gas being used is the pure organ gas in both the cases. The gas flow rate being
used is 16 liters per minute in both the cases. The arc energy developed is in this case 222.59
kilo joules per meter whereas, in this case is 256.8 kilojoules per meter. The welding speed of
the torch is 330 millimeters per minute for fabrication; whereas, in this case it is 320
millimeters per minute. So, these are the deposition parameters which is being used in
fabricating this FGM.

(Refer Slide Time: 45:04)

So, friends we have learnt in this lecture more details about the FGM process, the ingredients
of the experimental FGM we did, the properties of the constituent metals and weld
parameters. So, in the next lecture we will discuss more in detail about the welding process,
fabrication of FGM and the mechanical and structural properties of the fabricated FGM and
compare it with the parent metals and see how they have improved when FGM is formulated.

Thank you very much have a good day.

188
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 11
FGM for marine application - 2

So, friends welcome to the 11th lecture where we are going to discuss more in detail about
the FGM.

(Refer Slide Time: 00:23)

So, in the last lecture we discussed about the ingredients of FGM which has got a specific
functional requirement which is corrosion resistance. We have checked the code
comprehensive of the ingredients of the material. We have assessed the thickness requirement
of the material. The cross sectional diameter and then the ingredient materials depending
upon the strength and functional requirement. We have chosen duplex stainless steel and
carbon manganese steel with respective fillers.

Then we arrived at the weld parameters we do not call as weld parameter actually they are
fabrication parameters, why it is called as weld because we are using a torch to control the
rate of flow of the wire. So, we call that as a weld actually So, once you deposit this by
layers this is what you get.

189
(Refer Slide Time: 01:41)

So, this is a view a and view b of the build. So, one can see, that the deposition is by layer.
And you can see both the material carbon manganese steel and duplex stainless steel
deposited like this, to form an FGM. The mechanical characteristics of this is assessed using
a high speed video camera and we are able to find out the wire electro discharge machining of
this.

(Refer Slide Time: 02:48)

So, now when this is manufactured in a circular form this is what we get. So, we are able to
get a cross section where it is circular in shape. The inner material what you see is a duplex

190
stainless steel. Now, outer material what you see is the carbon manganese steel, which is
shown in different two colours you can say this is only a very small thickness of 3
millimetres and this is a large thickness of 14.67 millimetres. Friends very interestingly we
cut a portion from here to test it in UTM for mechanical characteristics, to assess the
mechanical characteristics.

So, what we do is we have taken a stratum from the FGM build which you see an enlarged
view here, which is cut from the build as you see here and placed in an UTM. Now, before
we put it in UTM we try to also place the markers for measuring strain. So, you can see there
are markers I am marking it in a different colour. So, this and this are for longitudinal strain
and the other one are for lateral strain .

So, these are placed in a UTM. So, please note here this is the direction of pull. So, very
interestingly friends, we are applying a tensile pull parallel to the layer right. So, this
becomes my longitudinal direction and this becomes my lateral direction . And we assess
now the stress strain characteristics of this material.

(Refer Slide Time: 05:19)

So, Poisson’s ratio is estimated based on this and we are using the Zwick Roell z 100 UTM.
So, the specimen is placed in the UTM I am just indicating the carbon manganese steel and
stainless steel in different colour. So, this is duplex stainless steel and this is carbon
manganese steel and we are going to apply the pull parallel to this . That is very important.

191
This is my longitudinal direction and this becomes my lateral direction. So, I am going to
measure the Poisson’s ratio based upon these two. Let us see how do we do it. So, the tensile
strain and stresses are observed when you place an UTM. So, we use a video extensometer to
measure this three specimens are tested and then we used to do the metallographic
investigations on the finished surface.

(Refer Slide Time: 06:41)

So, now to determine the elemental composition of the FGM build, scan electron microscope
which is quanta 200 F to check the elemental composition essentially at the interface of two
metals. Let us say a and b a is duplex stainless steel and this is carbon manganese steel at the
interface we want to see. So, the observations were recorded for about 35 micron length.

The micro hardness test is also carried out on the polished surface of the FGM build . And
hardness test is conducted using Wolpert Wilson Micro Vickers digital hardness test. Further
the samples obtained after tensile testing are subjected to x ray diffraction using Bruker D8
X-ray diffraction equipment.

Let us now see first the metal transfer characteristics which are obtained. The metal transfer
characteristics very clearly showed that there is a uniform distribution or uniform spread of
metal a with metal b. There is no distinct layer seen between these two. So, there is a good
mixture good composition mixture of these two which was verified from the metal transfer
characteristics of this.

192
(Refer Slide Time: 09:26)

(Refer Slide Time: 10:04)

Further, X-ray computer tomography was conducted, to check the porosity and formation of
micro cracks during the manufacturing process. So, the figure now shows the view what you
obtain using X-ray computer tomography, a and b shows a shows the 3 dimensional view, b
shows the top side view and of course, this is c and this is d c refers to the side view and d
refers to the front view of the FGM build .

So, from the figures and the x the photographs one can say that both duplex stainless steel
and carbon manganese steel are free from manufacturing defects. There are no micro cracks.

193
So, it confirms that the wire arc additive manufacturing method used is good for building
FGM and it can be used for engineering applications .

(Refer Slide Time: 11:52)

Furthermore, tension test was conducted under the axial tension the strain rate used to do this
was 1 x 10^(-3) per second. The stress strain curve of the FGM sample along both directions
longitudinal and transverse can be seen here.

(Refer Slide Time: 12:39)

194
So, this is longitudinal and this is transverse of sample 2. There were many samples tested
one of the sample. So, one can very well see here that the ultimate strength is approximately
602 megapascal. The yield strength was about 403 megapascal. The modulus of elasticity the
slope of this line is about 213 gigapascal and the material show very good ductility. In fact,
we want to compare this with the X52 steel and see what happens.

(Refer Slide Time: 13:58)

Let us show a comparison of this with all the 3. This is my X52 steel friends which is the
conventional steel being used for marine risers now. And the remaining three colour curve
show you the stress strain curve longitudinal of course, longitudinal for three samples of
FGM build. One can easily see here that FGM has got a similar ductility as that of X52 while
the ultimate strength the yield strength and the toughness is much higher, compared to the
conventional X52 steel which is currently being used for marine risers.

In addition to this the FGM build what we proposed had a exclusive corrosion resistance is it
not. So, it is functionally graded without compromising the strength of the parent material
that is the beauty of the whole research problem . Even at fracture which is pulled along you
will see there is no separation is it not. It is fracturing like a normal ductile material. There is
no layer separation which generally is expected when it would have been a composite.

So, FGM is not a composite . Though it is manufactured in layers, though the deposition is
done by wires, but because of the manufacturing process and technique the chosen two
materials got unified and you get a single material as an outcome . That is the whole

195
emphasis we want to make here. Having said this if you look at the fractured sample the
fractured sample does not show any layer interference between the stainless steel and carbon
manganese steel.

It also showed the percentage elongation higher by about 6 percent in compared to X52. So,
interestingly friends one will be very curious to know the test results of X52 and FGM.

(Refer Slide Time: 17:09)

So, we will compare this with X52, which is marine riser steel with our FGM under: Young’s
modulus, yield strength, ultimate strength. Let us also try to work out the strength ratio. We
will also work out the ductility ratio from the curves. Then we will also compute the
toughness tensile toughness then Poisson’s ratio.

The Young’s modulus in this case is established as 210 MPa whereas, we got 209.66 plus or
minus 4.48 MPa, as far as yield strength is concerned X52 exhibits 358 whereas, we got
39.66 plus or minus 12.23. That is the variation we have in samples. Ultimate strength
showed 453 MPa in X52, whereas, we had a very exorbitant value of 587.66 plus or minus
12.76 MPa.

So, therefore, the strength ratio in this case 1.265, which is actually the ratio of ultimate
strength yield strength whereas, we have 1.5 plus or minus 0.02. The ductility ratio usually is
32.207 for 52, but we have 45.47 in FGM. The tensile toughness which is expressed in joule
per cubic meters is 104.92 for X52. Whereas, in our case we got 120.5 plus or minus 2.84.

196
So, the marine riser values have a very strong reference which is taken from Lyapin et al
2015 and FGM is our own paper.

So, one can see here that the Young’s modulus being in the same range all other parameters
are excellent compared to the current steel being used as marine riser.

(Refer Slide Time: 20:28)

So, let us go one step ahead and check. What is happening to Poisson’s ratio. And percentage
elongation, because these are very required. So, the Poisson’s ratio classically for X52 is 0.3
whereas, we have 0.3 plus or minus 0.07 variation. Percentage elongation is 20 whereas, we
get 22.31 plus or minus 0.11. So, all parameters are excellent and in fact, more than the
requirement as recommended by the code compliance for marine risers.

So, friends advanced steel design also comes from material renovations. So, this is a very
interesting example research study which I am presenting to you which are of course,
available in our published papers and references given in the course material for this
particular course. You can have more information about this.

We have also authored a book on FGM for marine risers please look into the book it contains
more details about the manufacturing process, the constituents of FGM and other mechanical
and structural characteristics which are more promising compared to the marine steel. Having
said this let us compare the interface strength.

197
(Refer Slide Time: 22:06)

Now, let us compare the interface strength. So, we have used ASTM E8 sub size specimen.
So, do this test and compare. So, we are now comparing the material for yield strength which
is 0.2 percent proof strength in megapascal. We will also compare ultimate strength in
megapascal. We will also compare percentage elongation quickly for this material.

For X52 this is 358 is 455.05 and this is 21 whereas, for FGM we have 340 and 483 plus or
minus 2.5 and 16.02 plus or minus 0.4. So, they are in the acceptable range of variation with
respect to the parent material being used. So, therefore, we further investigated more
structural properties of FGM and compare it with X52.

198
(Refer Slide Time: 23:57)

So, we also investigated the toughness properties. So, toughness is actually the area what you
work out with the stress strain curve. The area has been computed and the toughness value
what you get from the stress strain curve of FGM is what you see here which is comparable.
So, the toughness value what we obtain for FGM sample is 123.611 joules per cubic meter
whereas, for X52, we also worked out this and that came to 104.92 joules per cubic meter. So,
FGM proved to be equally tougher compared to X52 steel .

(Refer Slide Time: 25:01)

199
We further did microstructure analysis on the FGM mould. The FGM mould is grinded to
high surface finish. The test section what we now consider consists of carbon manganese
steel, duplex stainless steel and interface . So, the microstructure analysis is carried out in
different sections.

(Refer Slide Time: 25:53)

So, there are different sections about 7 sections are cut. Carbon manganese and duplex
stainless steel are very clearly shown here. So, this is at Section 1. So, the microstructure
Section 1 shows an equiaxed grain formation . This comprises of pearlite matrix of ferrite.
Section 2 shows an equiaxed formation of pearlite matrix of ferrite, it also contains
martensite.

200
(Refer Slide Time: 27:25)

(Refer Slide Time: 27:47)

This is Section 3 which is here. Section 3 shows high concentration of martensite. then we
move on to the next section. Section 4 show a clear distinction between the pearlite formation
in a matrix of ferrite. There is a clear distinction. Section 5 shows, the face of carbon
manganese steel now. Section 5 is here and carbon manganese steel is this.

See the legend. It shows a perfect blending and mixture of both the metals is it not. This is
Section 6 is the location of interface, because you can see Section 6 is located exactly at the

201
interface between the stainless steel and carbon manganese steel. So, one can very clearly see
the interface in the image.

(Refer Slide Time: 29:03)

And Section 7 the microstructure shows the austenitic, which is shown in bright colour and
ferritic, which is shown in dark grains of the composition. So, friends the micro structure
analysis very clearly shows there is a uniform blending of both the material as we fabricated
it.

(Refer Slide Time: 29:48)

202
(Refer Slide Time: 29:58)

Furthermore is a comparison of all the images as we just now saw. Then furthermore we also
did energy dispersive X-ray analysis, what we call as EDX. So, this analysis done at the
interface of carbon manganese steel and duplex stainless steel at the interface. an elemental
EDX scan analysis is carried out. The purpose is to determine the chromium variation across
the interface. That is the idea.

So, the figure is shown on the screen now. So, at the interface this is the figure it shows the
chromium content as seen in the figure.

(Refer Slide Time: 31:10)

203
The chromium content in duplex stainless steel is about 20 to 22 percent at 32 to 35 microns
from the interface whereas, the chromium content is closer to the deposition process. So, it is
important to note that the chromium content which is responsible for corrosion resistance has
not moved away from the interface. It is not dispersed that property is protected.

So, we also wanted to check the chromium molybdenum and nickel content of this.

(Refer Slide Time: 32:41)

So, this is a chromium molybdenum a nickel content of the FGM build, which ensures that as
you move away the distance from the microns you will see there is a uniform spread of all
these metals which are all necessary for providing corrosion resistance. So, though the FGM
build is manufactured in wires, but still the parental characteristic of the duplex stainless steel
which offers corrosion resistance is still maintain its enrichment in the formulated FGM build
. That is what we observe from this image analysis.

204
(Refer Slide Time: 33:48)

A micro hardness test was conducted using Vickers hardness, the duplex stainless steel
showed Vickers hardness in the range 266 to 235. The carbon manganese steel showed
between 133 to 197 . At the interface in the vicinity of about 2 millimetre thick the hardness
number was seen as 307 to 320. So, the martensite composition in the vicinity of the interface
has increased the hardness at the interface.

So, it has never failed in interface, because of this reason friends you can see in tensile test
the interface separation did not occur. The increase in hardness contributes to the overall
increase in strength of the material. It is also seen in the region of duplex stainless steel is of a
higher oddness compared to carbon manganese steel which provides an additional strength to
the FGM material.

205
(Refer Slide Time: 35:45)

So, an image was also fabricated and tested, which shows at the interface the hardness is
relatively higher as see in the figure. Further more friends a fractography analysis was also
carried out.

(Refer Slide Time: 36:12)

On the FGM mould by extracting a section comprising both carbon manganese steel and
duplex stainless steel. Fractography is conducted at both this location of tensile sample. So,
the fractography image is now what I am going to show you.

206
(Refer Slide Time: 36:40)

This is duplex stainless steel, this is carbon manganese steel and this is the SEM image at the
failed portion or at the fractured portion and this shows the fractography indicating a ductile
failure. So, friends it is seen from the image is that the dimples are prominent on a large
diameter in carbon manganese steel compared to duplex stainless steel. A full smooth
cleavages present in duplex stainless steel is also noticed.

So, this shows an improved ductility. Furthermore, the stress partitioning was shown through
an SEM image between the two materials, which shows that the failure is purely a ductile
failure. So, the ductility characteristic of FGM is not compromised by mixing or by
functionally grading these two materials.

207
(Refer Slide Time: 38:48)

Furthermore, a pitting corrosion test was also conducted on the build using ASTM-G48-11
method A. The test solution is used is about 100 grams of reagent grade ferric chloride in 900
ml of distilled water. All the surfaces of the mould were ground clearly using 120 grade
emery paper and acceleration corrosion test was conducted for 24 hours.

The corrosion rate is determined by change in weight and surface area. The corrosion rate
obtained for duplex stainless steel is 70.77 grams per centimetre cube and that for carbon
manganese steel which is 823.35 grams per centimetre square.

(Refer Slide Time: 40:16)

208
X-ray diffraction analysis carried out, where you see the result on the screen now which
shows that there is a central peak of chromium 0.1, in ferric 1.9 indicating the presence of
chromium oxide at the interface. So, it indicates the presence of FGM. So, the formation of
chromium oxide gives duplex stainless steel its corrosion resistance which is now available at
the interface . So, that is a new material we have which has got a comparative value.

(Refer Slide Time: 41:23)

So, the more details of the studies can be seen in these two textbooks. One is the Design of
Marine Risers using FGM, which is written by me for Elsevier. There is another book which
talks about use of FGM in Offshore Topside which is written by me for Wiley. So, these two
books will serve as a very good reference material for you to know more about the material
properties of FGM.

209
(Refer Slide Time: 41:57)

So, friends as a summary we have learnt that FGM experimented using these two constituents
of duplex stainless steel and carbon manganese steel proved to be better than X52 steel. The
manufacturing process is using wire arc additive manufacturing detailed mechanical
characteristics and durability characteristics are assessed using various tests and the results
show a positive recommendation of FGM for marine application.

So, there is no compromise on strength there is only improvement on resistance of corrosion.

So, that is how materials can also govern the steel design in the present context.

Thank you very much, have a good day.

210
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 12
Design Methods - 1

(Refer Slide Time: 00:24)

So, friends welcome to the lecture 12 of the course Advanced Steel Design, wherein this
course we are discussing about the design procedures which are in the recent practices. We
are also looking for the advantages of steel as a construction material plus alternate materials
substitute steel for its functional value like functionally graded materials.

So, in this lecture 12, we are going to learn more about design procedure in compliance with
Indian and international codes. As I told you we will not discuss about the conventional
design methodologies which is being taught in steel design classes. We are going to talk
about the intricacies of the design methodologies, and we are going to learn more about the
complexities involved in those design methodologies.

So, our focus is not about the conventional design methodology, we are talking about the
advancement. So, we are looking at the complexities, uncertainties, limitations, problems,
issues, concerns related to the design procedures which are currently in practice.

211
(Refer Slide Time: 02:04)

When we talk about design, what is the basic aim or objective of structural design let us ask
this question. The basic objective of structural design is to obtain the dimensions of the
member, its layout etcetera such that the structural form as a whole, will sustain the loads and
deformations to which it will be subjected to during its service life. So, that is the basic
objective which we all know. Now, we do this whole exercise with a margin of safety. So,
structural design inherently involves embedment of safety in the procedure.

Now, one may ask me a question what does the safety accounts for? What does this safety
take care about? Does it take care about the human errors made in the structural design or
does it account for something else beyond the limitations and control of the structural design?
We will answer this question shortly from a while from now.

The moment I say it has got to fulfill certain obligations of the structure has to fulfill the
functional requirements for which it is designed. So, its performance become the primary let
us say outfit of the structure. Now, how do you assess its performance? So, the question
comes how then the performance is assessed, because we have to check whether the
performance is fulfilled satisfactorily.

212
(Refer Slide Time: 05:49)

Performance is assessed by various limit states. When I say limit states there is something
called ultimate limit state, fatigue limit state, limit state of progressive collapse and
serviceability limit state.

So, there are four limit states which can become a tool box to assess the performance of a
structure with an objective that the structure should perform its intended function,
satisfactorily and safely.

Having said this, let us try to elaborate these limit states quickly and see what do the limit
states suggest us.

213
(Refer Slide Time: 07:38)

Let us say ultimate limit state. Ultimate limit state corresponds to the maximum load carrying
capacity of the structure. Fatigue limit state corresponds to assessing the damaging effect cost
due to cyclic load. So, fatigue limit state takes care of assessment of the structural system of
course, only for one specific class or type of load. Limit state of progressive collapse – this
corresponds to assessing the progressive collapse of the structure after rupture or after
excessive deformation.

So, limit state of progressive collapse does not assess the complete structure, it only assesses
critical sections of the structure, where the deformations are excessive.

214
(Refer Slide Time: 10:32)

Serviceability limit state is related to assessment of the structure under normal use and
durability. So, friends, if you look at this scenario each limit state has got a specific role to
play limit states or tools to assess the performance criteria of the structural system.

There are four limit states ULS, FLS, PLS and SLS. Each one of them have their own role to
play. Ultimate limit state only assesses the maximum load; fatigue limit state assesses the
structure only under one type of load; limit state of progressive collapse assesses the structure
only at critical sections whereas, serviceability assesses the structure under normal use and it
also takes care of the durability of the structure.

So, they play very distinct roles there are no overlaps of these limit state functions in terms of
their objectives upon the structure. You should also understand that one limit state does not
counter check the other. There is no mutual interference between them. Each perform their
application rules as per the laid procedure. Let us try to look into the ultimate limit state and
the other limit state slightly more in detail.

What does ultimate limit state do? It includes collapse mode of the structure caused due to
excessive yielding. Excessive yielding actually transforms the structure into a mechanism.
We will address what is a mechanism slightly later in plastic design, but it transforms the
given structure into a mechanism.

215
b) buckling due to elastic or elasto plastic instability. Now, this will lead to loss of
equilibrium in certain parts of the structure or in the whole structure. So, these are all
consequences of this collapse modes, an ultimate limit state takes care of these and addresses
the capacity or the maximum load capacity of the structure when subjected to these collapse
modes. So, ultimate limit state estimates load capacity. Let us have this clue with us.

(Refer Slide Time: 15:22)

Let us proceed forward what does this fatigue limit state does more in detail. Fatigue limit
state is important to structures under repetitive or cyclic load. The moment I say cyclic there
are two issues in a load – one is the magnitude, other is the duration or I should say number
of cycles. So, in cyclic load which may lead to a fatigue failure magnitude is not that
important, even a lower magnitude loads at large number of cycles can cost fatigue.

So, the magnitude is of course, a criterion, but that is not the governing criteria of a fatigue
failure. The magnitude plus the duration that is the number of cycles the load is applied on
the structural system. So, a fatigue limit state takes care of those structures which are purely
subjected to or which are exclusively under repetitive or cyclic loads. So, they will look
forward for estimate stress concentration factors.

There are some examples where cyclic loads are very common on structures which are
reversal or repetitive in nature. Classical examples could be a wave load, even wind load,
current, machinery loads, etcetera.

216
(Refer Slide Time: 17:38)

Now, the progressive limit state of collapse is of interest as this dissects only the behavior or
it dissects the behavior of the structure only at critical sections. This is very important when
the structure is subjected to accidental loads.

Serviceability limit state is assessing the structure under normal conditions and durability. So,
this will help to formulate the governing criteria for deflection and limitations in vibration.

When we start assessing the behavior of structures looking through these windows of limit
states and assessing their performance or functional value or utility value of the intended
function of the structural system under different segments of limit states as just now we saw.
The whole procedure of this assessment has got some uncertainties.

217
(Refer Slide Time: 19:47)

There are some uncertainties present in the whole system. Now, these uncertainties can be
due to variability and other effects caused in the analysis. So, when we have some different
issues and concerns in the analysis, then what do we focus at? The focus is to obtain the
strength of the structure as a statistical quantity.

So, strength can be lower and higher, it can be minimum and maximum, it may not be
consistent throughout. So, we look for a variable we look for a variation band and we express
the strength not as a single number or a single quantity, but we try to express this using
statistical tools. Now, we will express the strength in terms of its mean and standard variation
deviation.

Now, let us ask a question what are those factors which can cause variability in the analysis.

218
(Refer Slide Time: 21:57)

The first factor could be variability in the material strength and cross-sectional dimensions
like modulus of elasticity, area of cross-section, etcetera. For example, the bar may not be of
a uniform diameter throughout. The depth or the diameter can slightly increase or decrease
along its length due to manufacturing problems. Young’s modulus or modulus of elasticity of
the material may vary from batch to another batch of production systems etcetera. So, they all
will cause or induce variability.

The 2nd source arises from variability in geometric properties of the components. It may be
arising from weight, diameter etcetera. So, this can be associated to manufacturing defect or
oversight and can be controlled if you have a quality product from a branded company.

The 3rd uncertainty arising from in loads with respect to the environmental conditions.
Environmental conditions can govern change in loads. For example, wind, wave, current,
earthquake. They are environmentally controlled processes. So, there can be a variation there
can be uncertainties in predicting or estimating or computing these loads because they are
governed by environmental conditions. So, there can be definitely a variation in their
estimates.

219
(Refer Slide Time: 24:53)

The 4th source arises from idealization process of a real structure to an equivalent
mathematical model. So, while doing an idealization from real system to a mathematical
system we can make or we used to make certain assumptions. These assumptions can also be
a source of uncertainty.

The 5th comes from accuracy of the analytical or numerical computation. Numerical means
the output from an existing software; analytical means output using your own computer
codes. So, there can be error, there can be variation, there can be disagreement, there can be
invalid outputs coming from the analysis. So, they can also lead to uncertainties.

Of course, as we all know quality of workmanship can also lead to a high degree of
uncertainty in certain class of constructions. Loss of strength in the material due to corrosion
– so, there can be a strength reduction with aging of the structure when the structure is placed
in marine environment. So, these 7 factors can contribute significantly to the uncertainties in
the analysis of structural systems under environmental loads.

220
(Refer Slide Time: 27:26)

Now, the next question obviously, comes is how these uncertainties are addressed because we
have to any way address them in the design, right. We cannot leave them to propagate to
continue to grow. We have got to address this. How are they addressed in the design? These
uncertainties are addressed during the design stage itself. How do we do this? They are
addressed using safety factors.

There are safety factors separately for material and separately for loads, because you know
some uncertainties come from the material. Let us talk about them. Let us say this comes
from the material, this also comes from the material or manufacturing, this comes from the
load, this comes from both load and material, this also comes from load and material because
there can be analytical issues in this, this comes from material, this also comes from material.

So, there are various factors coming from different sources. So, they come either from
material or from the load. We are not talking about issues coming from the environmental
conditions, we believe that these conditions influence either one of them and that is
accounted directly if account for material or load purposes. So, we use safety factors for
them.

Since each one of them are independently address the design we call them as partial safety
factors. So, this is gamma m this is gamma f. So, m stands for the material and f stands for
the forces. So, that is the load. Now, they are advised by the code based on experimental

221
studies, research studies and they are usually discussed based upon statistical outcome of
these studies.

(Refer Slide Time: 31:02)

Having said this, let us take for example, ultimate limit state and see how this uncertainties in
the form of factors are addressed in ULS. In ULS, let us see how the safety factors are
embedded in the design process. When we talk about material as well as load we do not talk
about a single quantity, as I said we always deal with statistical quantities of these variables.
Codes intelligently state them as characteristic values.

In fact, characteristic load, characteristic strength; the word characteristic refers to

probabilistic model of the variable. Probabilistic model always refers to statistical outfit. So,
what we can understand here is uncertainties are addressed in ultimate limit state procedure in
a semi probabilistic manner in the form of characteristic load.

222
(Refer Slide Time: 33:34)

And, characteristic strength one applies to the load and then apply other applies to the
material. So, there is a very nice equation which we all know. Let us recollect that equation.
Characteristic strength for the material divided by factor of uncertainty should be definitely
greater than or equal to characteristic service load multiplied by factor of uncertainty.

Friends, please see here very carefully. As far as material is concerned we decrease the
strength right because we are dividing the factor. Believe that the number is more than 1. As
far as load is concerned we increase the load and this seems to be a very unrealistic procedure
because on one hand you decrease the strength, on the other hand you enhance the load and
then you say they are equal. This seems to be unacceptable, is it not?

But, there is a very interesting conclusion which we can draw if we start exploring this
slightly a bit more.

223
(Refer Slide Time: 35:39)

Let us see that for example, there can be a case where a specific live load in a given system
will be exceeded during lifetime of the structure. Let us say live load intensity of a specific
value may be x kilo Newton is likely to be exceeded during the service life of the structure.

We are not looking for this intensity. We are looking for what is the probability of this
exceedance to happen within the service life of the structure. Friends, that is a very important
argument which you must understand. In the whole design procedure if you look at carefully,
we are not interested in quantifying the magnitude. We are choosing a magnitude and trying
to quantify it is probability of exceedance of that magnitude within the service life of the
structure.

So, there are two comparisons happening here. We always compare the chosen magnitude
with service level structure. For example, let us say the design live load is 4 kilo newton per
square meter. If you ask a question to you what would be the probability of the live load of
about 1 kilo newton per square meter occurring on the structure, I would say the probability
of the live load exceeding 1 kilo newton per square meter is 100 percent.

Similarly, probability of exceedance of the live load of 2 kilo newton per square meter 100
percent keep on doing that what would be the probability of exceedance in the live load more
than 4 kilo newton per square meter that is what we are looking for. So, we have to choose a
value and look for the probability of exceedance of the value that is the art of addressing the
uncertainty in the load in the probabilistic sense, .

224
(Refer Slide Time: 38:41)

For example, if the live load or if the load coming from the environment if the load is
encountered from the environment there is always a chance that this load can be exceeded
because environmental conditions which will be prevailing during the service of the structure
cannot be guaranteed.

Now, you can look into the previous data of 100 years and you can say the wave load or the
wave height or the wave period or the wind speed may not exceed certain magnitude. But,
there is no guarantee because these loads are governed by the environmental process.
Tomorrow the process becomes volatile, becomes violent, changes drastically there is an
expected increase or higher probability of exceedance of your chosen number because they
are governed by environment.

Look into the intensity here this problem was not there in the live load because live load can
be estimated with higher accuracy whereas; environmental loads cannot be estimated with
that higher degree of accuracy. So, there can be loads where the dependence of the load or the
exceedance probability can be violently changing because they are environmentally
controlled.

So, in such situation what do we do? Here we are looking for probability of exceedance in the
service life of the structure. Here we will look for probability of exceedance of the chosen
magnitude of the load during the service life saying that the more the service life more the
probability of exceedance.

225
So, this relationship we keep in mind when we express this kind of factors of uncertainty for
the loads.

(Refer Slide Time: 41:42)

Now, friends, now let us consider two components which are important in design – one is the
load, other is the resistance. Of course, resistance comes from the material, come from the
member, form dominant design – let us talk about all those things, we know them, right. So,
for a good design we all do agree that resistance for a good satisfactory ideal design
resistance should exceed yes, is that not the load?

If resistance lesser than the load applied on the system or encountered by the system will be
unsafe, it is going to collapse. So, so the ideal condition we all know that your design
procedure cannot be reversed on this condition, but let us try to express this slightly in a
statistical meaning.

Let us consider that both R and S are obeying normal distribution. I am talking about
statistics now. We all know normal distribution it is a very simple expression available in the
statistics, you will be able to get that. Now, let us consider resistance with two distributions.
They have the same mean, but different standard deviation. Let us plot this.

226
(Refer Slide Time: 44:10)

Let us see. I am plotting r and s and f s and f r in the y-axis. Let us say this is my load
distribution curve normal distribution and this becomes my mean for the load. Now, let us
consider resistance of two distributions. One distribution is what I am showing currently now
which is having the mean somewhere here.

Now, let us also use this is capital S, we call this as f R 1. Let us have one more distribution
with the same mean, but a different standard deviation. So, we call this as f R 2. Now, you see
there is an overlying area where at any point here strength or resistance is lesser than the load.
So, this is the failure area. So, looking at this figure we can say that the probability of failure
is higher for distribution f R 2 in compared to that f R 1 because f R 1 has no overlap with f s
.

So, indicated with a shaded region is given by probability of failure 0 to infinity.

∞
𝑃𝑓 = ∫ 𝐹𝑅(𝑥)𝑓𝑠(𝑥)𝑑𝑥
0

if FR and fs are independent distributions. Now, friends, we will make a very important
comment here we say that the mean value of the load which is m s and mean values or mean
value because there are same mean value of the resistance which is m R which is m s.

I think you are able to see it here are useless for me as characteristic values, then what should
I look for? I want to look for these values as specified points are important. What are those

227
specific points we are looking for? We are looking for the specific points like I am marking
here. One is I look for S p, then I look for r p and I would say this gamma p is actually
gamma times of S p. We will explain what are these. Now, S p and r p are essentially let us
write it here.

(Refer Slide Time: 49:50)

S p is 95 percent fractile value of the load . So, let us say it is 95 percent . So, the remaining
is only 5 percent whereas, r p is 5 percent fractile of the resistance. So, this is only 5 percent.
So, this is 5 percent of the resistance, this is 5 percent of the low and we draw lines here.

So, I am looking for the values at specified points of our interest, right. So, now, if we agree
that both the load and material or the strength or resistance follow normal distribution then X
p will be m x plus or minus k p of sigma x equation 2; the previous one is equation 1, where k
p is the value at the fractile points.

Let us say in terms of p as percentage we can write k p 20 percent 10, 5, 2.5, 2.275, 1, 0.135,
0.0032; k p are 0.842, 1.282, 1.645, 1.960, 2.0, 2.326, 3.0 and 4.0. These are the fractile
values of these points, .

228
(Refer Slide Time: 52:50)

Now, different codes recommend different fractile values. For example, euro code states k p
should be chosen corresponding to 2.0. What does it mean? If we choose correspond to 2.0 I
am looking for the probability of exceedance of load or probability of non exceedance of
resistance ok only as 2.275 percent right in 100.

So, they define characteristic load based on this. So, characteristic loads are defined based on
the values at the fractile points of our choice assuming that load and resistance obey or follow
normal distribution. So, therefore, the factor gamma which you see here this factor ok is
going to be r p by S p. So, which is nothing, but characteristic strength by characteristic load,
so, this is the measure of safety .

229
(Refer Slide Time: 54:54)

So, for design purposes this gamma for design purposes this factor of safety is divided into
two parts – one is gamma f, other is gamma m. Gamma f relates to load gamma m relates to
material. So, in simple terms gamma f gamma m should be r p by S p. So, therefore, friends if
you want to combine or join them we should say gamma f multiplied by S p should be
gamma p divided by gamma m.

So, this is what we call as a design load, this is what we call as a design strength. This is the
same equation what we wrote earlier see here, the same equation. So, one is a multiplier,
other is a divider. See here, one is a multiplier, other is a divider.

230
(Refer Slide Time: 56:09)

So, friends in this lecture, we learnt about limit states, we learnt about characteristic, values
of load and strength of the material. We also understood the fractile values of the probability
distribution function and we realize that different codes designate different fractile points to
estimate the characteristic values of the loads and the strength of the material.

Thank you very much. Have a good day.

231
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 13
Design Methods - 2

(Refer Slide Time: 00:20)

So, friends, welcome to Lecture 13 of the course on Advanced Steel Design, in this lecture
we are going to learn more about the Partial Factors. So, this lecture is also focusing on
γ𝑝
Design Methods. In the last lecture we said γ𝑓𝑆𝑝 ≤ γ𝑚
. So, we call this as the design load

and we call this as a design strength. whereas, the factors address the uncertainties which are
arising from various sources as we discussed in the last lecture.

232
(Refer Slide Time: 01:34)

So, having said that the factors present in the load and the strength are going to take care of
these uncertainties they are anyway addressed only as partial factors. So, let us take for
example, the load factor which is γf, in the literature γ f is actually a product of γ f 1, γ f 2,
γ f 3. So, it has got three subdivisions let us see what are these three subdivisions address.

γ f 1 accounts for load deviation from the assigned characteristic value. If you say the
characteristic value is 5 percent the exceedance is only about 5 percent, but we exceeds
beyond 5 percent what is that exceedance, beyond the characteristic value will be looked at γ
f 1 separately.

γ f 2 will account for reduced probability that several independent loads appear together with
the same characteristic value. When several loads with the same characteristic value appear
together there can be reduced probability of that combination.

233
(Refer Slide Time: 04:02)

So, γ f 2 looks for that, γ f 3 accounts for possibility of inaccuracy in the load value arising
from mathematical error or mathematical model error, variation from the real problem,
problem arising during erection or construction etcetera. So, amongst these if you look at
carefully γ f 2 addresses the effect of load combination.

So, even within the partial safety factor of load itself there are three sub classifications.

(Refer Slide Time: 05:15)

234
Similarly, if you look at γ m, this accounts for possibility of unfavorable deviations of yield
strength of the material from its characteristic value. There can be factors arising from local
weakening of the section due to or during eruption. There can be difference in the material
strength for example, σu in the lab and that observed in the field.

So, in the field also you conduct experiments, you find out ultimate strength and Young’s
modulus in the field can compare those data directly with the manufactured supplier’s data.
So, there can be a variation and that difference is accounted in γ m. Further there can be
dimensional inaccuracy which can arise during erection. There can be also error in
mathematical modeling of the material.

However, many codes do not distinguish these factors elaborately and do not give the
percentage contribution of γ m or γ f from independent factors, they give a single combined
number. So, for example, if γ f is 1.5 and γ m is 1.15 as in Indian codes, you will not come to
know this 1.5 will be contribution from various sources. What is the effect of every factor on
this γ? You will not know these independent contributions, but they are together contributing
to γ, we can only know that.

(Refer Slide Time: 08:32)

Let us now see something called load effects, see, what is the load effect? Based on the
effects caused by the loads on structures loads can be categorized as P class load, L class
load, D class load, E class load and A class load.

235
P class refers to permanent load, L class refers to live load, D class refers to deformation
load, E class refers to environmental load and A class refers to accidental load.

(Refer Slide Time: 10:02)

Let us see them slightly elaborate P class load, which are permanent load they include time
dependent loads. Examples could be gravity load, weight of any permanent machinery on the
structure and their variation over a period of time. So, the good thing about this class of load
is this class of load is generally estimated with higher degree of accuracy.

Hence, the characteristic value of P class load is generally taken as an average of the data
available and we can compute this with higher accuracy.

236
(Refer Slide Time: 11:42)

Let us come to L class load, L class load anywhere refers to live load, live load are associated
with nature and operational features of the structure. That’s why they are called live load. So,
now, the choice of the characteristic value depends on type of the structure not on the type of
the load.

(Refer Slide Time: 12:39)

When we talk about D class loads which refer to deformation loads, these loads are imposing
the imposed deformation on the members. It may be due to temperature effect, due to
differential settlement, due to settlement of supports, due to form dominance nature of the

237
structure. So, the characteristic value of D class is prescribed as the maximum expected
values, it is an upper limit.

(Refer Slide Time: 14:09)

If you come to the E class load which refers to environmental load, they are associated with
loads arising from the environment. Examples could be wave load, wind load, load arising
from current, and earthquake load etcetera. They are generally classified with the help of a
term called mean return period.

Mean return period is defined as the expected number of years between a seasonal maximum
to occur. For example, if the return period is 50 years.

238
(Refer Slide Time: 15:40)

It means that, it means that the characteristic load is expected to occur only once in 50 years
that is the return period. Now, inverse of the return period is a very interesting parameter, this
is the probability that the characteristic value will be exceeded in any one year of its
application.

For example, if the return period of wind load is 100 years, inverse of this is going to be 0.11.
So, we can now say that probability of the wind load that is the characteristic value will be
exceeded in any one year is 0.01 which is nothing, but 1 by 100. So, 99 percent it will not
exceed in any one year, but 1 percent there is a probability it may exceed. That is what the
meaning of inverse return period.

Now, friends we have the return period which is expressed for the environmental loads we
have seen the inverse of the time period also.

239
(Refer Slide Time: 18:05)

Now, we have seen that return period and its probability of exceedance in one year. Now,
what the question we want to know is; what is the probability that the characteristic load will
be exceeded during the lifetime of the structure? What is the probability?

So, now, since 1 by R is the probability of that load exceeding in 1 year 1 minus 1 by R will
the probability of non - exceedance in 1 year, is not.

(Refer Slide Time: 19:36)

240
So, the probability of non - exceedance of the load in 1 year is 1 minus 1 by R, what is the
probability of non - exceedance in n years will be 1 minus 1 by R to the power n. Now, we
are looking for what is the probability of exceedance of the characteristic load in n years.

So, I should say probability of exceedance in n years 1 minus 1 minus 1 by R to the power n
right.

(Refer Slide Time: 20:31)

For example, if the return period is 50 years and the lifetime of the structure is 20 years we
can now say the probability of design load exceeding at least once during the service life that
is 20 years of the structure could be P 20 which is 1 minus 1 minus 1 by R to the power n 1
minus 1 minus 1 by 50 to the power 20. So, it has got an exceedance of 33 percent, which is
an alarming number.

So, it is interestingly friends, the return period is 50 years, service level 20, you may possibly
think the exceedance will not happen at all see here it is about 33 percent. So, it is clear from
the example that there is a chance of design load exceeding at least once in the service life is
about 33 percent which is an alarming number. Therefore safety factors are important.

241
(Refer Slide Time: 22:11)

The last one is an accident class loads A class load refers to accident class, these loads are
exceptional cases, they arise when there is a collision, when there is an explosion or fire or
there can be accidents crossed due to dropped objects. Now, the characteristic load of the
dropped objects or accidental load of this class depends on the subject of investigation and
operational manner etcetera.

(Refer Slide Time: 23:03)

242
So, friends in this lecture we have discussed about different class of loads, we have discussed
about return period, we have discussed about the probability of exceedance of the return
period and we have seen how the return period is connected to service life of the structure.

So, please revise these concepts and take many examples from the textbook, try to look into
them, read research papers from various journals in this connection and update your
knowledge and know how on these friends.

Thank you very much have a good day.

243
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 14
Load combinations

(Refer Slide Time: 00:21)

Friends, in this lecture 14 on Advanced Steel Design course, we are going to learn more
about the Load Combinations in the design process. We are going to talk about the factors
that affect the material strength and we also talk about the FGM manufacturing process and
testing procedure as a part of the extended discussion we had in the previous lectures.

244
(Refer Slide Time: 01:30)

So, in the last lecture, we summarized of saying there are different class of loads which are
called or classified based on load effects and they are classified as P class load, L class load,
deformation load, environmental load, and accidental load is not; we discussed about this in
detail in the last. We also discussed about the effect of environmental load in the context of
return period and we learnt how this can be handled, .

(Refer Slide Time: 02:12)

Now, we will discuss about the load combinations. Several loads actually act on marine
structure simultaneously; not only marine structures, in general even land based structures

245
they act simultaneously. So, generally the combined action of these loads are accounted using
load combination factors; you must have seen and you have used in design.

For example, if dead load and live load are combined, we use generally a combination of 1.5;
if dead load and live load and wind load are combined certain course advises 1.2 and so on.
So, this factor affects for or accounts for the combination effects of these loads provided that
there is an assumption that these loads act simultaneously. So, now, when we talk about the
estimate of load factors, the combination effect should also be considered.

(Refer Slide Time: 04:08)

So, determination of load factors gamma f, in fact we already saw gamma f has got three
different components which are actually not explicitly given in many codes; but still it are all
inherent representation of different factors as we discussed in the last lecture. So, when you
talk about the determination of this factor gamma f, which is to be now used along with load
combination effects; then we will see how this is handled in the codes, .

So, now we know each load be it a dead load, be it a live load etcetera has a specific
probability of exceedance; they also have a specific characteristic value to be used in the
design, is not. And obviously, we all agree the probabilistic exceedance of these loads for
example cannot be same right, they can vary; because some of them are environmental
dependent, some of them are arising from the machinery equipments, some of them arise
from the gravity loads of the structural system.

246
Therefore, the probability of exceedance of these loads need not be same. So, I have to
account for the probability of exceedance of these forces or loads along with the combination
effects and that should give me the appropriate load factor which I should use in the design is
not. So, I should now say that the load factor apart from accounting for various uncertainties
should also consider the probability of exceedance of different loads, when they are
combined.

Now, the question is you do not have a choice whether to combine them or not, in nature they
will act simultaneously; for example, in a marine structural system wind load, wave load,
current load will always be there. You cannot imagine a situation where wave load is absent,
wind load is absent, you cannot have a situation.

Because these loads are essentially arising from the environment; therefore they will be
combined and they will act simultaneously. So, the load factor should account for this,
correct.

(Refer Slide Time: 07:48)

So, if we agree on making this statement, then the combination of such loads should take the
following form. Let us say S d is S times of gamma f of P capital P gamma f of l capital L
plus gamma f of e capital E and so on. Let us try to explain what do we understand by the
symbols. So, in this S d refers to the design strength, which we want to incorporate in
the design.

247
S is actually a subset of these loads; let us say P in this case refers to permanent load, L refers
to live load or I should say permanent load characteristic value, live load characteristic value
and E refers to environmental load characteristic value etcetera, . So, now they all should be
combined with different factors and these factors for example, gamma f should be a
combination factor of permanent load, which need not be equal to gamma f of that of the live
load.

So, you can have different combination effects of these loads and adding all of them together
should give you the design strength.

(Refer Slide Time: 10:11)

So, in this context let us see what international codes advise; let us take euro code for
example, if you talk about the type of loading and the effect of this loading on structures and
the appropriate gamma f advised by the euro code; if you say for permanent class of loads,
there are two cases they say unfavourable condition and favourable condition. For the most
unfavourable condition if you want to design the system, then use this value as 1.35; you
want to design it for a most favorable condition, then use this factor as 1.0 for permanent
loads.

Let us say for a variable load, it can be even the live load for example; then you look for the
most unfavourable condition only for this kind of variable loads and the factor is 1.5 and for
the favorable condition of design this is not suggested. So, friends looking back these values,

248
now looking back this equation; now I can say gamma p p can be either 1.5 or 1.0, whereas
gamma l L can be 1.5.

So, as I said they need not be equal. So, load factor cannot be equal to all kinds of loads when
they act simultaneously. So, the combination effect of these loads also play a role in
estimating the load factor, that is the point we want to emphasize.

(Refer Slide Time: 12:19)

Let us see what AP RP code says, American Petroleum Institute recommended practice;
LRFD load resistance factor design code, . So, go to the provision given by API. It indicates
for the purpose, for the type of load and the load factor. If you say the purpose is to calculate
the internal load effects, let us say that is the purpose and for all loads you can have a
common factor as 1.3.

If the purpose is to arrive at loads during specific operations, such as load out, launching
etcetera; these are all specific type of operational conditions which are in prevalence when
you commission an offshore platform. My textbook on ocean structures and materials
published along with Arvind Kumar Jain for CRC press; we will have more details about
what are these kind of load conditions and how do they actually occur and what do they
mean.

So, please refer to that text book for more information. So, what we learn here is these are all
different specific operations happening during marine construction installation time. For that

249
case if you have a gravity load arising during this operation, then your factor is 1.3; if you
have an environmental load arising during this operation, then the factor suggested is again
1.3.

If I have a gravity load acting along with the environmental load, where environmental load
is predominant ok; in that case for gravity load, the combination factor is 1.3 sorry 0.35 and
for environmental load the value is 1.10.

For towing operation you want to calculate; then for a gravity load the factor suggested by the
code is 1.10, for environmental load the factors suggest is 1.35. So, friends one can see here
very clearly that gamma f is not a single number, it keeps on changing depending upon the
combination, even depending upon the type of purpose what you want to compute.

(Refer Slide Time: 15:56)

Let us look one more insight what we see in the DNV code, which is offshore standards
which is DNV offshore standards c 1 0 1. In this case it clearly says, the combination of
design loads and the load category and the categories are as we see permanent load, live load,
environmental load, deformation load and accident load. And if it is an ordinary condition or
ordinary or operating condition, then the factors for different class are indicated as you see
here; obviously during operational conditions accidental loads are not present.

If it is extreme condition, then you reduce the combination from permanent live load and
increase the combination from environmental load and accidental loads are to be computed as

250
per actual,. So, you can see there are three different codes advice in different manner as the
load factor, which is to be multiplied by the imposed load the characteristic value to compute
the design load on a structural system.

So, we have seen in detail that the load factor accounts for the uncertainties present in
estimating these loads, the probability of exceedance of these loads, different load
classifications, their combination effects and their operational conditions with respect to their
combination effects how these loads or factors are influenced, .

(Refer Slide Time: 18:33)

Having said this, let us now go one step ahead and discuss what will be the fabrication
factors, fabrication factors that can affect the strength of structure. The strength of the
structure can also be affected by various factors during fabrication. Now we all know steel
consists of many constituents material and each constituent material plays a significant role in
affecting the strength. Let us see how these factors affect the strength. Let us try to plot
percentage carbon in x axis, we will plot the strength in megapascal.

So, let us say percentage carbon as 0.5, 1, 1.5 and 1.7; let us just to draw these grids for our
learning. Let us have these as the control points. Let us say this is 250 megapascal and this is
500, let us say goes till 1000. So, let us try to plot the tensile strength variation starting from
275, it increases with the carbon content; it goes as high as 1000, close to 1000, then drops,
this is the tensile strength.

251
Let us talk about the maximum strain or maximum elongation. The maximum elongation
decreases from 1000 with the increase in carbon content. So, we can say we will say the
tensile strength increases with increase in percentage carbon, maximum elongation decreases
with increase in percentage carbon.

(Refer Slide Time: 21:52)

Apart from that, the alloy of steel also affects the strength. Let us say steel alloy has got a
combination of iron, carbon; I am only writing the major components and various alloying
elements. These alloying elements are essentially manganese, silicon, nitrogen, hydrogen,
oxygen, etcetera phosphor, sulphur etcetera. Now, with respect to carbon, it increases
toughness, yield strength; the presence of carbon increases toughness, yield strength and
tensile strength.

We just now saw the figure, but it decreases ductility, maximum elongation and it also affects
the welding properties.

252
(Refer Slide Time: 23:31)

Now, with respect to the alloying elements, let us quickly see what are their contributions;
manganese increases the tensile strength, the presence of manganese increases tensile
strength, it also increases the weldability. Silicon increases toughness, tensile strength and
yield strength; but it decreases the maximum elongation and cold form ductility. Phosphor
decreases ductility and notch toughness; we have heard about this, we know how they are
important in classifying steel.

Sulphur decreases ductility, weldability and notch toughness. Nitrogen presence decreases
resistance against brittle fracture. Hydrogen decreases ductility, weldability; carbon increases
corrosion resistance. So, varying alloying elements also influence the characteristics,
mechanical and performance characteristics of steel in a different segment as you see on the
screen just now.

253
(Refer Slide Time: 25:45)

Now, friends when we start fabricating a structural system using steel as a construction
material, we all know there will be connections; there will be joints which are in unavoidable,
you cannot have a single piece for the entire system. So, these connections are joints
generally are welded or sometimes they are bolted; riveting is of course not in practice
because of various reasons, but welding and bolting is still in presence.

Now, these specific operations also induce something called residual stresses in steel
structures. So, residual stresses are generally caused due to manufacturing and or fabrication
process. Manufacturing is during the steel manufacturing itself; fabrication is during erection
fabrication in the site, they are two different things.

So, what are those common processes which influence the strength, residual stresses? The
common processes are 1 welding, which of course the most common process; 2 something
called flame cutting, this is a proper process used to make openings. 3, hot rolling; this is a
manufacturing process, this generally used for manufacturing plates or flats. So, they
introduce residual stresses. Now, if you ask me a question, introduction of residual stresses
causes what kind of problem?

254
(Refer Slide Time: 28:01)

So, residual stresses affects the strength, it induces brittle fracture fatigue, it can induce stress
corrosion. Stress corrosion is a induced corrosion caused because of high stress concentration
at the connections; it can cause buckling failure, it can induce inelastic behaviour in the
member, not on the material please understand.

Now the question comes; if I want to get rid of this residual stresses, what should I do? Now,
to get rid of residual stresses following are advised; 1, you can preheat the member before
welding; 2, you can post heat the member after welding; 3, can also do something called
peening, hammering the weld, this will help to elongate the weld and this will relieve the
tensile stresses which are residual in the weight, . So, these are some of the aspects what we
have, which I want to discuss about the factors affecting the strength.

255
(Refer Slide Time: 30:45)

In continuation to this, we also said that functionally graded materials is a combination of 2

or more metal manufactured using wire or additive manufacturing method using CRT torch.
So, we discussed this procedure very much in detail; we also shared the important research
findings and experimental outcome of FGM fabricated in our lab at IIT Madras.

So, now I want to show you a quick video, to impart you or to impress upon you that this
material manufacturing is very simple and the testing proves the data what we supported. So,
recollect this, we have taken a specimen of this nature if you remember. And in this one
portion which is our 3 mm thick is material A, which is approximately 3 mm thick in our
study, which is duplex stainless steel. The remaining material about 14.5 millimeter is carbon
manganese steel.

As I said very interestingly friends, the load is applied in tension along the boundary; this is
the boundary right, this is the boundary between the two alloying metals, it is along and you
will see there is no fracture notified along the boundary, but the fracture is happening here.
So, I would like to show you a video at this moment and share this idea, just give me a
minute; I will share the video, then come back to this.

256
(Refer Slide Time: 33:12)

So, friends you can see the CMT torch is doing the wire feed on the stainless steel strata base
strata, depending upon the x y coordinates laid for laying the wire and the wires are laid in
layers and this is wire arc additive manufacturing method, which you see on the screen at this
moment, . So, this is fabricated layer by layer and you get a specimen as we discussed in the
last lecture.

(Refer Slide Time: 33:51)

Where the specimen is placed in the UTM under axial tension; there are two pivotal points
which are placed along the longitudinal direction and two along the lateral direction, there are

257
four points you can see. So, these are the gauge points where you are measuring the
elongation and computing the percentage elongation. So, it is a high speed camera which is
recording this, the system is being loaded now; you can see the specimen very close at this
moment.

So, the system is applying an axial pull; now you can closely see the boundary, you can
closely see the boundary of the material. The edge is duplex stainless steel, the major part is
carbon manganese steel; there are two dots which are going to measure the elongation along
the longitudinal direction and there are two normal to this, which will measure the elongation
along the lateral direction.

Please look at how this material fractures. Can you see the fracturing happening here; it is not
fracturing along the boundary of this, but is fracturing like a simple tensile bar made out of
homogeneous material. So, this is a very interesting exhibit which we wanted to show you,
where we have learnt the fabrication method, the constituents of FGM and of course the
testing procedure.

(Refer Slide Time: 35:49)

So, friends as a summary in this lecture, we learnt about the load effects, about the
combination of loads, about the residual stresses and also we learnt about the alloying
elements of steel and their effect on strength. We have also seen a video, which shows the
wire arc additive manufacturing of FGM and we also saw how the FGM breaks in UTM,

258
which is a tensile fracture similar to a homogeneous material like mild steel. So, friends this
is what we wanted to add at this particular point of time in this lecture.

Thank you very much, have a good day.

259
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 15
Dynamic material strength

(Refer Slide Time: 00:21)

Friends, welcome to the 15th lecture on Advanced Steel Design course, in this lecture we are
going to talk about the Dynamic material strength and the response criteria with respect to
steel as a construction material. We already said the design of structures is to aim creating
resistance against the loads, in fact, to be very clear if the load effects arise out of accidents
like fire, explosion, and impact etcetera, Then, the material strengths need to be explored
under such conditions because the material strength varies under such impulse loading or
under such time variant dynamic loading.

260
(Refer Slide Time: 02:18)

For example, structures under blast loads undergo plastic deformation, because this is
required to absorb the explosion energy. Therefore, these types of loads also inherently affect
the dynamic characteristics and strength of the material.

On the other hand, structures under conventional loads like dead load, live load, etcetera shall
remain or continue to remain in elastic range. So, the design philosophy also intricately
address the type of load and the preferred response. So, now, the design procedure itself is
multi-dimensional, the design objective itself is varying, therefore, we cannot have a single
design procedure which can address all kinds of loads and all kinds of materials.

261
(Refer Slide Time: 04:25)

Having said this, let us quickly understand what is the difference between static and dynamic
response, because it is very important for us to learn this very briefly, so that we know how
the dynamic material strength is understood. Now, let us for example talk about conventional
loads; conventional loads such as live load, dead load in fact, I would also call wind load as a
conventional load.

They are generally applied relatively at a lower speed or lower rate of application on the
structure. And usually they remain constant for a long period of the structure. Now, when I
say I have to compare the lower rate, longer period, then this is generally in comparison to
the time period of the vibrating system.

So, compared to the time period of the vibrating system these are all lower rate of application,
longer period of stay, etcetera.

262
(Refer Slide Time: 06:16)

On the other hand, if you look at blast loads they are applied or act very rapidly and they will
give rise to the stresses on the member very rapidly. So, one can very well say blast load
looks like a transient phenomenon. So, what does it mean? The structure will or expected to
return to normal, operational conditions in a short period of time. So, time period plays a role
in understanding the effect of these loads on the structural systems.

Therefore, in relationship to this what will happen in the conventional design?

(Refer Slide Time: 07:50)

263
In the conventional design stresses are limited within the elastic range whereas, in blast
resistant design for example, yielding should be permitted because it is one of the form of
absorbing the explosion energy. So, since yielding is allowed member response is moved to
plastic region. So, during this process it continues to absorb the blast loading or blast energy
by balancing the kinetic energy of explosion against the strain energy of the member.

(Refer Slide Time: 09:54)

Now, when we accept this statement we say that now the total strain energy is a function of 1
dynamic material properties, and 2 section properties, and 3 the amount of plastic
deformation permitted.

So, the total amount of blast energy required to be absorbed let us say we look into the
demand. So, the demand is from the blast energy. So, the total amount of blast energy need to
be absorbed should be a function of the peak load and of course, as you correctly guessed
duration of the blast t and peak load let us say amplitude A bar.

So, now, the adequacy of the blast resistant design of structural systems is based on the
maximum deformation rather than the stress level. So, now the blast resistant design
philosophy should be based on the maximum deformation level it is not based on the stress
level. So, therefore, a conventional design, which talks about the stress as an indicator in the
design parameter is of no use under these kinds of loads.

264
Having said this let us see, how the material response, we have seen how the structure
responds.

(Refer Slide Time: 12:44)

Now, let us see how the material response to this. When a material is loaded rapidly, it cannot
deform at the same rate at which the load is applied. So, what I mean to say is the rate of
loading cannot be equal to rate of response or rate of deformation of the material.

So, this disagreement creates an increase in the stress level at which yield occurs in the
material. It also affects the ultimate stress achieved in the material. These are the two
influence caused by this disagreement. So, what will be the immediate consequence of this?
The immediate consequence will be ultimate stress will be reached before rupture that is the
problem. That is a very interesting, understanding which we must know about the material
behavior when this or such disagreement is seen.

265
(Refer Slide Time: 15:44)

Having said this, we now clearly say increase in strain rate leads to strength increase, it leads
further to a consequence that member develops structural resistance in excess of the static
loads. So, this strength increase is quite significant, but its effect on the design is not too
important because the rate of loading remains only for a very brief period compared to the
service life of the structure therefore, this benefit of increase in strength cannot be accounted
in the design as a significant parameter.

(Refer Slide Time: 17:19)

266
Having said this, when we compute the flexural response of the member, we generally do not
account for the strength increase caused by the increase in strain rate. I mean this statement is
kept in mind and applied to steel as a material. Please note that this in the background
information is we are talking about steel as a construction material.

Now, if the strength is ignored as we said in the previous sentence then the forces or the loads
at the joints or connections will be underestimated. Please understand why? Friends, please
note this is an internal effect caused by the member. So, I should say this is a reaction, now
this reaction will have a counter-effect on the joints.

So, whatever loads coming on the joints now whether it is external by the moment caused by
the loads or internally caused by the material or the member behavior, you will be
underestimating that there is the danger here there is a danger here you are underestimating
that.

Now, you ask me a question; how this is accounted in the design? Because this has to be
accounted in the design how? This is accounted in the design by using a factor called
Dynamic Increase Factor, DIF. This is very much prevalent in blast resistant design or design
under accidental loads.

Let us go at one step ahead.

(Refer Slide Time: 20:25)

267
And discuss about the resistance deflection function. Generally friends, in flexural response
stress rises proportional to strain in the membrane that is a usual practice, generally it rises
like that. Since, resistance offered by the material primarily and by the member because form
dominance is also a function of stress, it rises in proportion to strain.

So, after the material extreme fibers of the cross section reaches the yield point or the yield
strain, relationship between stress and strain becomes non-linear. Please note very carefully
here, this becomes non-linear only at that instance when the extreme fiber alone reaches yield
strength not the full cross section.

only just when the extreme fiber reaches the yield strain stress strain behavior becomes
non-linear immediately. There is no proportionality between the rate of increase of strain
and the rate of increase of stress, they are no more linear right, I am talking about steel is a
material in mind.

(Refer Slide Time: 23:20)

Having said this, further as the extreme fiber yields stress in the interior section also begins to
yield, because the load will be transformed the strain will be transformed to the next adjacent
fiber. So, that starts yielding now. And this will initiate a plastic hinge. We will talk about the
plastic hinge and plastic design in detail in the next module lectures, but let us say it initiates
a plastic hinge.

268
Plastic hinge is an imaginary section that yields and therefore, rotates without any additional
moment, that is, at plastic hinge, the moment capacity, which makes the hinge to rotate is Mp,
but any moment addition to this is not required we will keep on rotating at that moment . So,
plastic hinge will be initiated.

So, where they will initiate as you correctly guessed, these plastic hinges will initiate at the
sections of maximum bending moment. I am talking about flexural failure.

(Refer Slide Time: 26:04)

If, premature buckling is prevented we will talk about the buckling failure in detail, but let us
know a hint here how do you prevent a premature buckling.

You ensure that the section has desired compactness limit. Compactness limit is a factor
given in the code which depends on the cross sectional dimensions of the member and if that
factor is satisfied the member or the sizing of the member will not initiate premature
buckling. So, it is simply a geometric choice of a cross section right. For example, b/d, h/t,
b/tw, h/tf etcetera.

So, if this is satisfied then premature buckling will not happen. Let us say if premature
buckling is prevented, then the deformation at the critical section on the member can be
related to the strain producing a relationship between resistance and deflection.

Let us try to plot this in the x axis, let me plot the deformation in the y axis, let me plot the
resistance. Now, let the curve has three slopes, let us see different distinct points, we call this

269
as y maximum, this is y yield point and this y elastic . W call the corresponding values as R y
and Ru. So, area under this curve as we all know with respect to x axis will give me the strain
energy.

(Refer Slide Time: 29:48)

Having said this, let us say an elastic resistance is the level at which the material reaches
yield at the critical location or critical section of the member, where the moment is maximum.

So, this is termed at the 1st yield. Now, beyond the first yield point plastic regions will be
formed in the section and it will lead to elastic - plastic deformation. In this condition,
internal resistance continues to decrease because plastic hinges are formed first yield has
occurred internal resistance continue to increase as the stress in other locations or other
sections rise in response.

So, response now will shift to the adjacent next critical sections and the response keeps on
rising in adjacent sections.

270
(Refer Slide Time: 32:37)

But this will occur at a lesser slope than the elastic region. Please see the curve, the slope of
this if it is theta 1, if this is theta 2 theta 2 will be far lesser than theta 1 . It will occur it
means a large deformation which is plastic will occur at a small rise in strength gain. So, up
to this place up to this place is proportional beyond this it is not. That is what we are
emphasizing here.

So, during this period some sections of the member are already in plastic region, while
remaining are in elastic region and this variation will happen along the member not across the
cross section. Please understand, this will not happen across the cross section, can you tell me
why? Very simple the moment varies along the span.

The response is directly related to the moment demand. So, this variation will happen along
the member and there will be a continuous shift of critical sections from a b c and so on,
which will slowly get plasticized, while the remaining will become elastic.

271
(Refer Slide Time: 35:22)

So, friends, in this lecture we are learning about the dynamic strength variation and elastic
plastic response behavior under blast resistant design procedure, not necessarily blast I can
even call this as accidental loads. So, hope you will be enjoying this and you will be
following these lectures back to know more about this. Please have a look at my Advanced
Steel Design book which is a very useful handout and gives more information about these
processes for your comfortable pace of learning.

Thank you very much, have a good day.

272
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 16
Material properties variations

(Refer Slide Time: 00:17)

Welcome to the 16th lecture an Advanced Steel Design course friends. In this lecture, we are
going to learn about, the Material properties and its variation in the design perspectives.

273
(Refer Slide Time: 01:01)

As we said in the last lecture, the resistance deflection curve, which is idealized as shown
here - Deflection versus Resistance.

We have already seen, this becomes R ultimate and this becomes R yield. And of course, this
becomes my deflection at yield and this becomes elastoplastic, and that becomes my
maximum thing and we already know the area of this curve under the x axis will give me the
strain energy. So, when the material reaches the yield point, reaches the yield point beyond
that the slope changes, and the slope is lower compared to the earlier one which is in the
elastic region.

As a response continues, other critical sections reach yield value and that will cause
formation of additional plastic hinges. We will talk about this procedure much in detail in the
next module on plastic design, but we are now talking about the resistance deflection curve.
So, little bit let us try to understand what happens to my resistance curve.

So, at yield point changes the slope of resistance of deflection curve and hence, the section
starts yielding after the last section yields no additional resistance is available is it not,
because all critical sections I have been now yielded and therefore, the resistance deflection
curve becomes flat. So, having understood this, let us try to apply more understanding on
blast resistance with steel as a structural material.

274
(Refer Slide Time: 04:10)

So, let us take structural steel and we are specifically talking about low and medium carbon
steel. I think we already given this classification; you should be able to recollect them. We
know that these two grades of steel are ductile enough and therefore, they can be used for
blast loads or let us say blast-resistant applications. Because as I said, the structural form will
start yielding on successive cross sections, which will lead to elastoplastic behavior of the
material.

And therefore, we need to look for a material which is having a good ductility under blast
loading of the structure. So, we should also put a caution here use of high strength steel for
blast loads should be avoided. So, I think very well we notice here that based on certain types
of loads on certain application strength will not govern; it is the function that governs. One of
the functional characteristic of durability is that is a ductility and corrosion resistance.

Let us talk about ductility. So, therefore friends, to avoid large deformations, or to be very
specific to achieve or to ensure safety at large deformations, let us put it like this. Because,
when we talk about form resistant design, large deformations are large displacements; will be
anyway a part of presence you cannot avoid that. So, to ensure safety at large displacements
without failure, steel members must be sufficiently braced.

275
(Refer Slide Time: 07:20)

This will help; this will help to avoid buckling failure and ensure stability. So, now, let us go
one step forward and look at the dynamic material properties. In this we will first start
looking at the stress - strain relationship. Response of the material under static and dynamic
load is governed by the stress - strain relationship, is it not; both under static and dynamic
load will be governed primarily by the stress - strain relationship of the material.

We all know that steel is a ductile material and be able to and therefore, will be able to
achieve large strain before rupture.

(Refer Slide Time: 09:15)

276
Having said this, let us say look at specifically a low carbon structural grade steel. A low
carbon structural grade steel exhibits a well-defined yield point, which is followed by a flat
yield plateau, but high strength steel does not show this behavior.

They show a sharp break at the elastic limit and the yield region will be highly non-linear.

(Refer Slide Time: 11:06)

Therefore, we can say that low-carbon steel are well suitable for form dominant design are
structures, which undergo large displacements; one example could be structures under blast
loads.

277
(Refer Slide Time: 11:49)

So, let us quickly look at the typical stress - strain curve has given by ASCE Manual. So, this
is a typical stress - strain curve of ASCE Manual 42, I am just reproducing that here. It shows
the initial slope with a small kink and this is Fu 882 Mpa steel is called as Rail steel. The
other one has got the same slope, but goes flat plateau and then this intermediate grade steel,
whose Fu is 517 Mpa.

We have one more steel which is much below this which shows a plateau which is structural
grade steel; whose Fu is 441 Mpa.

(Refer Slide Time: 13:38)

278
So, one can very well see here, I must go for one should prefer a long resistance deflection
curve to absorb the blast energy and this will also ensure no brittle fracture.

(Refer Slide Time: 14:33)

Now, let us talk about something called strength increase factor strength increase factor. In
practice the average yield strength of the material is approximately 250% greater than the
specimen minimum values. In general, average yield strength is about 2.5 times more than
the specified minimum values for those grades of steel. Therefore, it is obvious that one
should account for this advantage by using what is called strength increase factor. Let us say
SIF.

This takes care of the additional strength, which is unrelated to the strain rate property of the
material. Let us see one cross reference of a core. Let us say what does it say.

279
(Refer Slide Time: 16:53)

UFC – 3 – 340 – 02 suggests a strength increase factor of 1.1 to the minimum yield stress for
structural steel with σ𝑦= 345 Mpa. Therefore friends, using strength increase factor is actually

to control the conservatism in the design, and make use of full capacity of the material.

The next point related to the design of steel will be dynamic strength increase factor.

(Refer Slide Time: 18:26)

Now, one can say that steel is experiencing an increase in strength under rapidly applied
loads. Steel cannot respond at the same rate at which the load is applied. There is a delay.

280
Therefore, yield strength increases with less plastic deformation that is the result. At a fast
strain rate, a greater load is required to produce the same deformation than that at the lower
rate.

I mean this is understood right the rate of loading is a very important parameter in stress -
strain curve.

(Refer Slide Time: 20:39)

So, for steel of structural grade the modulus of elasticity, let us recollect what we understand
by this; it is actually a slope of the stress - strain curve in the elastic region. So, the modulus
of elasticity is same in the elastic region and forms a yield plateau for or under static and
dynamic response.

In the strain hardening region, the slope of the stress - strain curve is different under static
and dynamic response but, let us see that this difference is not very significant. There is a
difference, but not very significant. But, there will be a difference, let us understand that.

281
(Refer Slide Time: 22:58)

Therefore the ratio of dynamic to static strength at ultimate load state is lesser than the
corresponding value at yield point. Let us try to see how does it look like using a figure. So,
this is my strain, this is my stress axis. Let us draw a vertical line in an enlarged manner. So,
approximately 0.002 and let us draw one more line which will be higher than 0.002, but
closely to 0.005.

' '
Let us draw horizontal axis as 0.5 fc dash. Let us say 𝑓𝑐 and 𝑓𝑑𝑐 let us say this mistress sound.

Let us draw the curve for rapid strain rate, it looks like this. It takes a slope and goes till here
and come backs here, let us say it goes here it takes a slope goes here and come backs here.
So, this is for rapid strain rate. Whereas, if you look at the ASTM standard strain rate there is
a change in slope, it touches this point and comes down. This is ASTM standard strain rate.

So, there is a change in slope; so, the change in modulus of elasticity. So, this curve shows
effect of strain rate on stress - strain curve for concrete, on the other hand let us see what
happens when you look at steel.

282
(Refer Slide Time: 26:32)

Let us draw a vertical line at approximately 0.001 to 0.02. Let us draw another line with large
ductility at 0.07 approximately to 0.23.

Let us say this is my 𝑓𝑦 and this is my 𝑓𝑑𝑦, and let us say this is my 𝑓𝑢 and this may 𝑓𝑑𝑢. I

will have control points, let us have this control points. Now, let us plot the stress - strain
curve under rapid strain rate and ASTM strain rate. Rapid strain rate and ASTM has the same
slope for steel. So, let us say it goes till 𝑓𝑢, it goes till it crosses 𝑓𝑑𝑢 takes a kink. It goes till

𝑓𝑑𝑢 takes a kink and develops a plateau and then, goes to 𝑓𝑑𝑢. This is for rapid strain rate.

When you draw this for ASTM strain rate, it stops at 𝑓𝑦, slope is same develops a yield

plateau, and goes till 𝑓𝑑𝑦 and comes. This is for ASTM strain rate. This is the curve which is

effect of strain rate on stress strain curve for steel according to UFC – 3 - 340 - 02.

283
(Refer Slide Time: 29:46)

One can very well notice that elongation at failure is relatively unaffected by the dynamic
response of the material.

The higher the static strength of the material, the lower the increase in dynamic strength;
thus, faster the material is strained, that is rate of strain is higher. The higher the increase in
the dynamic strength of the material; it also increases higher ultimate strength.

(Refer Slide Time: 31:34)

284
Therefore friends, magnitude of dynamic response also depends on static material strength
and very importantly STRAIN RATE.

Let us look into the next factor, which is very interesting for us.

(Refer Slide Time: 32:17)

Which is dynamic increase factor; now, we want to incorporate. So, one should incorporate
the effect of material strength increased with the strain rate, is it not. Just now we saw there is
an increase. So, therefore, your dynamic increase factor is incorporated is a ratio of dynamic
strength to static strength.

Now, this depends on various factors. 1 it depends on material type, it depends on strain rate,
and depends on type of stress state. There are different methods to compute this.

285
(Refer Slide Time: 33:53)

Let us see one by one quickly these methods. Let us say method 1, what we do is we know
that it is possible to determine the actual strain rate of a material during calculation of
dynamic response. This can be done using an iterative procedure.

What do we do this is - a rate of strain a strain rate is assumed and an appropriate DIF is
selected. Now, the dynamic strength is determined by multiplying the static strength with
DIF. We already know the equation see here, if we know the dynamic static strength. We can
get the dynamic strength if we know the DIF. So, I get this ok then, the time required to reach
the maximum response can be used to determine a new strain rate.

Therefore, DIF will be revised, because strain rate is changed for a given or a known assumed
strain rate, you select a DIF; when the strain rate changes DIF is also revised. So, once DIF is
revised you again obtain the dynamic strength. Now, try to find out the time taken for the
maximum response to reach.

286
(Refer Slide Time: 36:49)

So, this process becomes iterative. So, this procedure is iterative until, the computed strain
rate which is relatively new matches with the assumed strain rate at the previous step. So,
there is one method by which I can compute this. There is another method by which we can
do the dynamic increase factor. So, this is suggested as per UFC – 3 – 840 – 02 this code
suggests choosing dynamic increase factor, based on pressure range or scaled distance of the
explosion the method group’s blast loads.

(Refer Slide Time: 38:42)

287
This method groups blast loads, as low - pressure and high - pressure categories and suggests
appropriate DIF value for each category. DIF values also vary for different stress states.

Further, the strain rate for tension and compression are lower than that for flexure. Therefore,
DIF values for tension and compression are lesser than DIF values for flexure. So, there is an
advice given by the code,

(Refer Slide Time: 40:38)

Depending upon the stress the DIF value suggested as per ASTM A 54 steel a sorry, A514
steel, it says for tension and compression. It suggests 1.05 and for bending that is flexure, it
suggest 1.07 to 1.09. So, that is about 10% increase in the dynamic strength.

288
(Refer Slide Time: 41:15)

Having seen this, let us go to the dynamic design stress. So, the dynamic design stress
requires to model the strain hardening effect of the material. Strain hardening effect in steel is
modelled using classical single degree of freedom systems by using a design stress greater
than the yield.

During the dynamic response the stress level of critical sections vary with strain of the
section, we all know that. So, in the elastic region, strain occurs at a section at any section
varies with its location from the neutral axis, we all know that. Beyond this region, the
member experiences plastic response, because when the strain or the stress reaches yield
value then, the member will experience plastic response. So, at this point, the stress is
constant that is at σ𝑦 over the entire cross section.

So, the steel member experiences increase in stress.

289
(Refer Slide Time: 44:20)

In the strain - hardening region until the ultimate stress is reached. So, after this point, the
fibre stress decreases with increasing strain and rupture occurs we all know that. So, there
will be a difficulty in modelling using this scenario in single degree freedom system.

So, to predict the dynamic response, it would be necessary to continuously vary the material
stress with deformation. This variation is difficult to model in single degree of freedom
systems, because this requires a complex resistance deflection curve at each point of time to
be updated.

So, how to model this accurately. So, there is a difficulty to model this behavior using SDOF
systems. So, how to then model this how to solve this difficulty.

290
(Refer Slide Time: 46:12)

So, what do we do is it is desirable to represent the design stress of the material as a bi-linear
stress strain curve.

So, we will express this as a bi-linear stress strain curve which will be the initial slope is k e
and this becomes my strain at y e and this becomes my R u. So, this is my resistance sorry,
this is my resistance; this is my yield, deformation or deflection. So, it is a bi-linear resistance
deflection curve as you see here. So, it is desirable to represent the design stress using this
curve.

In this curve we can see that stress increases linearly with strain up to yield, and after that it
becomes constant. This produces a simple bi-linear resistance deflection curve, which can be
included to model the strain hardening effect.

291
(Refer Slide Time: 48:32)

So, therefore, the design stress need to be computed from this. So, at low response ranges, the
maximum design stress is equal to the dynamic design stress or dynamic yield stress.

At higher response ranges, the design stress is increased to account for strain hardening in the
initial portion of the response this increased design stress will result in over prediction. This
increase will result in over prediction, but at greater deformations at large deflections the
stress level and the corresponding strain level will be under predicted by the design stress.

(Refer Slide Time: 50:49)

292
So friends, in this lecture we learnt, the design related material properties. We have taken the
example as steel as a material. We discussed about its application for form resistant design;
we also touched upon the loads like blast loads. So, I think this content will help you to know
more about the variations in material characteristics.

Especially, steel when you look for its application which demands large deformations from
the material side, because structural system undergoes large displacement, because of its
either form dominance or because of the rate of strain of the load coming larger on the
material which is a practice, in case of blast loads and impact loads.

Thank you very much, have a good day.

293
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 17
Plastic design- 1

(Refer Slide Time: 00:20)

Welcome to Lecture 17, where we are going to talk about Plastic Design. This is actually a
part of 2nd Module of Advanced Steel Design course which is under NPTEL IIT, Madras.
Before we enter into discussing the plastic design fundamentals and its applications and
theories related to the conceptual development of plastic design, let us try to ask a question,
what will be the general design considerations in designing a steel structure? Let us ask this
question.

294
(Refer Slide Time: 01:13)

So, let us talk about design considerations for steel structures, when I say steel structures, I
am not talking about conventional structures I am talking about strategic and special form of
structures where the structures resist majority of the load by the geometric form. I am talking
I am keeping this in mind because we are always talking about advanced steel design
procedures. So, the conventional design of steel structures is not discussed under the purview
of this course at all.

So, for example, for our convenience let us take the design of topside of an offshore platform.
Offshore platform is a strategic structure, if it happens to be a compliance structure it will
have a form dominance. Let us talk about the design consideration which you have for
topside structures. So, the first condition should be that the top side members should be
designed to have sufficient reserve strength.

Why it is required? The reserve strength is required to withstand the imposed load during the
pre-service life and in service life. So, that becomes a first requirement. The second
requirement could be, the platform or the structure shall posses, optimum safety, reliability,
cost effectiveness and flexibility that is very important we are talking about form dominance.

295
(Refer Slide Time: 04:20)

The third condition could be the design shall comply with necessary serviceability criteria
that is very important as stipulated by the design codes. The fourth could be it should comply
anyway with the design standards as imposed by the codes. So, therefore, the design
optimization is in terms of member sizing, joint detailing and to produce overall economic
design about the materials being used and ease to fabrication.

So, these are the factors based on which a structural design of an offshore platform is
generally optimized. Now let us talk about what would be the general design criteria or
design acceptance criteria.

296
(Refer Slide Time: 06:28)

Let us talk about the design acceptance criteria. Now when you talk about design acceptance
criteria, we have to refer to the codes of practice, let us say with reference to AISC and API
RP codes. There are some basic criteria stipulated by these codes. Let us say the allowable
stress shall be based on the code provisions. If you permit any increase in the basic allowable
stress, it may be due to many reasons. It can be due to strain rate increase, it can be due to
dynamic increase factor, etcetera.

If you permit that in the design it should be notified that is very important you must have to
specify this very clearly in the design. That is one of the requirements of the design.

297
(Refer Slide Time: 08:21)

Furthermore, the second condition of course, comes is the slenderness ratio. The slenderness
𝑘
ratio 𝑟
should follow the code accepted values for different members. We all know this is the

standard procedure we are not discussing that here; however, you must note that the
slenderness ratio for compression members in form dominant designs, cannot be more than
120 is a limit.

The third conditions come from the sizing, which is diameter to thickness ratio. The
𝐷
minimum diameter to thickness ratio for tubular members is 20 the maximum 𝑡
ratio for

tubular members is 60; however, under exceptional circumstances we can increase this.

298
(Refer Slide Time: 09:58)

The next condition the design acceptance criteria is thickness limit of the member. There is a
limit imposed on the thickness of the member. The primary members the beams that is
flexural members cannot have a thickness lesser than 6 mm. If members if the member
thickness exceeds 65 mm, then its a special requirement.

So, its upper limit we have. So, anywhere from 6 to 50 mm is the recommended thickness of
the members. Furthermore, the chord thickness of the joints because we are talking about
tubular members the chord thickness of the joints can be increased up to 65 millimeters or
𝐷
here 𝑡
ratio of 20 if you have ring stiffness.

299
(Refer Slide Time: 11:52)

Then next condition comes as allowable deflection. Is one of the important serviceability
criteria. The deflection should be limited to 1 by 200 or to be very specific of the span. The
next condition is the fatigue life of the joint. The minimum fatigue life for the joints both
accessible, inaccessible or different. For accessible joints the minimum fatigue life is 60 years
for inaccessible joints the minimum fatigue life should be 120 years.

See these are some thumb rules, these are some basic guidelines, design governance which
international codes commonly follow when we talk about large displacement structures in
marine environment where the environment influences the material degradation. And of
course, it also causes special kinds of loads on structural systems where steel is considered to
be one of the most promising material for such kind of construction.

300
(Refer Slide Time: 14:00)

Having said this, let us now get into the design methods, before we enter into plastic design.
We all know that analysis and design is a closed loop. The structural design generally is
carried out to control the response of the structure under the combination of various loads.
Analysis reflects the response behavior of a designed structure. Therefore, friends it is a
common understanding the analysis is carried out to check the adequacy of design.

Structural design is more based on experience or understanding of the structural behavior

under combination of various loads. But analysis is a sophisticated process, which is
computer aided to solve high end mathematical equations with the help of various numerical
tools. The results of analysis; however, need to be checked with the designer to correlate this
with the actual behavior of the structure.

When we talk about strategic structures structural design is more focused on arrangement of
functional utilities to obtain a load balance geometry. If you talk about design strategy of
form dominant members, then we can say it is more focused on arrangement of functional
utilities and the members. So, the objective is to obtain load balanced geometry, that is the
objective. So, the design should not be seen as sizing of the structural members.

301
(Refer Slide Time: 17:04)

So, let us not do not look design of form dominant systems, as member sizing to encounter or
to sustain the encountered loads. It is not simply the member sizing, its a large picture in
reality. The focus should be more towards the overall shape and size and geometric layout of
the structure. So, the focus is towards the overall shape, size, layout of all utilities that locate
or that form the part of the structure. So, if you talk about the top side of an offshore
platform.

(Refer Slide Time: 18:49)

302
Then it is layout of various utilities, let us quickly see what are various utilities of an top
shore platform drilling, processing, electrical, mechanical equipment, pipes, cranes and other
lifting facilities, storage, housing or what we called as accommodation units, rescue units,
helipad or helideck, etcetera.

So, various utilities are first arranged based on that a load balance geometries obtain that is
actually the design it is not member sizing friends. So, we are looking design as much larger
perspective. So, the main objective of the design therefore, is to minimize weight and
maximize its functional ability. Therefore, friends the functional design of a top side or any
strategic structure preludes the structural design.

So, ideal design philosophy maintains the equilibrium between the applied load and
resistance that can offer by the structural system. Unlike in conventional design practices
where resistance is accounted only from the strength of the material for strategic structures or
form dominant designs.

The resistance is also offered equally from the geometric form. So, even the geometric form
arrangement of members, their sizing, placement of heavy equipment’s both vertical and
horizontal zoning, spacing of critical operations such as lifting loading, unloading, drilling
plays a very important role in balancing the load acting on the structure. That is very
important.

(Refer Slide Time: 21:31)

303
So, various functional activities their layout and their weight, etcetera also help in the design
of form dominant structures. Example top side of an offshore platform. The very examples
we have seen in different kinds. Therefore, friends a safe design is termed as resistance
exceeding the load capacity. Resistance exceeds the load demand.

As a resistance can also arise from geometric form, form dominance is carefully exploited to
maintain the dynamic equilibrium between the resistance and the loop. Therefore, it is
important to ensure that the large displacement structures like, form dominant systems
maintain geometric stability while they are under activation of loads.

Because they undergo large deformation these large deformation can get into a plastic
deformation of the material. So, there should be one to one correspondence and support from
the material to the structure and vice versa so, that the encountered loads are counteracted by
the capacity of the form and of the material string together.

So, please note that the deformations being discussed here are not at the material level we are
also including them at the structural level. So, now, there are different design methods which
we all know. Like working a stress design, ultimate load design, limit state design, etcetera.
Let us quickly look at an overall view of this before we get into the plastic design. So, in
working stress design method.

(Refer Slide Time: 24:02)

304
So, let us say design methods a comparison. This is only just to review what we already know
most of you know this, but still for the completion sake. Let us review this quickly without
spending much time and get into the plastic design. In working stress design, the demand at
the working load is made to be lesser than the capacity of the material or capacity of the
structure, let us put it like this.

Now, the capacity of the structure comes from two, one comes from the geometry which is
static indeterminacy redundancy we call, the other comes from the material. The one which
comes from the material is stated as allowable stress limit. Usually, this value is kept within
the elastic limit of the material.

So, we say σ𝑎𝑙𝑙𝑜𝑤𝑎𝑏𝑙𝑒 usually is lesser than the elastic limit or stress at elastic limit of the

material. That is the usual practice what we do in working stress design. Therefore, friends
the ratio of capacity the material which is taken at the σ𝑦𝑖𝑒𝑙𝑑 to the σ𝑎𝑙𝑙𝑜𝑤𝑎𝑏𝑙𝑒 is called as a

factor of safety. So, this has got something called factor of safety which is a ratio of
σ𝑎𝑙𝑙𝑜𝑤𝑎𝑏𝑙𝑒
𝐹𝑂𝑆 = σ𝑦𝑖𝑒𝑙𝑑
.

We do not take it till yield, we will limit it below yield and that is the factor of safety. So, the
margin between the allowable stress and the working load is actually termed as a safety
margin, right. So, there is an explicit safety margin in the design like this. Hence in this
design method there are 2 tire safety margins.

305
(Refer Slide Time: 26:53)

So, there are 2 tire safety margin. What are they? The first one is limiting by limiting the
stress to stress allowable which is lesser than yield. So, there is a reserve capacity in terms of
the material. The other one is the actual capacity of material. Now the material also has
reserve energy. So, these two together has to safety margin.

So, one should carefully note, that still the reserve capacity of the material is beyond its yield
which is completely ignored in the design procedure. Let us put this mark here. The reserve
capacity of the material beyond yield is completely ignored in this design position correct.
The second one, which is very common is ultimate load design.

306
(Refer Slide Time: 28:26)

In ultimate load design case, the ultimate load is computed by enhancing the demand at
working load by a load factor. So, in this case what we do we enhance or increase the demand
at working load by a load factor we all know for example, the load factor of 1.5 is being
recommended. But still the material strength is limited only to its yield value.

Hence the safety margin which was present in the workings stress method remains
unchanged, is it not? The safety margin present in the earlier design method remains
unchanged, but the load has been increased from working load to ultimate load, right. So,
what we did is we have increased the load from working to ultimate using a load factor, but
the safety margin of the material is not changed. Therefore, the factor of safety is now
replaced by load factor, is it not?

307
(Refer Slide Time: 30:38)

However, you must understand that the reserve capacity of the material beyond yield is still
unexploited even in this method, correct. In limit state design procedure, the ultimate capacity
of the material is also reduced by a partial safety factor. So, what we do is we reduce the
ultimate capacity of the material using a partial safety factor for materials, which is generally
indicated as γ𝑚. On the contrast; on the contrast the demand at the design load is enhanced.

We call this as γ𝑓. Using a partial safety factor for loads which we discussed in detail. So, the

material capacity is restricted only up to yield. As in other methods therefore, let us quickly
see this as a good graphical comparison on the screen.

308
(Refer Slide Time: 33:00)

We will say that, let us say its may 0 level and we say this is my working stress method. So,
what we do from here is, this is working stress value, working stress value. So, let us say we
try to increase or in working stress design.

(Refer Slide Time: 34:04)

This becomes my load criteria or the strength criteria. Whereas, in limit state design the load
at working load is enhanced, is enhanced by γ𝑙. But the material is applied with γ𝑚. So, this

becomes my parameter for working stress design and this becomes my parameter for limit

309
state design. If we talk about ultimate load design what I do is I enhance the material
characteristic further and take it till the top.

So, in all the three cases we can very well see the safety factor is inherited in the material
level itself. However, the reserve energy possessed by steel is relatively higher.

(Refer Slide Time: 35:37)

And that is not used is not used, is it not? And it is relatively higher in steel and that should
be accounted in the design, is it no? So, if you look at the reserve energy before, we do let us
make a comparison the working stress design does not permit or does not invoke even any
damage.

So, 0 damage condition. The limit state design invokes nominal damage, but meets
serviceability requirements. The ultimate load design invokes damage, but no collapse. So,
safety of the geometry is ensured. So, if you look at the reserve energy possessed by
structural steel its very amazing. Let us see this let us try to plot the stress strain curve of
structural steel.

310
(Refer Slide Time: 37:28)

So, the initial slope let it be same for all the types of reserve energies. So, initially this is for
mild steel. I am just drawing qualitatively. Now I will say this is going up to 0.25. If you look
at the high strength steel. This is the mild steel, this is the energy. We have, when you look at
high strength steel this is low carbon steel.

And the limitation stops at about 0.2 and this becomes there is the energy. When we talk
about special grade of steel which is heat treated this is furthermore, but stops at about this is
point this is 0.15 this is 0.2 this is 0.1. So, this is heat treated construction alloy steel which is
quenched and tempered alloy steel, this reserve energy is still high.

311
(Refer Slide Time: 39:55)

So, friends one can very well see the design methods can be now divided into 2 major
domains can be divided into 2 major domains; one is force controlled other is displacement
controlled. In force-controlled design method failure occurs if the imposed load exceeds yield
strength. So, in force-controlled method failure occurs if the load exceeds yield strength, but
in displacement control method no failure occurs even if the load exceeds yield displacement.

Please carefully understand this. Provided the material is ductile; however, damage occurs in
excessive yielding and will cause permanent deformation this is termed as plastic
deformation. So, let us say what would be the concept of plastic design now based on this.

312
(Refer Slide Time: 42:32)

So, the concept of plastic design is induced force can be more than yield strength it will not
cause damage but causes plastic deformation which is a permanent damage, please
understand that. So, in plastic design you are enabling you are permitting a permanent
damage, but; however, the force level will be much more than the yield strength of the
material. Therefore, friends material to be used in plastic design should possess high post
elastic deformability, steel has this property. So, let us look into this graphically. If we plot
this as deformation and load curve.

So, we can have a nice curve and like this. Let us say at 1 we can have a load level which is
load level 1 I call this 𝑃1 where no permanent deformation is caused. It is at this point which

level; I will call this as level 1 level 2 which has reached yield the load can be still more than
that, see here the load can be still more than that. But during this increase in load, you also
encounter large permanent damage.

And of course, it collapse, it will undergo a permanent damage or a failure. Now if you look
at the elastoplastic deformation of steel, I will draw it here itself is easy. Elastoplastic
deformation of steel, steel has got ideally a good yield plateau and then it goes. So, this is the
yield plateau. The strain and stress of course. So, this is now going to be idealized in my
plastic design as a bilinear curve where I am going to stop this at σ𝑦, I am going to and this is

my ε𝑦.

313
So, this is going to be my idealized curve which I am going to use for plastic design this is for
steel. So, steel has got good ductility, elastoplastic idealization which can be applicable to
steel. And still shows a very good yield plateau which suffices the design requirement of
increase from 𝑃2 to 𝑃3 with permission of permanent damage as envisaged here.

Of course, we are not entering into the region at all. This region is not entered. We are
limiting till here. So, if I say this is my XX line the XX line is here. I am limiting till here,
friends.

(Refer Slide Time: 47:37)

In this lecture, let us look at the summary what we learned. We learnt various design methods
quickly, we compared them and we understood something called factor of safety, γ𝑚 and γ𝑓

for load and material. And we realize that how the methods of design exploits the steel
capacity reserve energy up to a level. Then we moved on to the plastic design we states that
the design process controls the stress at yield, but allows increase in imposed load at a cost of
plastic deformation which is a permanent damage in the member.

So, if you want to allow this design to be incorporated you have to have the material
capability of ductility induced in the material plus when the structures are undergoing large
displacements because of the form dominance, it should be supported equally and parallelly
by the material therefore, the material should be also remaining elastic or elastoplastic
conditions should be available in the material with enact ductility.

314
So, that the one-to-one relationship between the demand from the structure, expectation from
the material matches each other to contract the load acting on the structure without any
failure. In all the time remember friends, the geometric stability of the structure even under
accidental worst scenario should be guaranteed.

There is no design procedure which is acceptable by any course of practice in the world
which says that the geometric stability can be over toppled at the cost of increase in load
carrying capacity of the member or the material usage effectiveness, friends. So, we will
close it here for this lecture, we will continue in the next lecture.

Thank you, have a good day, bye.

315
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 18
Plastic design -2

(Refer Slide Time: 00:18)

Friends, welcome to the 18th lecture at Advanced Steel Design which is on continuation
lecture to plastic design. We call this Plastic design 2, because in the last lecture we started
with the plastic analysis design. So, friends the plastic analysis has certain basic conditions
permits no shear strain and no warping.

Why? Because, there is a very strong assumption we deal with this which controls this. The
very strong assumption is plane sections remain plane and normal to the longitudinal axis of
the member even after bending. So, very interesting and very vital assumption we make in
plastic analysis which can be expressed diagrammatically like you see here.

316
(Refer Slide Time: 02:11)

Let us say I have a beam which is supported one end we just free the other end. It is a
cantilever beam. Let us assume that the beam is subjected to a tip load, some magnitude. As
we all know the beam will deflect. Let us say a reflected position of the beam right. So, let us
say the beam initially had different segments equally cut along its length, different segments.

Now, each segment will now also get transformed, is it not in the new position; am I right?
So, we clearly say that the segments in the old and new are same, they do not get disturbed.
So, there is no warping which is happening. So, only the horizontal axis is get twisted that is
all. So, the extreme fibers in the top will be under tension and the extreme fabric at the
bottom will be under compression, I think we all know that. And, somewhere there is a layer
which is the neutral layer which neither has tensile nor compression fiber.

So, that is a very interesting assumption that the plastic analysis does not permit any shear
strain or warping on the cross sections. It emphasizes that each fiber in the cross section is
free to deform. There is no restraint on them independently. So, that successive fibers in the
cross section can yield one after another successively upon increase of load. As you keep on
increasing the load, successive sections keep on yielding independently; each fiber is been
given this freedom to deform independently.

That is a great advantage, that is a great boon given to every section, every fiber in the cross
section, in the cross section. Every fiber in the cross section is permitted to do anything they
want freely, they are independent. They can deform independently; they have a full freedom.

317
And, this freedom enables successive fiber in the cross section to keep on yielding one after
another when you gradually increase the load; that is an assumption we make and that is how
it is implemented. Having said this let us go slightly one step ahead and discuss more in detail
about this.

(Refer Slide Time: 06:45)

So, friends according to plastic design philosophy, the capacity of the member; let us say very
clearly load capacity of the member is geometry dependent. It is not load dependent. I will
later on prove that the moment capacity of a plastic section is independent of load. I will
prove that the moment capacity of a plastic section will be independent of load frames is very
interesting you know.

For example: if you really wanted to find a moment capacity of any section, let us say simply
2
𝑤𝑙
supported beam you want to find moment, you will say it is 8
where moment is a function
2 2
𝑤𝑙 𝑤𝑙
of load. You want to find a fixed beam any section, you say 12
or 24
. So, the moment is

again a function of w is it not, but I will prove in the derivation later in this class that the
moment capacity of a plastic section will not depend on. It is purely a geometry dependent
parameter.

So, can I say that we are approaching towards form dominance? So, plastic design is in
existence for the past 25-30 years in analysis and design friends. So, as I said in the beginning
form dominance is not a recent philosophy. It is well established, well practiced in

318
engineering application, in structural engineering problems both in classrooms codal
provisions and in fact, industrial practice very well versed in the application. So, here is a
standing example where the moment capacity of the section is not your load dependent
criteria.

It is a geometric property, it depends purely on the cross-section property of the system and
of course, the material. It is purely not geometric dependent; it has got a component coming
from; the strength has to come from the material friends. You cannot ignore the material
strength anyway. For example, you have a shell, you have an arch which are all form
dominant; can it constructed with the paper to carry the load? You cannot say sir material is
going to be paper which is having no strength, but can the form alone this is the strength.

The answer is big no, we all know that engineering is not possible. So, material strength is
not ignored, but that is not the only criteria that is the emphasis in form dominance.
Geometric dominance also plays a vital role in sharing, distributing, using or dispersing the
load very economically, that is the point we are discussing here.

(Refer Slide Time: 10:38)

So, friends as seen in elastic design, the capacity you will say the procedure restricts or limits
the capacity of the section or the member or the material whatever you want to say up to first
yield is it not. So, we know look at the stress strain curve of classical structural steel which
we plotted n number of times after the first yield steel undergoes local yielding. This happens
because of the residual stresses. Now, we all know where these residual stresses come from.

319
They come from fabrication process, they come from preheating of the member. There are
many reasons, we discussed them in the previous lectures. Therefore, friends it is not a good
design practice to limit or to restrain the material capacity at the first yield is it not. So, I
should say here a design procedure which limits the capacity up to first yield strength of the
material is not a good design practice.

Why? There is a reason for this. Beyond the first yield it has got a lot of ductility, lot of
energy, lot of reserve energy present in the material. So, what the design procedures used to
do? The design procedures used to restrict the strength up to yield and used the redundancy of
the geometry as reserve energy.

(Refer Slide Time: 14:04)

You know static indeterminacy is a redundant strength in the present in the structural system.
This is purely developed from the geometry of the system. So, it is capable of offering great
internal resistance capacity. This is also a form, it is also a type of form dominance. But, we
are not talking about this, we went one step beyond this. We allow the structure to undergo
large displacement and by the virtue of large displacements, they interact or counteract the
loads. So, this is the form dominance what we expect which is happening in special kind of
structures like TLP, cable state bridges etcetera.

Offshore structures are most prominently formed dominant which undergo large displacement
and the relative displacement or relative force helps the structure to counteract the platform.
There is a big beautiful idea conceptually implemented in the design procedures of offshore

320
structures. So, I think it is very important for us to learn this. Coming back to that we
understood that the static degree of indeterminacy is capable of offering a good internal
resistance capacity.

(Refer Slide Time: 16:40)

Having said this the material reserve which is present beyond yield in the form of ductility
can also add additional resistance is not, its capable. So, but in elastic design this was
neglected. So, the material remain under use, its not used properly. So, now, this idea is
violated is. So, in plastic design apart from using the geometry, material reserve capacity
beyond first yield is also utilized.

321
(Refer Slide Time: 18:16)

And, let me also put a mark here. First yield in many σ − ε curve is a falsify indication of
yielding. This occurs primarily due to presence of residual stresses. So, therefore, first yield
cannot be controlled or cannot be allowed to govern the design, that is the first argument we
have. Secondly, steel as a material has significant reserve energy beyond first yield and that
should be utilized. So, that is the focus of plastic design. The focus of plastic design is that.

(Refer Slide Time: 20:24)

So, now let us extend this argument further. Considering the standard σ − ε curve of
structural steel, let us plot it here; we all know that. Let us consider the standard σ − ε curve

322
of structural steel. One can understand that when material fails under load following
observations are also important which are the failure could be also due to instability, can also
due to fatigue or it can be due to excessive deformation.

So, if any of these types of above factors does not initiate the failure then the member can
carry more load beyond the elastic limit, is it not. Therefore, friends in plastic design the gain
is the increase in load capacity. The loss is permanent which we call as plastic deformation.
So, there is a gain, there is a loss. It is also interesting note that plastic design misnamed even
though the stress is up to σ𝑦, please note that.

(Refer Slide Time: 23:18)

In plastic design the stress value is limited to σ𝑦. If you ask me a specific question what

would be the σ𝑦, upper or lower yield? Because, many steel many material does not show

both the yield points. So, we are talking about an idealized curve where. So, we are not
exceeding beyond this. There is an idealized curve, the red one is an idealized curve and the
green one is the actual.

So, the stress value is limited only to σ𝑦, then why we call this a plastic design? Because, we

still call we still name this procedure as plastic design because we permit plastic deformation.
So, the term plastic does not refer to load, it refers to deformation. To be very specific it does
not refer Y axis, it refers X axis. So, in plastic design the yield of the cross section is assumed
to be idealistic as we are going to see now.

323
(Refer Slide Time: 26:20)

So, let us try to take a beam. Before we get into the derivation of this, let us try to have a
serious note furthermore on plastic behavior of structures. We all know friends that ultimate
limit state deals with the collapse modes which will convert the structure into a mechanism.
Now, the formation of mechanism enables excessive yielding which finally, leads to collapse.
We all know friends that steel is a very common construction material which has got enough
ductility. Looking at the σ − ε curve of steel, it is evident that the material has got good
capacity beyond the first yield.

Therefore, any design procedure which limits the strength till the first yield like an elastic
design is not a good design procedure. So, the material strength beyond yield which is having
enough reserve value is not utilized. So, there should be a design procedure which utilizes
this excessive reserve strength beyond yield value, is it not. The load carrying capacity will
therefore, can be increased because of the strength.

But however, this does not mean that the structural capacity should be extended beyond a
value that results in a catastrophic failure. Therefore, σ𝑦𝑖𝑒𝑙𝑑 of a material should not be a

designing factor to assess the failure of the structure; because failure cannot be assessed
purely based on the strength. We should also look into deformation characteristics.

324
(Refer Slide Time: 29:59)

So, an important factor which will govern the behavior of material is as we know is a residual
stress. So, presence of this will have a false yielding point, because of imperfections in the
material which is developed during fabrication. So, let us recollect that residual stresses are
also responsible for the following changes in the material behavior. What are those changes?
Brittle fracture failure, stress corrosion, buckling strength reduction. We already said that, but
we are reemphasizing for us to understand more.

Last one is inelastic behavior in metals or let us say materials we are talking about steel
therefore, does not wrong in saying metals. But you know friends residual stresses
development cannot be avoided. It is an inherent part of manufacturing a fabrication process,
because when you start building the connections there is always a high probability that will
land up in development of residual stresses. So, this will now start prompting a long σ𝑦𝑖𝑒𝑙𝑑

which will govern the design.

So, elastic design of structures controls the design till this point. Therefore, they are not
considered to be a good method of design. Therefore, friends designer shifted the design
procedure from elastic to ultimate load design. The most advantageous factor of steel as we
all know is ductility which is a sign to use the reserve strength more conveniently and
confidently.

325
So, the ductility of the material enables a thorough redistribution of moments from highest
strength region to the next highest stress region. Thus, the load carrying capacity of the
member is kept on increasing beyond the σ𝑦𝑖𝑒𝑙𝑑 which is enabled in the other methods of

design like plastic design which is completely different and deviated from elastic design.

(Refer Slide Time: 33:05)

Therefore, friends plastic design is an effective design procedure. Let us quickly compare the
plastic design with ultimate load design. The ultimate load design uses the material ability to
the maximum positive point in design perspective whereas, in plastic design the stress is
limited to yield value, but enables plastic deformation. So, one method uses the material
ability the maximum strength, other method uses the material ability to the maximum
deformation.

So, simply look at this curve. One method uses it to the maximum strength, other method
uses it to the maximum deformation. So, I mean we are looking for a different perspective of
conservatism. So, now, I want to design the plastic moment carrying capacity which we call
as a plastic design you know of sections. How do we do that?

326
(Refer Slide Time: 35:27)

Let us imagine that I have a beam subjected to moments at the ends. So, I have a beam, I
subject this beam to a moment at the end. The beam will bend. There will be a neutral layer
which will have neither tension nor compression. Above the neutral layer things will be under
compression, below the neutral layer things will be under tension. So, we call this as neutral
axis. This is the bending action of the beam. Let us take an arbitrary cross section.

In this cross section, let us mark the extreme fibers. So, when the member is fully elastic, we
all know that this fiber will be in compression and this fiber will be in tension. You see from
here this is tension, this is compression. So, when the member is fully elastic. So, the stress
value in compression and stress value in tension that is also mark the neutral axis and by here
the neutral axis.

So, this becomes my here the stress is less than σ𝑦, here also the stress is less than σ𝑦. And,

this is tensile and this is compressor. So, this is purely elastic case. Now, I apply more
moment, I increase the moment now. There is a stage possible where the stress will become
equal to σ𝑦.

I say the stress is now equal to σ𝑦. So, I call this as strain hardening ray stage. Now, as I said

independent fibers can elongate or contrast with the full freedom available. So, what happens
is apart from this fiber extreme becoming σ𝑦, we will have additional fibers also becoming σ𝑦

327
. We will have additional fibers also becoming σ𝑦. So, the stress distribution curve. So, there

is a small region in tension and compression which has become σ𝑦. So, I can call this as

elastoplastic condition. Why elastoplastic condition? Some part is still elastic, because the
stress is less than σ𝑦. Some part is plastic, because σ = σ𝑦.

As I said please recollect, we are restricting the stress only up to yield. We are only using the
X axis capacity of the material, not the Y axis capacity to the material to be very precise
mathematically. Now, if you still further increase the moment applied on the beam, a stage
will come where the stress diagram looks like this. So, this is C, this is tensile, this is
compressive. But, one can notice here that the interface of tension and compression force
does not happen at neutral axis.

Now, it is a new axis being formed here. We call this as equal area axis which means that
above this line, the compressive force is equal to tensile force. So, I can call this section as
fully plastic. How can I say it is fully plastic? Each and every fiber equal to σ𝑦 whereas, in

this case only few fibers where equal to σ𝑦. Each and every fiber in this case equal to σ𝑦

equal area axis.

So, this sketch expresses how a plastic deformation occurs in a given system. The stress
variation is shown like this. Now, I extract only this figure and this figure, from here. Let me
do that.

328
(Refer Slide Time: 42:36)

So, I am drawing an arbitrary cross section, taking the extreme fibers, drawing the equal area
axis. I mark this as equal area axis, then I only draw the compressive force and the tensile
force. So, this is my total tension force, total compressive force. So, this will have a centroid.
Let me say the centroid is here. This will have a centroid. Let us say the centroid is here, they
are opposite.

I call this centroid as y bar 1 from the equal area axis. I call this centroid as y bar 2 from the
equal area axis. And, we all know that at this stage the stress is σ𝑦 in the entire cross section.

So, we can call this as force equilibrium condition is the cross section is the stress diagram.
When the cross section was loaded and the stress distribution is happening by stages from
here to here, we had made some assumptions.

329
(Refer Slide Time: 44:08)

What are those assumptions? The valid assumptions could be material obeys Hooks law until
the stress level reaches the first yield point. On further straining stress remains constant
which is equal to σ right. The second assumption we make is the upper and lower yield points
if shown in the σ − ε curve both in tension and compression are same. Then material is
homogeneous, isotropic in both elastic and plastic states. The last assumption which is the
foremost important assumption.

The transverse plane section will remain plane and normal to the longitudinal axis of the
member even after bending.

330
(Refer Slide Time: 46:29)

So, friends in this lecture, we have started understanding the importance of plastic design
philosophy. We learnt the difference between plastic design and compare it with elastic
design with ultimate load design. We learn that plastic design enables plastic deformation,
that is why the name plastic has come. The freedom is in the X axis, it is not in Y axis.

So, in the next lecture, we will talk about how to work out the moment capacity of a plastic
section. And, we will show that the moment capacity of the plastic section is load
independent.

Thank you very much, have a good day.

331
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 19
Plastic design - 3

Friends, welcome to the 19th lecture on the course Advanced Steel Design. Here we are
going to continue with the Plastic design. I will call Plastic design 3, 3rd lecture. In the last
lecture we learnt the comparison between the plastic design philosophy with elastic design
and ultimate load design. We have learnt that in plastic design the material strength, which is
in terms of ductility or plastic deformation is utilized which is one of the basic requirements
in a form dominant structural system.

Apart from utilizing the reserve capacity of the geometry, in terms of structural
indeterminacy, the material reserves strength in terms of ductility is also being used in plastic
design effectively. So, we said that in plastic design we have got a new concept called equal
area axis where the total compressive force in a cross section, meets exactly in magnitude
with the total tensile force acting in the cross section.

So, each and every layer in the cross section is having freedom to reach the σ𝑦, because we

have assumed that each and every fiber is independent to elongate or contract. So, that the
stress value can be reached till σ𝑦, once the stretch σ𝑦 is reached then the next section of the

fiber is increase the stress of σ𝑦 and so. So, the whole cross section gets plasticized.

So, let us now derive the plastic moment capacity of any cross section. To start with we will
assume a rectangular cross section then we will go to arbitrary sections. So, to start with we
will do with the rectangular cross section.

332
(Refer Slide Time: 02:38)

So, let me draw a rectangular cross section; a solid cross section of dimensions b and d. So,
let us draw the stress diagram at elastic state and plastic state so, in the elastic state, because
only the extreme fiber is yielded. As usual this is my compressive force and tensile force.
This is my σ𝑦. In the fully plastic state we already know that the total tensile force will be

equal total compressive force and every fiber will be σ𝑦. So, we call this centroid as 𝑦1and

this centroid as 𝑦2. So, now, this axis is termed as equal area axis.

Because above this axis you see the total compressive force is equal to the total tensile force.
Let the material strength is taken as σ𝑦. So, we can easily find the section modulus I will just

rub this and write it here, because I need some space.

333
(Refer Slide Time: 05:19)

Let me write it here, I put this equal area axis. So, we can quickly find 𝑍𝑒, which is the elastic

section modulus which can be given by the moment of inertia about the Z axis by the distance
of extreme fiber.

3
𝑏𝑑 𝑑
So, for this cross section we can easily find out it is going to be 12
, divided by 2
which
2
𝑏𝑑
becomes 6
; let us call this equation number 1. If you want to find the elastic moment

capacity this is given by a simple expression this section modulus multiplied by the
corresponding stress we know that. Let us look at the plastic stress distribution. Let us look at
this is let us look at this figure 3.

So, now with reference to figure 3, let us find out the moment 𝑀 = (𝐶𝑦1) + (𝑇𝑦2), since C

is equal to T we can say this is 𝑀 = 2(𝐶𝑦1). Can I say that? So, what is the compressive

force, which is going to be 2 times of so, you know here the breadth of section is b is it not
see here. So, b and d by 2, because this is d by 2 am I right?

So, I am looking at this area now, b into d by 2 into. The centroid y bar will be d by 4, I think
that is simple geometry and I multiply this with stress, because this has got to be force into
2
𝑏𝑑
this stress. So, which gives me 𝑀 = 4
σ𝑦. So, if I say this equation number 3. So, I replace

334
this M as 𝑀𝑝 and I will term this as 𝑍𝑝 to σ𝑦, this is only to keep the similarity of this

equation with that of this.𝑀 = 𝑀𝑝 = 𝑍𝑝σ𝑦.

Instead of keeping some bd etcetera, I want to keep it like this. So, 𝑍𝑝 is called plastic section

modulus. 𝑀𝑝 is called plastic moment of resistance and of course σ𝑦 is the yield stress. You

can easily see there is a perfect compatibility between the equations 2 and 3a, both of them
indicate in the same manner and one refers to elastic section modulus, other refers to plastic
section modulus.

Therefore the corresponding moments are elastic moment of resistance and plastic moment of
resistance is it not. So, it is.

(Refer Slide Time: 09:29)

But we do not want to say please understand, we do not want to say do not state that
2
𝑏𝑑
𝑍𝑝 = 4
. We will not do this. We will not do this; we will not remember 𝑍𝑝 in this manner.

Well, simply say plastic moment of resistance is plastic section modulus multiplied by σ𝑦.

We will come to this argument slightly later, how do you get this. Now friends, we can easily
find out the moment capacity. Let us take these 2 equations 2 and 3a. Let us consider
equations 2 and 3. So, equations 2 say it is 𝑀𝑒 = 𝑍𝑒 σ𝑦 and this is 𝑀𝑝= 𝑍𝑝 σ𝑦. The ratio of

𝑍𝑝 to 𝑍𝑒, the ratio of 𝑍𝑝 to 𝑍𝑒 is called shape factor. On the other hand, I would say 𝑍𝑝 is

335
 2
𝑏𝑑
shape factor multiplied by 𝑍𝑒. That is why I said let us not remember 𝑍𝑝 = 4
of rectangular

section.

So, our job is to now find out the shape factor for various cross sections. So, for a given cross
section which is T L circle or any arbitrary, if I am able to find out the shape factor. I already
have the second moment of area divided by the 𝑦𝑚𝑎𝑥, I get 𝑍𝑒 I can easily find the moment

capacity. And also please note friends the moment capacity of plastic section is plastic section
modulus multiplied by σ𝑦.

2
𝑤𝑙
It is not a function of applied load, am I right? Usually 𝑀 = 12
etcetera whatever maybe. It

is not a function of load it is purely a geometric property of course, σ𝑦 is present. So, that is

why we always wanted to introduce a factor called shape factor. So, shape is related to
something of geometry. That is a reason why we want to introduce a term called shape factor.

So, for different cross sections, if I am able to find the shape factor derive them or is it
available in steel tables or any standard handbooks I can easily find the moment capacity,
because we all know in structural steel handbooks. The moment of inertia or second moment
of area of the cross section and depth of the section; therefore, 𝑍𝑒 section modulus is

available in the steel table is it not.

Please look at structural engineering handbook IS-SP 6 1. If you look at for different cross
sections of i c l angles, you will find 𝑍𝑒. That is elastic section modulus this is available

readily. If by any chance if the table also gives you a shape factor of that section, you also
have 𝑍𝑝 with you and if you have 𝑍𝑝 you can have 𝑀𝑝 with you. So, it is just a minute to

calculate the moment capacity of the section for plastic design.

336
(Refer Slide Time: 13:54)

So, let us try to understand a very important concept here. That the additional moment
capacity, from M elastic to M plastic is obtained purely as a geometric capability, provided
the material is ductile. So, the additional gain comes from the structural redundancy, which is
degree of static indeterminacy is one part of it form dominance is other part of it.

(Refer Slide Time: 15:32)

Now, friends let us also make a comment that; the maximum stressed section is identified as
the critical section. So, what do you mean by a critical section? This critical section cannot

337
carry any additional moment or let us say load beyond 𝑀𝑃. This is because at this critical

section the full cross section has already yielded. So, such sections are termed as plastic
hinges and the corresponding section is fully plasticized.

But please note at plastic hinges, the stress level is only σ𝑦, but still the term plastic is being

used, because the deformation at the section is in a plastic stage. Having said this let us try to
derive a shape factor.

(Refer Slide Time: 18:02)

Let us quickly see what is the shape factor? Shape factor is a factor that enables higher
moment carrying capacity, which is strongly geometry dependent. That is why it is called
shape factor.

Let us derive a shape factor for an arbitrary section. Let us derive a shape factor for an
arbitrary section, then we will apply this logic to all sections. We will take an arbitrary
section let us say extreme fiber we have a field of plastic stress distribution and we also mark
the equal area axis. This is C compressive force and this is tensile force. Let us say the
corresponding area here is 𝐴1.

And the corresponding area here is 𝐴2. So, now, at equal area axis, the section is divided into

equal halves, is it not? So, that is the total compressive force is equal to the total tensile force,
T. So, I can now say the total area A = 𝐴1 + 𝐴2. Can you say this as A by 2?

338
(Refer Slide Time: 21:19)

𝐴
That is we can also say 𝐴1 = 𝐴2 = 2
. Now, let us take moment about of C and T, about

equal area axis. See what happens.𝐶 = σ𝑦𝐴1; 𝑇 = σ𝑦𝐴2; σ𝑦𝐴1𝑦1 + σ𝑦𝐴2𝑦2 = 𝑀;

We should say C into y bar 1 plus T into y bar 2 should be the total moment. We also know C
𝐴
is σ𝑦 into 𝐴1
and T is σ𝑦 into 𝐴2. Let us substitute them let us substitute equation 3 in

equation 2. So, σ𝑦 𝐴1 y bar 1 + σ𝑦 𝐴2 y bar 2 is M. Let me carry it in the next page.

(Refer Slide Time: 22:43)

339
 σ𝑦𝐴1𝑦1 + σ𝑦𝐴2𝑦2 = 𝑀;

𝑀𝑝 = σ𝑦𝑍𝑝;So, can I say this as σ𝑦 into 𝑍𝑝, if I use this term as 𝑍𝑝 then this becomes my 𝑀𝑝

am I right. Now, interestingly friends the shape factor is 𝑍𝑝 by 𝑍𝑒, 𝑍𝑝 by 𝑍𝑒 is the shape

factor. So, now, we can try to find out the shape factor for different cross sections. Before we
do that let us try to find out what is an elastic core?

(Refer Slide Time: 24:19)

So, let us try to understand what is an elastic core and what its depth? So, for a plastic design
to be effective, the depth of elastic core at the critical section should be zero. Is it not?
Because at that section it is fully plasticized so, elastic core should be zero. Then only at
these, sections plastic hinges can form. Now, the number of plastic hinges, that can form
depends on the static degree of in determinacy. Let us call this as n. So, the number of plastic
hinges in a given structure should be equal to number of degree of static indeterminacy plus
1.

When so many hinges have been formed the structure will get or will get converted into a
mechanism. So, more the degree of static indeterminacy more the possibility of plastic
hinges, because we need n plus 1 hinges. So, it becomes a mechanism. It will also enhance
the moment carrying capacity of the structure apart from enhancing it is attained the shape
factor. In simple terms the moment carrying capacity is achieved by the plastic design in two
ways.

340
(Refer Slide Time: 27:06)

1. By having a higher shape factor, which is geometry dependent. 2. The structure should
have higher degree of static indeterminacy. Friends, you will recollect that both of these are
related to the structural form is it not. Therefore, one can say plastic design is a form
dominant concept. It utilizes strength till σ𝑦 of the material and it invokes the reserve energy

from the material in terms of ductility from the structure in terms of static degree of
indeterminacy and form compliancy. That is very interesting.

So, let us try to derive the elastic core and check, what is the condition for a plastic design.
So, to do an elastic core let us take again a rectangular cross section.

341
(Refer Slide Time: 29:22)

The cross section has the dimensions b and depth as d. Let us h, let us design a hybrid section
where the section is partially plastic and the remaining is elastic. Let us have an elastoplastic
section. So, let us have an elastoplastic section.

We call this elastic part as e, where e is called depth of elastic core. Of course, we know that
the stress at the extreme fiber remains σ𝑦. And we also know from this figure that this

dimension is h by 2 and this dimension of core is e by 2 am I right. Let us consider this cross
section and the stress distribution diagram, which is elastoplastic as shown in the figure right.

First let us find out the depth of elastic core. So, this is the elastic part and this is my plastic
part. Let us find out the moment of resistance of elastic part or elastic core at any cross
section. How do you get that? Let us say that is 𝑀1, which will be area of a triangle. So, half

base height is e by 2 and the cg of that is going to be 2 3 e by 2 am I right it is a triangle

know? I am taking moment about this point.

2 3 e by 2 and I have two such things 1 on the top and 1 on the bottom. And there is only the
force. So, I will multiply this with the stress to get my moment right. So, half base e by 2, 2 3
e by 2 and twice of that. So, let us cut the common values. So, can I get this as be square by
6, σ𝑦 be square by 6, σ𝑦 equation number 1. Now, let us work out the moment of resistance of

the plastic section or plastic core.

342
I call this as 𝑀2 which is the red one right. So, let us find out that. So, let us say it is a

rectangle. So, b into h by 2 minus e by 2 right and the cg of this from here will be equal to e
by 2 plus h by 2 minus e by 2 of half am I right, e by 2 plus h by 2 minus e by 2 half of that
correct. So, that is going to be the distance and I multiply this with the stress and there are
two such pieces one on the top and one on the bottom.

Let us simplify this. So, 2 h by 2 minus e by 2, e by 2 plus h by 2 minus e by 2 half of that

correct fine. So, let us simplify this please simplify and see what happens. So, it becomes 2
σ𝑦 b into h minus e by 2 of e by 4 plus h by 4, which will further become twice σ𝑦 b h minus

e by 2 h plus e by 4 am I right, which can be further 2 σ𝑦 b or let us remove this 2.

Because this goes away, a plus b into a minus b we can say a square minus b square by 4. So,
that becomes maybe equation 1, equation 2, equation 3, equation 4, equation 5. So, that
becomes my 𝑀2. So, now, friends the total moment carrying capacity of this elastoplastic

section will be sum of these.

(Refer Slide Time: 35:24)

So, the moment capacity of the elastoplastic section will be M=𝑀1 + 𝑀2. So, let us sum

these two what happens let us see. So, M is going to be σ𝑦 is anyway common in both cases

343
 2 2 2)
𝑏𝑒 𝑏(ℎ −𝑒
the first one was 6
, the second one was 4
, am I right, which on simplification will
2
2 2 σ𝑦𝑏ℎ 2
become σ𝑦 {
𝑏ℎ
−
𝑏𝑒
}, on simplification ⎡1 − 𝑒 ⎤.
4 12 4 ⎢ 2 ⎥
⎣ 3ℎ ⎦

Please check this let us call equation number this was 5. So, I call this as 6, this is 7 these as
8, this as 9. So, that is my total moment capacity of the elastoplastic section, which is
2
σ𝑦𝑏ℎ 2
⎡1 − 𝑒 ⎤.
4 ⎢ 2 ⎥
⎣ 3ℎ ⎦

(Refer Slide Time: 37:25)

We also know, we also know that 𝑀𝑝 is σ𝑦 into 𝑍𝑝 for rectangular section, 𝑍𝑝 is bh square 4.

We already said that. So, we can now write.

2
𝑒
We can rewrite the equation 9 as M equals this part I am replacing as 𝑀𝑝(1 − 2 ). That is
3ℎ

the equation number 10; where e is the depth of elastic core in the section. So, for the section
for a plastic section, e will be 0 for a complete elastic section, e will be equal to what? Depth
of the cross section is it not ok. So, one can easily find out this. So, we will also try to expand
this.

344
(Refer Slide Time: 39:17)

𝑀𝑝 σ𝑦𝑍𝑝
Let us say 𝑀𝑝 = σ𝑦. 𝑍𝑝. 𝑀𝑒= σ𝑦 Z elastic. So, let us say 𝑀𝑒
= σ𝑦𝑍𝑒
, which gives me that

shape factor. So, shape factor which is a geometric property is also the factor of additional
moment capacity of the plastic section beyond elastic section correct, because 𝑀𝑝 is shape

factor of 𝑀𝑒. So, shape factor is an additional capacity indicator of the fully plastic section

from the fully elastic section.

Please understand friends, at this stage stress is not exceeding yield, but still the term plastic
is used, because the deformation is plastic not the stress. That is what it is. Having said this
let us try to work out shape factor for a rectangular section.

345
(Refer Slide Time: 41:09)

The geometric parameter let us do it for rectangular section. Let us take this h. So, this is top
𝑏ℎ
area bottom area. We call this as 𝐴1= 𝐴2 = 2
. This is also h by 2 is it not? So, we know
𝑏ℎ ℎ
𝐴1= 𝐴2 = 2
. We also know y bar 1 is y bar 2, which is 4
. So, 𝑍𝑝 is actually the equation is

A by 2 of y bar 1 plus, y bar 2 is it not, that is a 𝑍𝑝 value. We had somewhere here this. So,

this 𝑍𝑝 is it not this is 𝑍𝑝. So, that is what you are trying to say here y bar 1 plus y bar 2

correct.

2
𝐴 ℎ ℎ 𝑏ℎ
Let us substitute that here which is going to be 2
(4 + 4
), which becomes 4
and
2
𝑏ℎ
𝑍𝑒 = 6
. So, shape factor as we know is 𝑍𝑝 by 𝑍𝑒 which is bh square by 4, bh square which

becomes 1.5. So, shape factor for rectangle is 1.5. So, the plastic capacity of rectangular
section is 50% more than elastic capacity of the same section.

So, friends in this lecture we discussed the importance of shape factor, the important
philosophy of plastic design, we also learned how to find out the depth of elastic core and we
understood how to arrive at the shape factor for different geometry is it not. I hope you will
follow these lectures and revise them regularly and you will apply for more examples look
into my advised textbooks recommended textbooks for this course acquire a copy learn them
there are many MATLAB programs available in this, which will help you to solve many
problems as exercises and for intensive learning.

346
Thank you very much have a good day.

347
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 20
Shape factor examples

(Refer Slide Time: 00:17)

Welcome to the 20th lecture on the course Advanced Steel Design. In this lecture we are
going to do more examples on Shape Factor. In the last lecture we already said that shape
𝑀𝑝 σ𝑦𝑍𝑝
factor is expressed as 𝑆 = 𝑀𝑒
= σ𝑦𝑍𝑒
and therefore, it is actually the ratio of section moduli

of plastic to elastic.

So, on this example we applied and worked out a rectangular section and we found for a
rectangular section we got the shape factor as S=1.5, we derived that.

348
(Refer Slide Time: 01:38)

Now, we will derive further more for different cross sections. Let us take a circular cross
section as you see on the screen the radius is r. So, let us say this is classical example b. A
solid circular cross section radius is r and we know the center of this particular part will be
4𝑟
𝑦1 = 𝑦2 = 3π
as you see here. The standard expression which we know.

2
π𝑟 𝐴
So, let us say the upper half is 𝐴1and lower half is 𝐴2. Let us say 𝐴1 = 2
= 2
. Because
4𝑟
you know this is actually equal area axis, correct. And if I say this as 𝑦1 = 𝑦2 = 3π
where r

is the radius of the solid cross section.

𝐴
So, we already have an expression for 𝑍𝑝 section modulus plastic which is 𝑍𝑝 = 2
(𝑦1 + 𝑦2)
2 2 3
. Let us substitute that here A is
π𝑟
2
. That is 𝑍𝑝 =
π𝑟
2 ( 4𝑟
3π
+
4𝑟
3π )= 4𝑟
3
.

Let us find out the elastic section modulus for this which is I by y max. So, which will be
4 3 3 3
𝐼 π𝑑 1 π𝑑 π(2𝑟) π𝑟
𝑍𝑒 = 𝑦𝑚𝑎𝑥
= 64
. 𝑑 = 32
. Which will be 32
= 4
.
2

𝑍𝑝 4𝑟
3
4 16
So, as a classical definition shape factor is 𝑆 = 𝑍𝑒
= 3
. 3 = 3π
. which can be 1.70.
π𝑟

So, that is my shape factor for a solid circular bar.

349
(Refer Slide Time: 05:19)

Having said this let us do one more example, which will do it now for an annular ring tubular
section. Let us say a tubular section you know the outer radius is 𝑟1 and the inner is 𝑟2 so;

obviously, (𝑟2 − 𝑟1 ) will be the thickness of the tube, which is t= (𝑟2 − 𝑟1).

Let the upper centroid be at y bar 1 from the equal area axis and the bottom centroid be y2
and we call this as 𝐴1 and this as 𝐴2. So, from the figure you can very well easily calculate
2 2
π𝑟1−π𝑟2
𝐴1 = 𝐴2 = 2
=
π
2 (𝑟
2
1
2
)
− 𝑟2 . We will also know 𝑦1 = 𝑦2 =
∑𝐴𝑦
∑𝐴
.

Σ𝐴𝑦
Let us do this way Σ𝐴
we will employ this equation. So, let us try to find out. So, I am

taking a semicircle. So, this has got 2 areas. So, this one is separate and this one is separate
we know this is at the radius 𝑟1 and this is at the radius 𝑟2. So, with the help of this we will

employ this equation and we will calculate this 𝑦1 which is as same as 𝑦2 will be

∑𝐴𝑦
𝑦1 = 𝑦2 = ∑𝐴
.

2 2

Let us say 𝑦1 = 𝑦2 =
π𝑟1
2 ( )− ( ) .
4𝑟1
3π
π𝑟2
2
4𝑟2
3π
am I right. Which we simplify will be
π 2 2
2
(𝑟1−𝑟2)

3 3
4𝑟1 4𝑟2
𝑦1 = 3π
− 3π
.

350
(Refer Slide Time: 08:54)

3 3
4𝑟1 4𝑟2
−
Let me write it this way.𝑦1 = 𝑦2 = 3π
2
3π
2 . Let us take it this way. Which will be said as
(𝑟1−𝑟2)

3 3
4 𝑟1−𝑟2
𝑦1 = 𝑦2 = 3π
( 2 2 ).
𝑟1−𝑟2

We will retain this equation we will call this equation number 1 which is 𝑦1, which is also
𝐴
equal to 𝑦2 instantaneously. So, we can find quickly the 𝑍𝑝 = 2
(𝑦1 + 𝑦2), which will be
3 3
4 (𝑟1−𝑟2)
𝑍𝑝 =
𝐴
2 (𝑦 1
+ 𝑦2 = ) π
2 (𝑟
2
1
− 𝑟2 [
2
) (
3π 𝑟2−𝑟2
1 2 )
]. Which gets simplified to 𝑍𝑝 =
4
3
3
(𝑟1 − 𝑟2).
3

So, that is my 𝑍𝑝.

Let us now quickly find out 𝐼𝑍 =

π
64 (𝑑 4
1
− 𝑑2 =
4
) π
64 ( 4
)
4
[ 2𝑟1 − 2𝑟2 ]. So, I can now find Z
𝐼𝑍 π 4 4
equivalent as 𝑍𝑒 = 𝑦𝑚𝑎𝑥
= 4𝑟1
(𝑟1 − 𝑟2).

𝑍
So, I can now quickly find the shape factor as 𝑆= 𝑍𝑝 . Let us say equation this we call as
𝑒

equation number 3. So, I should say now this is equal to equation 2 by equation 3. So, which
3 3
4𝑟1 16𝑟1 (𝑟1−𝑟2)
you simplify which will be 𝑆 =
4
3 (𝑟 3
1
− 𝑟2 .
3
) 4
π(𝑟1−𝑟2)
4 . Which will become 𝑆 = 3π 4
(𝑟1−𝑟2)
4 .

351
(Refer Slide Time: 13:13)

𝑟2
Now, let 𝑟1
= 𝑘. Then shape factor S was actually equal to

(𝑟 −𝑟 )
3 3 ⎡ (𝑟3−𝑘3𝑟3) ⎤
16𝑟1
( ).
3
1 2 16 ⎢ 1 1
⎥. Which will now become 𝑆 = 16 1−𝑘
𝑆 = =
(𝑟 −𝑟 ) ( )
⎢ 𝑟4−𝑘4𝑟4 ⎥
3π 4 4 3π 3π 4
1−𝑘
1 2
⎣ 1 1
⎦

Where k is a ratio of the radii. Yes, interestingly friends by substituting 𝑟2 = 0. This will
16
reduce to solid circular tube, is it not? So, therefore, the shape factor s for a solid tube is 3π

which we already have. So, we landed up in the same equation same as derived earlier. So,
we have a very interesting example of the solid section and an annular section.

352
(Refer Slide Time: 16:05)

Let us do one more example as you see in the figure, which is a channel section. Because
these are common sections used in steel design. So, you are trying to work out the shape
factors for all of them. So, the design becomes easy and can handle it comfortably. So, look at
the section available on the screen. Now let us say this dimension the overall size is 200 mm,
and the thickness is 5 millimeters, thickness is 5 millimeters through and through, right.

Let us try to find out the total area of cross section. You can see this is going to be I have
divided like this, this is my first piece. This is my second piece and my third piece. So, which
is going to be now 100 into 5, there are 2 such pieces plus 200 minus 10. So, 190 into 5. So,
the total area becomes 1950. Since the section is symmetrical equal area axis passes through
95
(100 × 5 × 97.5)
= ⎡⎢ (100 × 5)+(95 × 52 ) ⎤⎥ = 73. 14 = 𝑦
Σ𝑎𝑦
the CG. 𝑦 = Σ𝑎
⎣ ⎦

Σ𝑎𝑦
One can easily find out 𝑦 which can be 𝑦 = Σ𝑎
. We will try to find out look at this figure b

here. So, that is going to be 100 into 5 into 97.5. That is the CG of this piece, right. 100 into 5
97.5 because this is 100 from here till here it is 100. So, now, this will not be. So, from the
CG to here this will be 97.5 plus then the second piece is this which is 95 into 5, that is the
area of this piece and CG will be 95 by 2.

Divided by total area 100 into 5 plus 95 into 5. If you work it out you get this value as 73.14

which is y bar which is indicated here. Now, I can work out 𝑍𝑝 =

𝐴
2 (𝑦 1 )
+ 𝑦2 is already

353
there. So, I can simply say it is 1950 by 2 73.14 into 2. Because 𝑦1 and 𝑦2 are identical. So, if
3
you do that, I will get this value as 142623𝑚𝑚 .

𝑍𝑝 =
𝐴
2 (𝑦 1
+ 𝑦2 =) 1950
2
(73. 14 × 2) = 142623𝑚𝑚 .
3

Let us try to find out the 𝐼𝑍 value of this channel section. We can use parallel axis theorem.

So, let us do that 100 into 5 cube by 12 plus 100 into 5 into 97.5 square that is my first piece.
I have 2 such pieces plus these are for the flanges. Now for the web going to be 5 into 190
cube by 12, you know if this is 200 this will be 190. Because the thickness is 5 in both cases.
6 4
So, if it make the total of 𝐼𝑍 this becomes 12. 366×10 𝑚𝑚 .

{ }
3 3
2 6 4
⎣
100×5
𝐼𝑍 = ⎡⎢ 12 + 100×5×97. 5 ( ) ⎤×2 +
⎥
⎦
5×190
12
= 12. 366×10 𝑚𝑚 .

𝐼𝑍 12.366×10
6
I can find Z elastic which is 𝑍𝑒 = 𝑦𝑚𝑎𝑥
which will be 𝑍𝑒 = 100
. Now this is the

4 3
extreme fiber from the neutral axis which will be 12. 366×10 𝑚𝑚 . So, I have 𝑍𝑝, I have 𝑍𝑒

can I quickly find the shape factor.

(Refer Slide Time: 22:38)

𝑍
So, the shape factor will be now equal to 𝑆= 𝑍𝑝 . which will be you can work it out S=1.153,
𝑒

friends. So, very simple example which is easily understandable.

354
(Refer Slide Time: 23:08)

Let us do one more section the T-section the dimensional T are given, 150 and the overall
thick depth is 100 mm, thickness of the flange and the web are different in this case. Flange is
5 mm whereas; web is thicker 10 mm. Now in this case the first step is to locate the equal
area axis.

In the previous example this problem was not there because there was an axis of symmetry.
In this case of course, you have vertical axis of symmetry there is no doubt, you have vertical
axis of symmetry. But I want to find what would be this value. So, we would like to locate
this, right. So, the first job is to find out the equilibrium axis. So, let us divide this into 2
parts.

Let us say this is my first piece and this is my second piece. Considering these 2 areas has got
to be equal because an equal area axis no. So, let us find out area of the flange which is 150
into 5. Let us find out the area of the web which is 95 into 10 which is 950. So, area of the
web is more than the flange. So, therefore, the equal area axis has to lie at a distance 𝑦 down
into the web.

So, we have to find out this distance 𝑦 now, which is measured from the. Let us mark 𝑦 like
this not here let us mark 𝑦 here. Let us mark y bar which is going to be from the flange web
junction let us say this is my 𝑦. Let us do this. So, let y bar be measured from the intersection
of flange and web.

355
 𝐴 (750+950) 1700
So, I can now say 2
= 2
= 2
= 850. And that should be equal to area of the

flange plus 10 times of 𝑦 am I right? Look at this figure. 850 = 𝐴𝑓 + (10𝑦).

So, I have this the governing equation I have area of the flange I know which is
850 = 𝐴𝑓 + (10𝑦) which will be telling me that 𝑦 is 10 mm. So, at 10 mm from here I

locate the equal area axis. So, I have located the equal area axis. One can also check 𝐴1 going

to be area of the flange plus 10 into 10 because I am working out this area. Which is 750 plus
100 which is 850 which exactly equal to A by 2 as you see here.

𝐴
𝐴1 = 𝐴𝑓 + (10 + 10) = 750 + 100 = 850 = 2
.

Now I want to locate the CG of both the sections. Now step number 2 will be to locate the
CG’s y bar 1 and y bar 2 of upper and lower sections. So, look at the figure b. So, 2 sections
Σ𝑎𝑦
are drawn separately let us try to find out and apply this equation. So, 𝑦 = Σ𝑎
.let us apply
(150×5×12.5)+(10×10×5)
this and try to find out for 𝑦1 first. So, let us say 𝑦1 = 850
= 11. 62𝑚𝑚.

So, it will be lying in the flange, as marked in the figure. Let us do 𝑦2. 𝑦2will be you know
85
this dimension the total is 95 you lost 10 there. So, its 85. 𝑦2 = 2
= 42. 5. So, can I quickly

find 𝑍𝑝? 𝑤ℎ𝑒𝑟𝑒 𝑍𝑝 =

𝐴
2 (𝑦
1 )
+ 𝑦2 = 850(11. 62 + 42. 5) = 46002𝑚𝑚 . We got Z p.
4

356
(Refer Slide Time: 31:13)

𝐼
Let me find 𝑍𝑒 we know 𝑍𝑒 = 𝑦𝑚𝑎𝑥
. Let us now consider the full T-section. Now to locate the

CG of the T-section. So, let us say the CG is taken as y bar in this case. So,
95
Σ𝑎𝑦 (150×5×2.5)+95×10×( +5)
𝑦= Σ𝑎
= 1700
2
= 30. 44𝑚𝑚. So, if you work out this 𝑦 will become

30.4 millimeters which is indicated here.

So, now, let us find out the moment of inertia of the section about the z axis, which will be
3
𝑏𝑑
using parallel axis theorem. Which can be said as 12
for the first piece
3
150×5 2
𝐼𝑍 = [ 12
+ 150×5 + (30. 44 − 2. 5) , that is for the first piece. Do it for the second
3
10×95 2
piece. [ 12
+ 95×10 + (52. 5 − 30. 44) .How do you get this 52.5? Is exactly this. So,
6 4
can you find out this? Which will be 𝐼𝑍 = 1. 767×10 𝑚𝑚 . So, now, I can find
6
𝐼 1.767×10 3
𝑍𝑒 = 𝑦𝑚𝑎𝑥
= (100−30.44)
= 25402. 5𝑚𝑚 . Because this dimension is 30.44 and this value

will be 100 minus 30.44. So, I substitute that here I get 25402.5, this should be 𝑦𝑚𝑎𝑥

remember that friends this is 𝑦𝑚𝑎𝑥.

𝑍
So, now I have say d with me. So, can I find the shape factor shape factor 𝑆= 𝑍𝑝 .
𝑒

357
(Refer Slide Time: 35:55)

𝑍 46002
𝑆= 𝑍𝑝 = 25402.5
= 1. 81(Refer Slide Time: 36:11)
𝑒

We will do one more example for an L-section. Let us say then L-section the dimensions are
given on the screen the overall thickness is same which is 5 millimeters. So, let us first locate
the centroidal axis of this.

Is an unequal angle. So, it has no axis of symmetry. So, we need to find out the equal area
axis and the CG both. So, let us find out the area of the shorter arm. The area of the shorter
arm is 60×5 the area of the longer arm is 95×5, the total area is

358
 2
𝐴 = 𝐴1 + 𝐴2 = 300 + 475 = 775𝑚𝑚 . So, now, let us locate the equal area axis first. Let

the equal area axis be at a distance y from the intersection phase as shown in the figure, y.

So, we know that 60 into 5 plus 5 into y. Should be equal to 775 by 2, by this logic y will
become 17.5 millimeters. So, y will now become 17.5 millimeters. We can check also. So, 𝐴1

plus 17.5 into 5 should be equal to 300 we will call this as 𝐴1 and this as 𝐴2, total area is 𝐴1
775
plus 𝐴2. (60×5) + (5𝑦) = 2
.where y=17.5mm.

So, 𝐴1 plus 17.5 into 5 which is 87.5 which is 387.5 mm square is actually equal to 775 by 2.

So, its, is it not? So, we have located now the equal area axis now, our job is to locate the
centroid, centroid of upper and lower areas. So, let us look at figure b. So, I want to find the
centroid of the upper part centroid of lower part then I should find out the 𝑍𝑝, right.
2 775 𝐴
𝐴1 + (17. 5×5) = 300 + 87. 5 = 387. 5𝑚𝑚 = 2
= 2
.

(Refer Slide Time: 40:26)

Let us take the upper part. Let us find out the lower part, let us say let us take the lower part.
Σ𝑎𝑦
So, we will apply this equation 𝑦= Σ𝑎
.So, we get
17.5
[60×5(17.5+2.5)]+(17.5×5× )
𝑦1 = 387.5
2
= 17. 46𝑚𝑚.

359
Let us work out 𝑦2. 𝑤ℎ𝑒𝑟𝑒 𝑦2 is very simple because we know the overall dimension. So, this
77.5
is going to be , 𝑦2 = 2
= 38. 75𝑚𝑚. So, can you quickly find 𝑍𝑝friends, using this

equation which is 𝑍𝑝 =
𝐴
2 (𝑦 1 )
+ 𝑦2 =
775
2
(17. 46 + 38. 75) = 21781. 38𝑚𝑚 . So,
3

similarly 1 can locate the CG. I am not giving the answer; I mean I am not giving the
procedure I think you should be able to work it out.

Σ𝑎𝑦
𝑦= Σ𝑎
= 33.15 mm. Once I get 𝑦 I can find 𝐼𝑍 which will be
3 3
60×5 2 5×95 2 5 3
𝐼𝑍 = ⎡⎢ 12 + 60×5×(33. 15 − 2. 5) ⎤⎥ + ⎡⎢ 12 + 95×5×(32. 5 − 33. 15) ⎤⎥ = 8. 175×10 𝑚𝑚
⎣ ⎦ ⎣ ⎦
𝐼𝑍 8.175×10
5
3
, 𝑍𝑒 = 𝑦𝑚𝑎𝑥
= (100−33.15)
= 1228. 87𝑚𝑚 . Therefore, shape factor will now become 𝑆=

𝑍𝑝
𝑍𝑒
= 1. 78.

(Refer Slide Time: 44:48)

360
(Refer Slide Time: 44:54)

So, friends in this lecture we learnt, the shape factor calculations for rectangular section, solid
cylinder, tubular member, L-section, T-section and we learnt that shape factor varies with the
geometry it is not same, right. The maximum value of the shape factor is about 2.0 which is
for a diamond section, you can try and find out.

So, shape factor is a geometric property and if I know shape factor, I can easily find the
capacity of the section. Which is very simple to find out from the derivations what we
discussed in the last lectures.

Thank you very much have a good day, bye.

361
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 21
Plastic analysis -1

(Refer Slide Time: 00:20)

Welcome to the 21st lecture on the course advanced steel design, in which we are going to
talk about the moment curvature relationship and Plastic analysis I. We are going to learn
these two things in this lecture now. Now, let us talk about the moment curvature relationship,
which is M-𝜑 curve. Let us consider a simply supported beam of a rectangular cross section b
cross d.

Say this is b and this is d, now let us keep it as h. So, let us draw the stress distribution
diagram of an elastoplastic section. Let us say this is partly plasticized on the extreme tension
compression fibers, then the remaining is elastic. So, let us say this is the plasticized part and
the remaining is elastic part; it is an elastoplastic section and this is the cross section.

And of course, we know this is h/2 and this is my y axis, this is my z axis and this is my x
axis, . We know according to theory of simple bending, Mz/Iz =E/R=𝛔yield/y, that is equation
number 1. Now, if you derive this further from this; we can say , Mz/EIz =1/R=ϕ, which I call
as the curvature and this is my z axis .

362
(Refer Slide Time: 04:46)

Now, one can see from equation 2 that, the curvature and the moment are proportional. Any
marginal increase in moment which is caused by the loads will increase the curvature. But
there is a big issue here, the linear relationship between moment and the curvature is valid
only up to limit of proportionality for the elastic material. When the moment reaches yield
value, that is the moment reaches moment at yield value, which I called as M yield; then the
relationship between M-ϕ is modified.

So, I should say My/EIz =1/R= 𝛔yield/y =ϕ, which is the curvature, call this equation number 3.
Now, look at the cross section here; the overall depth of the section is h.

363
(Refer Slide Time: 07:28)

So, the overall depth of the beam is h. So, hence y=h/2, that is the distance of extreme fibre;
therefore, 𝛔yield /(h/2) which should be equal to, let us let me write the equation again here, let
us say My/EIz =1/R= 𝛔yield/y =ϕ. So, let us write , 𝛔yield /(h/2)=E/R or 2𝛔yield /(h)=E/R, 2𝛔yield
/Eh=1/R is the curvature. Let us call this as equation 4, equation 5 and equation 6.

(Refer Slide Time: 09:14)

Now, after reaching yield, if further moment is applied; the section will get plasticized. So,
this state is called as elastic-plastic state.

364
So, at this stage partially the section is plastic and partially it is elastic; therefore, the depth of
elastic core is identified as e. Let us look at this figure; we call this as depth of elastic core, so
e is depth of elastic core. So, for elastic sections or for the elastic core following expression is
valid; because there is a linear relationship existing, therefore I can say 2𝛔y /Ee =1/R.

We already said 2𝛔y /Eh=1/R; I am replacing this h with e for an elastic core right, equation
number 7.

(Refer Slide Time: 12:06)

So, at this stage the linear relationship between moment and curvature will not hold good. So,
as stated in plastic design, let us keep on continuing to load the beam. This will now increase
the moment, beyond M y. But the section will continue to carry the load, why?

It is because the section is not fully plastic, still some part of the section is elastic. So, the
section will keep on continuing to carry the load until the complete section or the complete
cross section is plasticized, at that stage moment will become Mp. So, when Mp is reached,
what would be the depth of elastic core? Elastic core will be zero; because this elastic section
will go away, therefore I can say e will become 0.

So, let us substitute, let us substitute e as 0 in equation 7. So, if you put e as 0 in equation 7,
what happens?

365
(Refer Slide Time: 15:08)

So, for e become 0, equation 7 will tend to become infinity; that is 1 by R which is a
curvature, which is M by E I z will become infinity, right. Therefore, we can say 1 by R; let
us write this equation as 8, 1 by R at yield will be equal to curvature at yield will be equal to
M, at yield by E I z.

Also, M p by M yield is phi p by phi yield; we already have this relationship with us, moment
and curvature are directly proportional till the elastic limit. So, we can use this relationship
now. So, therefore, friends M p we know it is Z p into 𝛔 y and M y is Z e into 𝛔 y, which will
be phi p by phi yield. Now, friends we all know that this relationship is actually shape factor,
equation number 11.

So, having said this, let us rewrite this equation of elastic core; we already know M equals
Mp(1 – e2/3h2).

366
(Refer Slide Time: 17:05)

We also know twice of 𝛔 y by E e is 1 by R; this equation we already have, see here, is it not.

Which tells me 2 R into 𝛔 y by E is the depth of elastic core. Let us substitute this value here.
So, I call this as equation number 12.

(Refer Slide Time: 18:35)

So, equation 12 now becomes M by M p = 1 minus; e will be 2 r 𝛔 y by capital E. So, 2 R 𝛔 y

by capital E the whole square by 3 h square, which becomes 1 minus 4 R square 𝛔y square by
3 E square h square, equation number 13 and 14.

367
So, therefore, M by M p is 1 minus 1 by 3 of, I am rewriting equation 14 slightly in a
different form, 2 𝛔 y by E the whole square 1 by h by R the whole square. 2 𝛔 y by E the
whole square into 1 by h by r the whole square.

So, let us see this equation of curvature. We already know 2 𝛔 yield by E is h by R at yield,
we already know this equation. So, let me substitute h by R yield from here to here. So, can I
write this as now M by M p will become 1 minus 1 by 3 of h by R yield square pi h by R the
whole square, can I say this?

(Refer Slide Time: 21:52)

Which I can say now as M by M p is 1 minus 1 by 3 of h by R-at yield divided by h by R in

general of the whole square.

So, what does it mean? If you have an elastic plastic section, whose radius of curvature and
depth of section are known or if I know the M p of the section which is Z p into 𝛔 y, which
we already know that Z p is Z elastic into shape factor into 𝛔 y. In a given tables steel table Z
e for different sections available and shape factor is also known. So, M p is known; applied
moment M is known, because you know the load, depth of the section is known, you can find
the radius of curvature and 1 by R will give you the curvature.

So, moment curvature relationship can be easily defined and derived for various pairs of M-ϕ;
you can easily know this here, you can plot this. That is a very interesting understanding we
learnt from the moment curvature relationship using plastic design concept.

368
(Refer Slide Time: 23:57)

Let us go to another important segment of plastic design, which we call as a load factor. load
factor is generally defined as; let us call this as Q is defined as the ratio of collapse load to
working load. Let us call this Q as collapse load to working load. We also know that applied
moment M is proportional. Let us say moment M is proportional to the applied load, hence
moment is proportional to W.

(Refer Slide Time: 25:30)

So, therefore, Mw is sometimes of Ww. Let us equate it to some proportionality constant; we

also know Mp is k times of collapse load. So, where Mw is the moment caused due to

369
working load and Mp is the moment caused by collapse load. So, in plastic design, we always
estimate the load as a collapse load because it is M p. Now, Q is W c by Ww; by this logic,
this can be simply M p / M w.

We also know M p is yield stress of Z p and M working is 𝛔 allowable of Z e; we are talking

about working loads. Hence, Q now will become 𝛔 y of Z p by 𝛔 allowable of Ze, which can
be 𝛔y by 𝛔allowable into Zp by Ze, which is actually equal to the shape factor multiplied by
factor of safety.

(Refer Slide Time: 27:28)

So, we have a very interesting relationship, load factor is a product of shape factor into factor
of safety. So, the load factor which is used in the plastic design is enhanced by the shape
factor. So, in plastic design, factor of safety is enhanced by the shape factor. So, this ensures
that the design procedure is safe. So, plastic design claims one more advantage. Without
replacing the material, one can improve the moment capacity by simply choosing an
appropriate geometry.

So, I will put it as not the word geometry, I will put it as a better shape factor; can I use this?
So, now, we have done the shape factors for different rolled sections like open sections of L,
T, channels, etcetera.

370
(Refer Slide Time: 29:55)

So, since tubular sections which are closed and hot rolled have higher shape factor, if you
recollect; It was approximately equal to 1.7, higher shape factor they are commonly used in,
for example offshore platforms.

So, offshore structures are mostly tubular members; this is because of the reason that they
have a very high shape factor. So, any strategic structure where the moment capacity needs to
be enhanced, can employ plastic design procedure for geometric centric design; that is for a
form dominant design, one can use plastic design and one can use cross sections of higher
shape factor.

371
(Refer Slide Time: 31:35)

But there is only one catch here, the material what you use should be ductile, should not fail
suddenly; it should give enough warning before failing. So, if you think about a material
immediately in mind, you will find the steel possess this advantage; therefore, steel is the
most successful and most prominent material recommended by engineers for plastic design of
sections.

That is the reason, otherwise there is no very fancy attraction about steel as a construction
engineering material. Steel almost possesses all type of advantages in the geometric side, in
the material side, in the fabrication, in the recycling, quick return on investment and of course
in maintenance; but the only issue what steel has is an environmental influence on the
material, which is strength degradation because of corrosion.

So, if that is addressed by suitable alternate suggestions on the material of steel, like
functionally graded materials; for example, which we studied in detail in the previous
lectures, then the construction practices use of steel or FGM on strategic structures can go
very sound and that can give you a very good performance as well as economic design, that is
the idea what you want to emphasize in this.

372
(Refer Slide Time: 33:52)

So, now let us take it forward; how do I do a plastic analysis? Let us do that load factor which
is used in plastic design is a function of shape factor, right. So, the Q of S is shape factor 𝛔 y
by 𝛔 allowable; if I say this is my factor of safety, so I can say that in plastic design the load
factor is enhanced shape factor times the factor of safety. So, it is a very safe design. In
plastic analysis as we go further, we have a concept called mechanism.

(Refer Slide Time: 35:06)

Let us ask a question what is a mechanism? If a body is subjected to any load, it will offer
resistance; we call this as internal resistance to the applied loads, is it not. This process of

373
offering resistance is otherwise termed as load carrying capacity; in general, this is the
capacity. Now, if by any chance the body is unable to offer resistance; there may be many
reasons for this, then it is called a mechanism. So, mechanism is a simple definition of no
resistance to load.

(Refer Slide Time: 37:22)

Then what is the collapse load? The load at which the body stops to offer resistance to the
external load is termed as collapse load. So, the structural system will become a mechanism
under certain conditions. So, the question comes, when a structural system can become a
mechanism? A structural system can become a mechanism only when sufficient number of
plastic hinges are formed.

374
(Refer Slide Time: 39:07)

So, what is the plastic hinge then? Plastic hinge is a connection or a joint or a cross section or
a section; not cross section or a section, whose moment capacity is M p.

Remember, you have a structural hinge like roller support, like hinged support, here the
moment is equal to 0, whereas a plastic hinge is a section where the moment is not 0, the
moment is equal to Mp. Now, the question comes if sufficient number of plastic hinges are
formed in a structural system, the system can become a mechanism.

(Refer Slide Time: 40:46)

375
Then let us ask a question for a system to become a mechanism, we agree that sufficient
number of hinges are to be formed. How many plastic hinges are required? The answer is
very simple; if the degree of static in determinacy is n, then n plus 1 plastic hinge are required
to convert a structural system into a mechanism. Then the question comes, where these
hinges can form, where plastic hinges can form?

(Refer Slide Time: 42:00)

They can form at specific locations; like at fixed supports, at points of concentrated loads, at
sections where bending moment is maximum, at sections where moment of inertia is changed
in non-prismatic members. Let us say at fixed support, plastic hinge can form; at point of
concentrated load, plastic hinge can form; at sections of maximum bending moment, plastic
hinge can form; if a section is having a non-prismatic cross section, plastic hinge can form.
So, these are the possible locations where plastic hinges can be formed.

376
(Refer Slide Time: 44:06)

Let us add one more point to our knowledge; there is a great caution with plastic hinges, the
caution is the plastic hinges have limited rotation capacity.

So, now there are some comments, observations what we can make, at plastic hinge, moment
is M p; just because it is a hinge, do not say moment is equal to 0, moment is not equal to 0, it
is Mp. It can form only at specific locations; where are they, we already said that. They have
finite rotation capacity; they are necessary to convert a structure into a mechanism.

Mechanism is that condition of a structure which cannot offer any resistance to the load; the
structure is completely out from its performance, it is collapsed, the structure cannot offer any
resistance when it becomes a mechanism.

377
(Refer Slide Time: 46:00)

A structural system by the virtue of plastic hinges is converted into a mechanism. So, my
question is, are we designing a system which is going to collapse?

The answer is not, we are designing a system going to collapse; we are going to estimate the
load at which this collapse can happen, that is called as the collapse load. The process which
helps us to estimate the collapse load is called plastic analysis. So, plastic analysis is a
procedure which helps us to find out the collapse load on a given structure.

(Refer Slide Time: 47:19)

378
So, friends in this lecture, we have learnt the moment curvature relationship, we have learnt
what is plastic analysis.

We have also understood what is a plastic hinge, where can it form, if it how many of them
can form; if it forms, it converts itself into a mechanism and what is a mechanism. So, in the
next lecture, we will start extending the learning process of estimating the collapse load using
different theorems, which are static theorem, kinematic theorem and uniqueness theorem.

Friends, you please revise these lecture notes, try to access to my textbooks referred in the
website of this particular course and keep on doing lot of exercises for shape factor and
estimating for different cross sections of geometric novelty; because they are very helpful in
making you to hold the answers for this.

Friends, there is one good news; for working out the shape factor, there are MATLAB
equations available for different problems in my textbook. You can copy paste them and run
in MATLAB window, you will get the answers of shape factor directly; you can also write
simple programs of this order for different cross sections and try to practice this for your
understanding.

Thank you very much, have a good day, bye.

379
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 22
Plastic analysis -2

(Refer Slide Time: 00:22)

Friends, we will have this lecture on Plastic analysis-2 where we are going to talk about the
theorems which are useful in estimating the collapse load. So, friends we are going to learn
about various theorems that can help to estimate the collapse loads. So, you should realize an
important statement when we talk about plastic analysis, you must realize that we are talking
about the stress level at σ σy and allowing plastic deformation in the material.

This has no direct relationship with inelastic displacement of the structural member.
However, large inelastic displacements of the structure can invoke the requirement of plastic
deformation. So, from the structure to the material there is a request passed on that the
material should undergo or should be capable of undergoing plastic deformation.

But from the material there is no request transferred to the system. that is important. So,
therefore, we keep on repeatedly saying plastic design or the design of large displacement
structures like form dominant does not depend on the material strength alone. they also
depend on the geometric configuration.

380
So, the geometric configuration enables a system to resist the loads to a significant extent of
course, you cannot have a good geometric configuration and the material can be made of zero
strength like paper we cannot have a structural member like that. So, we are not oaring that
extreme condition we say the material strength is important, but the design is not material
strength centric alone.

(Refer Slide Time: 03:30)

So, we also learnt in the last lecture that plastic hinges will be formed at critical locations we
have discussed what all these possible locations where plastic hinges can form. So, at critical
locations these hinges will be formed, but they have a limited rotation capacity and they have
a limited moment carrying capacity. What does it mean?

The moment carrying capacity of the plastic hinges is referred as M p which we can recollect
now confidently the σy *Zp where we can also say that is σ y in shape factor into Z e where
we already know for a given standard cross sections the elastic section modulus and the shape
factor can be easily computed.

So, M p is not the function of load, but it is a geometric centric property. So, plastic hinges
will be formed whose rotation will be limited and based on the theorems which we are now
going to discuss, one can estimate the collapse load. So, please be noting that we are referring
this load from the plastic analysis.

381
So, plastic analysis applied on to a structure statically indeterminate structure will give you
estimate of collapse load. Please understand very carefully the structure which is subjected to
plastic analysis the moment capacity is already known, it is not dependent on load. So, from
the moment capacity you calculate the load. So, it is actually a reverse engineering problem.

In convectional analysis load is given to you estimate the moment then design the section and
then check the section for exceedance of its bending stress shear stress etcetera that is the
conventional design procedure you know, but here the design procedure is entirely
challenging. You have the moment capacity of the section you want to find the collapse load
estimate.

(Refer Slide Time: 06:32)

So, one can ask me a question then what is the use, what is the necessity to estimate a
something called collapse load? This is a load which should not be exceeded at all. You may
ask you a question what is the difference between a collapse load and the characteristic load?
Let us say and the design load a collapse load is that which you get from the plastic analysis
and this load should not be exceeded in the service life of the structure.

Whereas, characteristic load is the probability that the chosen load intensity will not exceed
by 95 percent, for example, Indian codes, but there is a probability that this load is permitted
to exceed at least to a maximum of 5 percent. So, there is a possibility that this load can be
exceeded, but this load cannot be exceeded. So, there is a difference between this.

382
Then you ask me a question what is the design load? Design load is the characteristic load
multiplied by partial safety factor for the load, you get a design load. We already saw this in
the discussions earlier. So, we have a distinct values for understanding that what is a collapse
load. Having understood this let us see what are the theorems which will help us to do plastic
analysis of structures.

(Refer Slide Time: 09:08)

See we have got three theorems, they are static theorem, kinematic theorem and uniqueness
theorem. So, there are three theorems which are useful in estimating the collapse load and we
do plastic analysis of structures we will see one by one now.

383
(Refer Slide Time: 09:48)

Let us talk about static theorem. Static theorem is also called as lower bound theorem. the
statement of the theorem is as follows. For a given frame and loading if there exists any
statistically admissible bending moment distribution throughout the frame which is safe,
Then value of load W is lesser than the true collapse load.

What does it mean is, from the analysis you will be getting a load W which will always lesser
than the true collapse load. So, W is a load obtained from the analysis from the plastic
analysis and W c is the true collapse load.

(Refer Slide Time: 12:05)

384
So, we could say that according to this theorem collapse load W obtain from this theorem
will be lower than or equal to the true collapse load W c. So, what are the steps involved in
estimating W? So, we should put a condition here W is always lesser than or equal to W c
since it is always lesser than or equal to this theorem is called lower bound theorem, what are
the steps involved in this?

First for a given frame under set of loads, draw the statically admissible bending moment
diagram. So, once you draw the bending moment diagram identify the critical sections where
the bending moment is maximum that is very simple because looking at the magnitude you
can identify them. 3. Assume plastic hinges at these critical sections. Please note carefully
“assume”, there is no guarantee that the plastic hinges will form only at these sections in this
theorem.

Assume plastic hinges at these critical locations. So, we also remember that if n is a degree of
indeterminacy static. So, one need n plus 1 plastic hinges to become a mechanism.

(Refer Slide Time: 15:08)

Then once you do this in the last step calculate the collapse load. Since we say this method
assumes formation of plastic hinges at all critical locations this method is iterative. Further,
this method shall give only the lower bounds of the true collapse load, W will be either equal
to or lesser than W c what we get from the static theorem is W.

385
So, out of all the Ws of time you can choose the maximum and say that is means W c. So,
that at any given point of time the W obtained from the iterative mechanism will not exceed
W c. So, there are some comments about this method ,this method is tedious, is complex. I
will tell you why because one should be drawing the bending moment diagram prior to
calculating the collapse load correct.

So, this method has a prerequisite. You need to draw a bending moment diagram for the
given problem now you may ask me a question why do you need to draw a bending moment
diagram? Go back here. You have to draw the bending moment diagram identify the sections
has got the maximum bending moment and those sections are the locations of a plastic
hinges.

So, just to know where the hinges will form you have to draw the bending moment diagram
they therefore, this method is prone to errors. So, that is the limitation here.

(Refer Slide Time: 17:52)

Now let us talk about the next theorem. Next theorem is named as kinematic theorem.
Kinematic theorem is also called as upper bound theorem the statement of this theorem is as
follows. The statement says for a given frame and loading under set of loads assume a
mechanism a mechanism is assumed, value of the collapse load is then obtain based on the
assumed mechanism . The computed load will be either higher or equal to the true collapse
load.

386
So, we can say W will be either greater than or equal to W c since it is greater this is called as
upper bound theorem.

(Refer Slide Time: 19:55)

So, therefore, friends the collapse load determined from this theorem will be greater than or
equal to the true collapse load. Mathematically, W will be either equal to or greater than W c.
Let us see what are the steps involved in computing W from this method.

(Refer Slide Time: 20:43)

387
Step number 1. Assume certain collapse mechanism. We already know mechanism will form
in a given structure only when n plus 1 hinges form the structure will become a mechanism.
But there is something called partial mechanism. This is the condition where the number of
plastic hinges formed is lesser than n plus 1 there is something called complete mechanism.

It is a case where the number of plastic hinges found will be exactly equal to n plus 1 there is
also something called over collapse mechanism or over complete mechanism where the
number of plastic hinges are more than n plus 1 where n is the degree of static indeterminacy
of the frame.

So, you can assume either a partial mechanism a complete mechanism or an over complete
mechanism depending upon the number of hinges what you assume and we all know that
hinges cannot be assumed at any location of your choice, hinges can only form on certain
critical locations which I already discussed in the last lecture. So, we know for a given frame
what are those critical geometric sections, where hinge plastic hinges can form likely? So,
choose those points, assume plastic hinges there and assume therefore, a collapse mechanism.

(Refer Slide Time: 23:14)

Now, step number 2. For the assumed collapse mechanism find the collapse load to find the
collapse load use principle of virtual work. So, one may ask me a question what is the
principle of virtual work which we all know, but still we will write this principle states that
the work done by external forces during the collapse mechanism is equal to the work
absorbed by the plastic hinges.

388
So, mathematically external virtual work is equal to internal virtual work. External virtual
work is actually done by load multiplied by displacement, internal virtual work will be
moment multiplied by theta. Since we are talking about plastic analysis the work done will be
by the plastic hinges. So, I should say this M p this is theta p where M p is the plastic moment
capacity of the section and theta p is the rotation of that plastic hinge at that location.

It is not plastic rotation it is rotation of the plastic hinge at that location

(Refer Slide Time: 25:55)

So, now, friends the question comes different types of mechanisms can be assumed for
kinematic theorem. Sometimes, it will be very difficult for you to choose which the
appropriate mechanism is. Sometimes it will be confusing to decide the most appropriate
mechanism, because some of them can be inappropriate also because some of the assumed
mechanisms can be inappropriate that may lead to a confusion or it may lead to estimate of
wrong collapse load.

Since you are assuming the mechanism and estimating collapse load this procedure is
iterative. So, friends, both the theorems demand an iterative scheme because in one case you
plot the bending moment diagram and keep on assuming the location of plastic hinges and
derive the collapse load in the other case you assume the mechanism itself and estimate the
collapse load. So, if you make a mistake in either of the procedure which is prerequisite to
estimate the collapse load then the estimate of collapse load can become wrong that is
number one there is a danger there.

389
Number 2. In both cases some of them may be inappropriate also of assumption of hinges
therefore, the system requires iteration. Having said this let us try to see what the possible
independent mechanisms which are available are.

(Refer Slide Time: 28:29)

So, if I have a beam subjected to a load then I can always assume plastic hinge the fixed
support plastic hinge here and plastic hinge here. So, the deflected profile will become like
this. If you have a frame then I can assume and subjected to a lateral load like this I can also
assume a collapse mechanism like this where hinges are formed here that is not here and
here or here . So, these two are called beam mechanisms.

390
(Refer Slide Time: 29:38)

Suppose, if you have a frame subjected to a lateral load like this the frame can sway and there
can be hinges formed here, here, here and here. this is called a panel mechanism, some
literature advises this as sway mechanism.

(Refer Slide Time: 30:30)

Suppose, if you have a frame which is having a combination of loads like it has got a lateral
load and the gravity load then the deformation of the structure could be as marked now where
the hinges can form here, here, here either here or here. So, this is called a combined

391
mechanism. why it’s called as a combined mechanism? You can see there is a sway
mechanism here and there is a beam mechanism. So, it’s called as a combined mechanism.

So, these are all different collapse mechanisms other than partially collapse mechanism over
collapse and perfectly collapse mechanisms and so on. So, kinematic theorem demands to
assume a mechanism first then estimate the collapse load.

(Refer Slide Time: 31:39)

Whereas, there is a third theorem called uniqueness theorem. let us see what does this
theorem does. The statement of the theorem says for a given frame under set of loads, at least
one statically admissible bending moment distribution can be determined. If this bending
moment distribution results in insufficient number of plastic hinges, so that a mechanism
cannot be formed.

Let us ask a question how many numbers of plastic hinges are required to form a mechanism?
If n is the degree of indeterminacy static degree of indeterminacy, we at least need n plus 1
hinges to convert the frame into a mechanism. But there is a possibility in the given bending
moment distribution you may not get n plus 1 critical sections where you can assume plastic
hinges.

So, there can be a possibility that the bending moment distribution results in insufficient
number of critical sections where plastic hinges can be assumed. Therefore, it cannot become

392
a mechanism in that case, the collapse load obtained will be equal to the true collapse load,
mathematically W will be equal to W c.

(Refer Slide Time: 34:14)

So, in static theorem or lower bound theorem we have W less than or equal to W c in
kinematic theorem or called as upper bound theorem W is greater than or equal to W c in
uniqueness theorem, we say W is equal to W c. We can also plot this graphically as shown
here. let us say in the x axis I try to plot W c in the y axis I am trying to give W is what we
estimate from the theorem W is what we estimate from the theorem W c is the true collapse
load.

So, the kinematic theorem gives you upper more than W c or equal W c, the static theorem
gives you a lower bound and this itself is a uniqueness theorem . So, that is what graphically
we can explain.

393
(Refer Slide Time: 35:49)

So, now, let us quickly see what we have learnt from this particular lecture. This lecture
helped us to learn three theorems of plastic analysis. It also told that plastic analysis will help
to estimate collapse load. We also learnt different steps involved in estimating the collapse
load from two distinctly different theorems. Both the theorems only help us to approximately
guess the collapse load because both the theorems or the solution from both the theorems are
iterative in nature.

So, the moment I say iterative there is a possibility that there can be some error in
computation. We also learnt that worse by worse the collapse load estimated will be either
lesser than or equal to true collapse load will be greater than or equal to true collapse load
uniqueness theorem is an hypothetical theorem it’s not applicable. So, we do plastic analysis
either by static theorem or by kinematic theorem and estimate W by iteration. So, that our
objective is to then conclude what will be the W c.

394
(Refer Slide Time: 37:57)

In terms of W c will be computed in terms of M p, very interesting in conventional analysis

M is a function of load know this is conventional. In plastic analysis W c is a function of M p,
because M p is known to me which is known independent of load it’s a section property it of
course, depends on material it’s the section property . So, that is the beauty what we have
learnt so, far in this lecture.

Thank you very much have a good day.

395
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 23
Plastic analysis -3

(Refer Slide Time: 00:19)

Friends, welcome to lecture 23, which is Plastic analysis III where we will learn few
examples to estimate collapse load. We will use both the static theorem and the kinematic
theorem to solve the problems. We will start with simple examples and see how we can
understand this theorem applications.

396
(Refer Slide Time: 01:14)

So, example 1, we consider a fixed beam with central concentrated load, we call this load as
W. So, I have a beam which is fixed at both the ends with the central concentrated load W, let
me name the points as A B and the midpoint as C and we say this span is equal and this is l E
comma I A .

So, as per the static theorem we have first to draw the admissible bending moment diagram,
for a fixed beam we know if this is my base line. So, there is a fixed moment which is equal
to Wl/8 negative bending moment, then there is a positive bending moment which is the
midspan is equal to Wl/4.

I superimpose these on the same side. So, I draw the negative bending moment diagram here
and the positive bending moment diagram I just swap it like this. So, this is negative and this
is positive again negative, we already know that this value is Wl/4.

So, now the critical sections are at three different places. The critical sections can be here, can
be here, and can be here. So, what is the static degree of indeterminacy for this problem let us
ask a question what will be the static degree of indeterminacy for this problem.

The static degree of indeterminacy for this problem is 2 by neglecting axial deformation that
is 4 are the unknowns moment, reaction, moment, reaction. There are two equations of
equilibrium Σ fy = 0 and Σm=0. So, the degree of indeterminacy becomes 2 by neglecting
axial deformation.

397
So, I need 3 hinges to get a mechanism, let us say these are my 3 places where the hinges can
form. So, the governing equation now going to be 2 Mp because this is also Mp this is also
Mp hinges is formed here. So, 2 Mp should be equal to Wl/4. Therefore, Wc which I said W
will be 4 into 2 Mp by l which is 8 Mp by l. So, that is my collapse load.

Ultimately we assume that, hinges are formed at three places hinges are formed at A B and C.
making it as a complete mechanism, is it not? I need 3 hinges I got 3 hinges. On the other
hand let us try to solve this problem using kinematic theorem. So, I have a fixed beam
subjected to central concentrated load W the locations are A B and C.

Let me now assume a mechanism, so the assumed mechanism is the beam mechanism. So, I
have one rotation here another rotation here and another rotation here by symmetry you will
see this is 2 θ, is it not? So, I need 3 hinges let us assume the hinges here, here and here
which are marked in this figure also.

So now, let us see what is the external virtual work? External virtual work is load into
displacement let us call this as Δ. What is the internal virtual work? Internal virtual work will
be the work done by the plastic hinges and its rotation which will be Mp into θ which is at A
plus Mp into 2 θ which is at C plus Mp into θ is at B which gives me 4 Mp θ.

So, let us equate external virtual work to internal virtual work. So, there are two unknowns
here Δ and θ, but in the figure they are connected. One can say very well if the span is l, you
can say for small rotations tan θ which is θ which is actually Δ by l by 2 which is 2 Del by l.

So, substituting back here W into δ will be 4 Mp into 2 δ by l. So, δ goes away I get Wc as 8
Mp by l which is same as this. So, friends both the theorems will ultimately converge to give
you the same answer if the problem is simple. So, either methods or both procedures give you
the same answer in this example. Now let us take another problem.

(Refer Slide Time: 09:22)

398
Let us say a fixed beam is loaded with uniform distributed load of intensity W per unit length,
obtain the collapse load. So, let us do that. Let us say I have a fixed beam subjected uniform
distributed load over a span of l E I let us mark this as A and B. There is a critical section at
the center which will be see because we all know for a beam under uniform distributed load
the maximum moment can also occur at the mid span of the beam.

So, what we should do in static theorem, we must first draw the static admissible bending
moment diagram, the static method and this is a kinematic method. Let us draw that, so I am
drawing statically admissible bending moment diagram. So, let me draw the fixed end
moment diagram separately we all know this intensity is Wl2/12 .

Then let me superimpose the bending moment diagram of this which will be equal to Wl2/8.
So, now, the net bending moment diagram is what I have here right out of which this is
positive this is negative. So, this neutral line has shifted to this line.

So, now we have drawn a statically admissible bending moment diagram, let me then do for
the statically admissible bending moment diagram. Let us first ask a question how many Mp
is required for this beam? it is a fixed beam, it has got static degree of indeterminacy as 2. So,
we need three hinges to make it as a mechanism.

So, let us assume these hinges are at A, B and mid span which is C. So now, I have a hinge
here, I have a hinge here, I have hinge. Now I say there is a hinge here, there is a hinge here
also, is it not? So, I should say, 2 Mp= Wl2/8. So, which will give me Wc = 16 Mp/l2, which

399
is my collapse load. Let us try to do this problem using kinematic theorem a fixed beam,
under uniform distributed load for a span of l and E I .

Let me try to draw the mechanism. Let us say the hinges are allowed to form at 3 locations
one at the mid span and therefore, this becomes my assumed mechanism. So, one hinge here,
one hinge here and one hinge here. Now let us take this deflection as δ let us say this rotation
is θ.

Now let us see what the external virtual work is. So, the external virtual work is done by the
load on this area. So, it is nothing, but if it is udl it is half into base into height into W which
will be Wlδ/2. Internal virtual work will be equal to Mpθ which will be at A, plus Mp into 2 θ
which will be at C which is the mid span of the member plus Mp θ that B which becomes 4
Mpθ/l.

So, by principle of virtual work I must equate external to internal virtual work. So, Wlδ/2
should be 4 Mpθ/l. So, from the figure we know that tan θ is θ = δ/(1/2). So, θ becomes 2δ/l
let us substitute that. So, W*l*δ/2 is 4Mp(2δ/l), so δ goes away. So, that becomes 16Mp/l2
which is Wc, which is same as this.

So, I get the same collapse load by both the methods. if the problem is simple where I can
draw the statically admissible bending moment diagram easily, I can assume a perfect
collapse mechanism easily and I can get the same answer by both the methods. And you will
also notice that there was no iteration involved. You may be wondering when will the
iteration come, when you have got more than one collapse load obtained from the analysis,
then the iteration will start.

400
(Refer Slide Time: 16:22)

Let us do another interesting problem, where I have a simply supported beam subjected to
eccentric load. So, let us say I have a simply supported beam one end hinged, other on roller,
I have a load which is eccentric. So, this is A and this distance is B, of course, the span of the
member is l and the member has got EI property. Let me draw the statically admissible
bending moment diagram, let us first ask what is the degree of indeterminacy for this beam it
is 0.

So, how many plastic hinges I need? I need only 1 plastic hinge, where it can form? It can
form only at C. It cannot form at A and B, because A and B are already hinged connections.
So, let us draw the bending moment diagram.

So, the hinge can form only at one point, that is here and we all know that value will be
Wab/l, So, this is the point where the hinge is formed therefore, I can say Mp = Wab/l
therefore, Wc is a collapsed load will be Mp*l/ab which will be Mp(a+b)/ab. So, that is
collapse load, no iteration, and straight forward solution.

Let us do this using kinematic theorem. Let us draw the admissible collapse mechanism here
is going to be beam mechanism. So, hinges can form only at one location that is here. Now
you may wonder how I am drawing the hinges very interesting friends.

If I draw a circle which is unfilled, this is a structural hinge. If we form a circle with filled
this is called plastic hinge. So, these two are structural hinges where the moment is equal to 0,

401
this is a plastic hinge where the moment is equal to Mp. So, now since it is unsymmetric the
load this will be θ1 and this angle will be θ2.

And this rotation will be θ1 + θ2 by simple geometry. Now we also know that θ1 let us call this
value as δ, so δ/a. So, that is δ =a*θ1. θ2 is δ/b which means δ is also equal to θ2*b. So, let us
say the external virtual work done for this problem is W*δ, the internal virtual work for this
problem is Mp (θ1+θ2), which will be equal to Mp*δ/a plus δ/b where we say it is Mp δ
(a+b)/ab.

So, we should equate this Wδ = Mp δ (a+b)/ab, so δ goes away. So, Wc is Mp(a+b)/ab which
is same as we have here. So, we are able to estimate quickly the collapse loads using both the
theorems without iteration, because in both the assumed mechanisms or the bending moment
diagram the critical sections are easily identifiable and we are able to mark the location of
plastic hinges conveniently without any ambiguity.

The confusion will come only when it is either a partial collapse mechanism or over collapse
mechanism.

(Refer Slide Time: 22:35)

Let us quickly find out another example, where a simply supported beam with central
concentrated load. So, I have a simply supported beam with a central concentrated load W
over a span of l comma E I. So, we know the number of plastic hinges required for this

402
problem is 1 because the degree of indeterminacy is 0 and that will be forming the hinge at
this point.

Let us draw the statically admissible bending moment diagram which will be this and this
value is Wl/4. So, let us quickly find out what will be Mp? Mp will be Wl/4 therefore, Wc=4
Mp/l. Let us do the same problem using kinematic theorem, let us draw a mechanism. We
know that the hinge will form here and here at these two locations, there will be structural
hinges or let us draw it with the other way let us do it with the other way.

(Refer Slide Time: 24:32)

Here there will be structural hinges and here there will be plastic hinge. So, this is θ this is 2 θ
this is θ I call this as δ and the load is W. So, the external virtual work will be W into δ the
internal virtual work will be Mp into 2 θ which is at C, so this is A B and C. We also know θ
is δ/(l/2), so θ is 2 δ/l. So, equating external virtual work to internal virtual work Wδ is
Mp(2θ) which will be Mp* 2δ/l, so δ goes away. So, that is 4Mp/l which is exactly same as
this, so that is Wc.

So, these are simple problems where we are able to find out the collapse load straight away
without any iterations. Now let us quickly compare with this understanding the plastic and
elastic analysis.

403
(Refer Slide Time: 26:16)

Let us say what are the advantages and disadvantages of plastic analysis.

1. The plastic analysis enables effective utilization of the entire cross section of the
member by completely plasticizing the section.
2. Plastic analysis increases the load capacity of the system.
3. Material strength is well utilized.
4. Factor of safety is enhanced by the shape factor.

What are the demerits of plastic analysis? The disadvantages could be the member will be
subjected to excessive deformation. In fact, it is subjected to plastic deformation. since this
method demands redistribution of moments between the critical sections, This method is
useful only to statically indeterminate structures.

Of course it is applicable to form dominant design structures as well. The system undergoes
excessive deformation.

404
(Refer Slide Time: 29:25)

Let us quickly compare the plastic and elastic analysis. Let us say an elastic analysis what do
we do is, we first satisfy something called equilibrium condition. It states that the structure
under any load combination should remain in static equilibrium.

The second is compatibility condition, which states that deformation of different fibers in a
given cross section should be compatible with each other to deform freely. The third
condition is called limit stress condition, this condition states that the maximum stress in the
extreme fiber should not cross yield stress. Relatively what are the equivalent conditions in
plastic analysis, the first condition equivalent to this is mechanism condition.

According to this condition ultimate load or collapse load is reached the mechanism is
formed. Two, equilibrium condition this states that the structure should remain in equilibrium
with the applied loads even after the formation of mechanism. Third condition is the plastic
moment condition. According to this condition in any fiber at any cross section the developed
stress, should not exceed sigma y and the moment cannot exceed Mp. So, quick comparison
between these two method of analysis.

Having said this let us now look at examples where some tricky iteration is also involved.

405
(Refer Slide Time: 33:32)

Let us do special examples, where obtaining the collapse load is not that easy slightly tricky
we will take an example 1. Let us say find the true collapse load of a proper cantilever
subjected to or under uniform distributed load . Let us take the example where static theorem
is applied we will do a static theorem now. So, let us take a cantilever which is propped at
one end subjected to uniform distributed load W per unit length. Let us say the span of the
beam is l and has got EI properties.

Let us call this section as A and this is B. Now we really do not know where will be the
maximum moment happening in the span A B. Let us call that section CC and we want to
find this. So, it is that section where the moment is going to become maximum. Obviously, I
think it is very clear for all of us that CC cannot be at the mid span because one end is
propped either end is fixed. So, there will be an uneven distribution of the critical section it
cannot happen at the mid span.

So, by the way let us see what is the degree of indeterminacy of this problem, neglecting
axial deformation. So, let us say there are two reactions at A and there is one reaction at B is
it not. So now, the total is 3 equations of static equilibrium is 2, so I need the number of
plastic hinges to be 2. They can form at A and they can form at C. They cannot form at B ,
because B is already a hinge where the moment is 0.

So, now it is very important to locate the section C, is it not? Then only we can. So now, the
problem is I cannot draw quickly a statically admissible bending moment diagram and locate

406
the hinge because I do not know the distance at which this is going to form let me call that as
x.

(Refer Slide Time: 36:46)

So, let me draw the bending moment diagram. So, we know this is going to be Mp this is
going to be Mp. So, one hinge will anyway form here. So, let us say the other diagram is
here. So, the section is going to be cc which is a distance x from here and there also the net
diagram is going to be Mp, so this is the bending moment diagram.

But the difficulty is what is c or what is x, we do not know this is it not? So, the first job is to
find out that. So, let us say bending moment at c or a section cc. So, this is Mp let us call this
as small y this is l, so use similar triangle. This is l and this is Mp and this is y. So, I want this
and of course, this is x. So, for l it is Mp for x it is y. So, y will be actually equal to Mp into x
by l.

So, the bending moment at a critical section Mcc will be equal to Mp+y which is
Mp+Mp(x/l) which is Mp(1+x/l) which is Mp(l+x)/l. That is my bending moment at cc. Let
us also find out this bending moment from the first principles.

407
(Refer Slide Time: 39:50)

We know Mcc is also equal to Wlx/2 – Wx2/2. So, this reaction is Wl by 2 and this x . So, Wl
by 2 I am taking anti clock as positive. So, which will become Wx(l-x)/2. So now, this
equation 2 we call this as equation 1, let us equate 1 and 2. So, we can easily do that.

So, Mp(l+x)/l should be equal to Wx(l-x)/2. So, I can straight away say Mp is Wxl
(l-x)/2(l+x). Can we say this? Equation 3. Now for this bending moment to be maximum the
first derivative of this with respect to the variable should be 0. Because I want the bending
moment to be maximum here. So, differentiate this, set that to 0 and find x.

(Refer Slide Time: 42:15)

408
You will find x as 0.414 l. Once I know x let me find Mp. So, substitute in equation 3 we get
Mp as which is Wxl (l-x)/2(l+x) which will become 0.086 Wl2. So therefore, wc is now going
to be 11.66 Mp/l2 or wc* l will be 11.66 Mp/l, where this I can say as Wc= 11.66 Mp/l.

So, friends you can see the procedure of finding out collapse load if the system is
complicated and you cannot draw the statically admissible bending moment directly to locate
the plastic hinges then the solution is tricky. So, you know in this case Wc is given by 11.66
Mp/l. And we all know very quickly that Mp= σy * Zp. σy for the material is known, Zp for a
cross section is known, because Zy* shape factor is known.

So, the right hand side equation is known. therefore, Wc can be computed, so collapse load is
estimated . So, look at this example very quickly that how we used the static theorem in a
roundabout manner to estimate the collapse load. In the earlier examples they were straight
away because no such complication was there. I could easily draw the statically admissible
bending moment diagram and mark the required number of plastic hinges, readily at those
cross sections and get the collapse load directly.

But in this case I could locate the section, but I do not know where it will happen. So, I found
out that and I could locate this.

(Refer Slide Time: 44:52)

Let us try to solve the same problem using kinematic theorem.

409
Let us see is it easy for us to use kinematic theorem in this case. So, kinematic theorem says
for a propped cantilever under uniform distributed load over a span of l E I. I have to draw
the mechanism, we know the degree of static indeterminacy in this case let us say this is A
this is B and this is C.

So, we know there are two unknowns at A plus, 1 unknown at B, we are neglecting the axial
deformation, this will be minus number of equations. So, equation of equilibrium will be 2.
So, number of plastic hinges required will be 2 because the degree of indeterminacy is which
is 3 - 2 = 1.

So, any two plastic hinges where will they form? They will form at the fixed support they
will form at C but C location earlier was not known using the static theorem, now I know
this. So, I get hinges here and here. Now, I can draw the deflected profile and draw the
hinges, say this is θ1 this is θ2 this is θ1 +θ2.

So, from the figure we know tan θ1 is θ1 for small deformation is δ/0.586. Then θ2 is δ/0.414
l. So, let us say the external virtual work is half into base into height into Wc, the internal
virtual work is Mp into θ1 at a plus Mp into θ1 plus θ2 at c, which will now Mpθ1 plus Mpθ2.

Let us substitute for θ1, θ2 from here.

(Refer Slide Time: 48:59)

internal virtual work will be 2 Mp( δ/0.586l) plus Mp times of θ2 which is δ/0.414 l , which
becomes 5.828 Mp/l of δ . So, we equating external virtual work to internal virtual work we

410
get Wc *(l δ)/2 is 5.828 Mp δ/l, so δ goes away. So, it will become 11.66 Mp/l2 or this is
small Wc. Wc * l is 11.66 Mp / l which is capital Wc.

So, friends you will see that this value is same as we obtained here, but there is a very
interesting conclusion in this problem we have used static theorem solution to check the
kinematic theorem solution, How can you say this see here I have borrowed this data from
the static theorem.

So, please friends note that all the time these two theorems are not independently applied
some of the problems they are dependent they will be helping each other. So, please
understand that it is important that I must know both the theorems, how to employ solution of
finding both the theorems then only I will be able to successfully find out the collapse code
there is a very standing example we have here. Let us do quickly a design problem and see
how this can be helpful.

(Refer Slide Time: 51:55)

Let us say a fixed beam with eccentric load of magnitude 30 kilo newton over a span of 5
meters is to be analyzed for collapse load. So, the beam is here fixed beam there is a load
which is W which is 30 kilo newton and the load is 2 meter from the end. So, this is 2 meter
from the left end and this is 5 meters.

411
So, now the degree of static indeterminacy is let us say this is A this is B this is C. So, there
are 2 unknowns at A and there are 2 unknowns at B neglecting axial deformation minus
equations of equilibrium. Equations of equilibrium will be 2 that is sigma f y0 and σM 0.

So, therefore, degree of indeterminacy is 4 minus 2= 2. So, the number of plastic hinges
required is 3 where will they form they will form at A, they will form at B, they will form at
C; it is a straight forward solution. Let us draw the deflected profile hinge here hinge here and
hinge here θ1, θ2, θ1 + θ2.

(Refer Slide Time: 54:02)

So, I could say now the external virtual work is W * δ, internal virtual work is (Mp * θ1) A +
Mp * (θ1+ θ2) C + (Mp * θ2) B. So, from the figure θ1, is δ/a which is δ/2 meters and θ2 is δ/b
which is δ/3 meters.

So, external virtual work is W * δ which is equal to internal virtual work which is Mp times
of 2θ1 +2θ2, is it not? So, 2 Mp*δ/2 plus δ/3 is W * δ. So, δ goes away. So, I can find the true
collapse load as 2 Mp* (1/2 plus 1/3) .

For a given cross section if I know the shape factor if you know Z y I can find Mp. So, I can
find the collapse load. So, plastic analysis is very simple using these theorems if they are not
iterative.

412
(Refer Slide Time: 56:04)

So, let us write down the summary what we learned from this lecture. This lecture helped us
to learn example problems to determine the collapse load we have used both static theorem
and kinematic theorem to estimate Wc. We have also seen examples where 1 theorem
supports the other in estimating Wc.

So plastic analysis is simple, easy and here the load is function of Mp where Mp is a function
of material and shape factor it is a geometric property of course, material is also involved that
is interesting. And I am sure you will find more examples in my book and other literature
papers and you will be able to enhance, your interest towards estimating plastic analysis or
doing plastic analysis on collapse load estimates for varieties of problems which we will be
doing furthermore in the coming lectures, but however, I wish that you should do more
examples by taking these case studies or looking at examples illustrated in my textbook and
my reference notes.

Thank you very much and have a good day friends. Bye.

413
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 24
Plastic design -1

(Refer Slide Time: 00:19)

Welcome to the lecture 24. In this we are going to learn more examples on Plastic Analysis
and Design. So, in the last lecture, we learnt two theorems, static and kinematic theorems
which are lower and upper bound theorems respectively, which are useful in computing the
collapse load. We already saw, that collapse load Wc is obtained using either static theorem
or kinematic theorem. Mostly, but we have also seen as example where both theorems are
used.

So, please do not mistake yourself in learning that these theorems are substitute each other.
No, in fact they support each other. So, do not think that there is a substitution of this. I can
only learn only one procedure. I can get rid of the other and I will still be able to get Wc. That
is the first thing we learned.

Second thing we also learnt that Wc is said to be obtained using iterative procedure. But we
have not seen an example so far, how this will be done from an iterative procedure. So, today
we will see some examples where this will be learnt.

414
We have also understood that collapse load is a function of Mp and Mp is known for a given
section and material, because Mp is actually yield stress multiplied by Zp, where Zp a shape
factor of Zy. And for standard sections given in steel handbook especially with steel, Zy is a
known value.

Of course, we have also given you the MATLAB codes. We explain you the procedure for
finding out the shape factor for different standard rolled steel sections. So, I believe that Mp
can be easily computed for a known material for a known cross section which is more or less
identified as a geometric property.

So, now, the collapse load or the load at which this geometry is intended to collapse can be
obtained by these two theorems namely static and kinematic theorems. So, having said this,
let us do few more examples to learn more in detail about the plastic analysis, and also an
example of the plastic design.

(Refer Slide Time: 04:07)

So, we will in this lecture, we say example 1, let us say we have a fixed beam subjected to a
load as shown in the figure is an UDL of 20 kN/m. There is also a point load of 10 kN and
the geometric details are given like this.

So, let us name this end as A and B, and this point as C. Let us quickly ask a question before
we solve this problem. Let us ask what the degree of static indeterminacy of this problem is.
If we neglect axial deformation, then degree of indeterminacy is 4 minus 2 which is 2. So,

415
how many plastic hinges do I need? The number of plastic hinges required to convert this
beam into a mechanism is n + 1 which will be 3.

Now, the question comes where these hinges will fall. For sure these hinges shall form at
support A and B, there is no doubt because they are fixed ends. But we have a doubt whether
this will form at C being a concentrated load or anywhere on the span BC. So, let us try to
examine this.

(Refer Slide Time: 06:50)

So, we know that the beam is subjected to two set of loads. One is a concentrated load of 10
kilo newton, then an UDL of 20 kN/m for a span of 1 meter and, 3 meters, respectively.

Let us use kinematic theorem. So, for kinematic theorem I must draw the plastic deformation
of this beam. Let us do that. So, we will assume a hinge here for sure, we will assume a hinge
here for sure, I will also have a hinge here for sure. You may wonder that for a uniform
distributor load span, I am drawing a straight line.

Friends, this is not the bending moment diagram. This is the deflected shape of the beam. So,
now, let us mark these rotations. Let us call θ1. If you call this as θ2 obviously, this will be
θ1+θ2 by simple mathematics. And let me call this vertical displacement as δ. Then, let us
write down the external virtual work.

416
There are two loads here, so the point load multiplied by this displacement. When there is a
UDL, you have to find the area. So, half into base into height into intensity of the load, so
which will give me 40 δ.

So, second one is internal virtual work. We have assumed plastic hinges at 3 locations as
marked in green color. So, this is going to be Mp of θ1 which is at A, plus Mp of θ1 plus θ2
at C, plus Mp of θ2 at B. So, which gives me it is 2 Mp (θ1+θ2).

So, from the figure, I can always relate θ and δ. So, from the figure, we can say for small
displacements tanθ1 is θ1 which is δ/1 and tanθ2 is θ2 which is δ/3.

Let me substitute these θ1 and θ2 as a function of δ in this equation. So, we substituting and
equate, external virtual work to internal virtual work, there is principle of virtual work. I
should say it is 40δ =2 Mp (θ1 + θ2) = 2 Mp (δ/1 + δ/3).

So, now 40δ =2 Mp δ (/ + 1/3). I can find Mp now. So, this matches with the assumptions
what we made in the number of plastic hinges formation and I am able to get my Mp, where I
consider this load will be the subjected load on the system. So, obtaining Mp for the analysis
is plastic design. So, now, I am estimating the demand.

(Refer Slide Time: 12:08)

The demand is Mp. The capacity of this beam is σy into Z p. If we know the cross section, I
can find this, if I know the material strength. I can check, demand should be always lesser
than the capacity for safe design.

417
(Refer Slide Time: 12:59)

Let us do one more example. Let us say we have a portal frame. We say estimate the collapse
load of the portal frame. Let us draw the portal frame. Let us say the length or the span of the
beam is l, and height of the column is 2l, and the column has a capacity Mp, whereas the
beam was also a capacity Mp.

I am marking Mp because I know Mp, because Mp is actually σy into Zp. Zp depends on the
cross section. So, I know Zp, I know the material strength, I know Mp therefore, I can mark
these values. Let us say that the beam and column has the same strength.

Now, this beam or this frame is subjected to two loads, one is lateral load W, other is a
vertical load which is 2W which is eccentric. This distance of the load is l /3 and of course,
this distance is 2l /3. Let us name these joints or nodes, this as A, this as B, this as C, this as
D, and we call this point as E. Let us try to find out the degree of static indeterminacy for this
frame.

So, there are two fixed supports. So, there will be 3 unknowns. So, the degree of
indeterminacy will be 3 reactions at A, plus 3 reactions at D, minus number of equations of
static equilibrium, will say it is 6 minus 3, so 3. So, the number of plastic hinges required to
form or to convert this portal frame into a mechanism will be n +1 which is 4. So, we need 4
plastic hinges.

418
Let us ask a question where these hinges will form. For sure they will form it A and D,
because they are fixed ends. And there is a possibility it can further form at B, C and E. So,
there are 5 possible locations, I need only 4. So, plus recollect that there are different
mechanisms available for analysis like beam mechanism, column mechanism, or sway
mechanism, combined mechanism, partial collapse mechanism, complete collapse
mechanism, over collapse mechanism and so on.

We learnt all of them in the previous lectures, please turn back and learn and understand.
Now friends, this problem is slightly tricky because we have got 5 possible locations
whereas, I need only 4. So, we will use kinematic theorem to solve this problem. The
kinematic theorem stays I must assume a collapse mechanism first.

(Refer Slide Time: 17:30)

Let us assume a beam mechanism which is going to be on the segment A, E, C or on the

segment BC, on the beam BC. Let us draw that beam separately. So, there is a load 2W here,
this is l /3, this is 2l /3. This is my deflected shape of the end B. So, let us assume that I have
3 hinges formed here, so it is a partial collapse mechanism. Why a partial collapse
mechanism? So, for a complete collapse mechanism I need 4 hinges, I assume only 3.

Let me redraw this figure with this load 2W. I call this as A and this as B. So, let us say I will
have hinges here. I can call this angle is θ1, this angle as θ2, and this displacement as δ and
this angle will be θ1 plus θ2.

419
Let us write down the external virtual work which will be 2W *δ. The internal virtual work
will be the work done by the plastic hinges. So, I have one plastic hinge formed at B, I have
one plastic hinge at C.

(Refer Slide Time: 19:51)

So, this multiplied with, ; let me Mp*θ1 this is at B, plus Mp*θ2 this is at C, plus Mp*θ1 plus
θ2 this is at E, so which becomes 2Mp*θ1 plus θ. So, from the figure we also know θ1 is δ/a
and θ2 is δ/b. So, let us equate external virtual work to internal virtual work, which is 2Wδ
should be equal to 2Mp times of δ (1/a + 1/b), which will be 2 Mpδ(a+b)/ab. Can I say a + b
as l from the figure? So, now, 2Wδ is 2 Mp δ l / a b. Therefore, I can say now the collapse
load W is Mp l / ab.

But I cannot guarantee is the final answer of the collapse load, because this has been obtained
based on the partial collapse mechanism. I will looking only at the beam mechanism, right.
There are other mechanisms possible.

Let us extend this further let us say a is l /3 and b is 2l /3, I can quickly get Wc as 9 Mp/2l.
That is my answer. I call this equation number 1.

420
(Refer Slide Time: 22:30)

Equation 1 cannot confirm the collapse load for two reasons. One, it is based on a partial
collapse mechanism. Two, it is based only on a beam mechanism. There are other
mechanisms possible. Let us look into the second option which is a sway mechanism.

(Refer Slide Time: 23:13)

Let us draw the deflected profile of the frame in the sway mode. So, I assume the plastic
hinges are at A, B, C and D. That is I think I can mark them here. Let me call this angle as α
which is also α, let me call the displacement as delta. Let us write down the geometry this is l
and this is 2l. And this end is A, this is B, this is C and this is D.

421
There is a load applied which is W here; and of course, there is a load applied here which is
2W. Let us write down the external virtual work, which is going to be only W into delta
because this will not do any work as it has not undergone any displacement.

And now let us estimate the internal virtual work. So, one work done is here it is rotated,
other work done is here is rotated, the angle between the normal’s should be same, but you
know there is only a stick to sway here. The frame has just to swayed. Therefore, I can now
say the internal work virtual work is Mp*α at A plus Mp into α at D. So, can I say this as 2
Mp α?

So, external virtual work equated to internal virtual work W into Δ should be 2 Mp α, but α
and Δ are related. You can write down the relationship from the figure. We know α for small
angles is Δ by 2l. So, substituting this in the above equation we get W into Δ is 2 Mp times of
Δ by 2l.

So, therefore, Wc is; there will be work done at sections B and C which is again Mp α plus
Mpα at C. So, let us say this is also added to this. So, let me put this as 4 Mp. So, this
becomes 4 MpΔ /2l, so its 2Mp/l. So, friends, one value gives me collapse load as 9Mp /2l
which is about 4.5Mp /l, the other one gives me it is 2 Mp/l. So, now, second equation.

(Refer Slide Time: 27:26)

422
The collapse load given by second equation is again only based on sway mechanism. But it is
a complete mechanism because there are 4 hinges, but still there is a question whether the
hinge will form at E or B or C. See the possible locations are A, D, B, C, and E.

We have 5 locations, but we need only 4 out of which A and D are sure because they are
fixed supports. But out of B, C and E, we have assumed that B and E, but we left out C
because there is a question. So, this is also may not be the correct answer.

(Refer Slide Time: 28:52)

Let us do one more attempt, one more iteration to find out using a combined mechanism. Let
us draw the frame and let us draw the combined mechanism including the sway and the
beam. Let us assume the hinges of course, at A and D for sure. Let us assume the hinge here
and here. We will not assume hinge here.

So, we are assuming hinges at A, D, C and E. So, there is a point load of 2W here, there is a
lateral load W. So, we call this value as α which is also α here, also α, is also α, is also α. And
we call this as Δ and we call this as δ. And we call this angle as θ1 and this angle as θ2, let us
not put this α.

Let us write down the external virtual work. The external virtual work will be 2W * Δ that is
by the vertical load, 2 W * δ + W * Δ that is the horizontal load. Let us write down the
internal virtual work, we have 4 hinges,

423
So, we should say Mp * α which is at A, plus M p Say, this is θ1 plus θ2, θ1 plus θ2 is going
to occur at E, plus Mp into α plus θ2 plus this is at C, plus Mp into α at D. So, which can be
said as Mp times of 3 α plus θ1 plus 2 θ2, the internal virtual work.

Now, I can find the relationship between Δ and α. So, α is Δ/2l. So, I can replace α in terms of
Δ.

Let us also find out relationship between θ and δ and θ2. That is θ1 is δ by l by 2 which is 2 δ
by l and θ2 is also δ by l by 2 which is 2 δ by l. So, we will substitute this. So, we have θ1, we
have θ2. and we have α and δ. So, we have equations in terms of δ and Δ. So, we now we will
replace this in δ and Δ. So, let us do that. We know α is Δ by 2 l.

(Refer Slide Time: 34:44)

So, I should say Δ is 2 α l. And we have θ1 as 2 δ by l and θ2 also as 2 δ by l. So, let us write
down external virtual work which is equal to 2 W δ plus W Δ. Internal virtual work which
was this one, Mp times of 3 α plus θ1 plus 2 θ2. Mp times of 3 α plus θ1 plus 2 θ2. Let us
substitute them back.

We will equate external virtual work to internal virtual work. So, 2 W δ plus W Δ should be
Mp times of 3 α plus θ1 plus 2 θ2. So, let us express everything. So, let us say W times of 2 δ
plus 2 α l which will be equal to Mp times of 3 α plus θ1 plus 2 θ2.

424
So, in this figure if you look at friends, I can say α and θ2. and α and θ1 because the angle
between the normal has got to be same. So, using that condition I can say α should be
equivalent to θ1 to θ2. because angle between the normal’s is same. That is in the displaced
portion we consider this angle as still 90 degrees.

We simplify this and try to find Wc. We will give you the answer in the next lecture, what
would be this value. Please work on this, and check up and tell me what is the Wc from the
combined mechanism, and therefore, what is the true collapse load of this.

Thank you very much. Have a good day. Bye.

425
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 25
Plastic design -2

(Refer Slide Time: 00:20)

Friends, welcome to the 25th lecture on Advanced Steel Design where we will continue to
discuss some more examples on plastic analysis and of course, application of analysis in
design perspective I will put this II because it will be a continuation of the last lecture what
we discussed. In the last lecture we are trying to apply kinematic theorem for a frame
example to find the true collapse load.

(Refer Slide Time: 01:14)

426
(Refer Slide Time: 01:36)

This was the problem where a frame is subjected to a lateral load of W and a vertical load of
2W at an eccentric portion of this. So, we computed from the beam mechanism that may
collapse load is 9Mp by 2l.

(Refer Slide Time: 01:42)

427
(Refer Slide Time: 01:45)

(Refer Slide Time: 01:52)

428
We also did the sway mechanism where we said the collapse load is 2Mp/l. We further went
down to understand there are 5 possible locations where the hinges can form, but we need
only 4 for a complete mechanism. So, we wanted to be sure using iteration procedure where
this hinge will form. For A and D they are sure because they are fixed supports. So, now the
hinge can form either B C or E. So, we have a doubt. So, we are going for an iteration to find
out the true collapse load.

(Refer Slide Time: 02:30)

We started doing this and we said that the hinges will form at A, E, C and D it is an
assumption, the hinges will form at A, D, C it is an assumption we are assuming this. The

429
external virtual work is force*displacement. So, 2 W*𝛿 and W*Δ the internal virtual work is
Mp of α which is at A. So, that we can check it Mp*α which is at A then again there is a
hinge here at E.

So, Mp*𝜃1and 𝜃2, this is at E, then again at C you got 2; 1 is Mp for α other is Mp for 𝜃2 and
we do know that these inclinations will not be equal because the load is eccentric it is not at
the concentric position of the beam. Then the fourth 1 is Mp α at D. So, this gives me an
internal virtual work let us say this is equation number 1 we will take this 2.

So, we want to equate 1 and 2 and find out the true collapse load, but there is a problem here
equation 1 has got 2 variables 𝛿 and Δ and equation 2 has got different variables α 𝜃1and 𝜃2,
this is the issue what we had. So, then we started working out from the figure α will be Δ by
2l this is very simple straight forward and 𝜃1and 𝜃2 need to be found out. So, now, 𝜃1cannot
be this way. So, let us try to correct this and go back.

(Refer Slide Time: 04:40)

So, we will simply say that the correction is going to be there for this case and let us do this.

(Refer Slide Time: 04:56)

430
So, now α is Δ by 2l, 𝜃1is 3𝛿/l and 𝜃2 with 3𝛿/2l from the geometry. So, let us call this
equation number 3. So, from this equation we can also say trying to look at 𝜃1and 𝜃2 we can
straight away say 𝜃2 is 𝜃1by 2 is it not? Because 3 𝛿/l is 𝜃1. So, 𝜃2is 𝜃1/2 . Now, for the
frame to sway towards right and maintain this angle as 90 if this is α and this is 𝜃1. The
frame has to sway towards the right.

So, let us assume that the sway Δ causes symmetric rotation at B what does it mean this is Δ.
So, we say now 𝜃1is α. So, we maintain its angle 90 degree. So, this is true. Now, we already
know the internal virtual work from equation 2, let us copy that equation. So, internal virtual
work is Mp times of 3 α + 𝜃1 + 2 𝜃2, let us see this equation.

So, which will be Mp times of α is here. So, 3 times of Δ by 2l plus 𝜃1is already α. So, can I
say this is α plus let us say 𝜃2 is 𝜃1/2 which will become Mp times of 3 Δ by 2l plus α plus
𝜃1, which I can now say is Mp times of 3 Δ by 2l plus α plus α, because we have this
equation here.

(Refer Slide Time: 08:16)

431
So, now I can say internal virtual work is Mp*(3Δ/2l) + 2 α which is Mp * (3Δ/2l + 2Δ/2l )
which is Mp *5Δ/2l, equation number 4. I have modified the internal virtual work in a
simplified form. Look at the external virtual work there are two variables, but we know they
are connected to each other.

So, let us say external virtual work which was 2 W*𝛿 plus W*Δ which can be said as,
2W(𝜃1l/3). Plus W*2lα, which will now become 2W𝜃1l /3 + W(2l𝜃1) because we know α and
𝜃1 are same.

So, can you simplify this as W*2𝜃1l /3 + 2l𝜃1 which will become W𝜃1*8l/3. So, I simplified
this equation further in this form W𝜃1*8l/3. So, now, you see I have internal virtual work in
terms of Δ, but here I have this in terms of 𝜃 1.

(Refer Slide Time: 11:30)

432
Let us convert this which now become W𝜃1/3 (8l) which can be said as W*8l α /3 you can see
here 𝜃1is α. Now α and Δ are connected, by this equation. So, let us do that which will now
become (W/3) *8l (Δ/2l).

So, l goes away and this becomes 4. So, that becomes 4WΔ/3, that is my external virtual
work. So, now, you see external virtual work is in Δ internal virtual work out in Δ let us
compare this. So, external virtual work should be equal to internal virtual work. So, 4WΔ/3
should be equal to Mp*5Δ/2l. So, now Δ goes away, I can state to be find W, as Mp*15/8l.

So, now, we have different values. We have W1 which is from the beam mechanism which
was equal 9Mp/2l which is 4.5 Mp/l. We have W2 from the sway mechanism which is here
which is 2Mp/l. We have now W3 is a combined mechanism which says it is 15Mp/8l. So, we
must choose the least of all. So, now, the true collapse load will be 1.875Mp/l which is given
by here.

(Refer Slide Time: 15:17)

433
So, I have a frame loaded by lateral and vertical these are the points A, B, C, D and E this
was the mechanism what we had which gave us the true loads. So, we have hinges at here, at
here, here and here is it not. So, now we say there is no hinge at B. So, M at B is lesser than
Mp.

So, we have to prove this, because we do not have a hinge at B. So, let us draw the bending
moment diagram on the tension side. So, this is Mp and again this is MB this is MB and this is
of course, Mp and Mp and Mp I am drawing B M D on the tension side. So, I am just joining
this.

So, I want to find MB. So, if you look at this MB will be actually equal to (W/2)*2l - Mp. Now
if I say this W is my Wc, let us say Wl - Mp which is 1.875 Mp by l*l minus Mp. So, you get
minus 0.875 Mp, negative is due to a sagging moment and it is lesser than Mp.

So, no hinge is formed at B, hinge will only form when M is equal to Mp.

(Refer Slide Time: 18:24)

434
Let us say, we have an un symmetric frame both ends fixed, there is a later load W and there
is a central concentrated load 2W, which is at the center and the capacities of columns and
beams are same. So, let us say this height of the column is l and the span of the beam is also l
and the height of the other column is twice of this. So, the question is found out the true
collapse load. So, let us quickly see the degree of static indeterminacy for this problem
neglecting axial deformation will be 4 - 2 which will be 2 because we neglected the axial
deformation.

So, this is A, this is B, this is C, this is D and this is E. So, we need number of plastic hinges
as 3 to make it as a mechanism. So, we will start with the beam mechanism. So, we assume
hinge at B, E and C. So, it is a beam mechanism. So, there is a hinge here, here and here. So,
this is 2W and this is 𝜃 and 𝜃 and this is 2𝜃 and of course, this displacement is 𝛿.

So, the external virtual work in this case is 2 W*𝛿, the internal virtual work in this case will
be Mp*𝜃 this is at B plus Mp*2 𝜃 this is at E plus Mp*𝜃 this at C which makes it as 4 Mp 𝜃.

(Refer Slide Time: 21:08)

435
So, equating external virtual work to internal virtual work 2W𝛿 should be equal to 4 Mp𝜃 and
we also know from the figure that 𝜃 is 𝛿/l. The span of the beam is 2l. So, 𝛿/l is 𝜃. So,
substituting 2 W*𝛿 should be 4 Mp*𝛿/l. So, Wc in this case is going to be 2 Mp/l the first
iteration.

(Refer Slide Time: 22:23)

Let us go for the next one which is going to be a sway mechanism. So, the frame is going to
sway to the right this point is A, B, C, D and somewhere it is E there is a load here and now
there is a lateral load here W. If I call this as Δ and this as α this is also going to be α and this
is β this is going to be β and we know this is 2l this is also 2l.

436
So, let us assume hinges at A, B, C and D. So, now, the external virtual work is going to be
equal to W*Δ the internal virtual work is going to be equal to Mp*α this is at A, plus
Mp*again α at B, plus Mp*β this is at C, plus Mp*β at D. So, which makes Mp*(2α+2β). So,
we know external virtual work should be equal to internal virtual work by principle of virtual
work.

So, WΔ should be equal to Mp*(2α+2β). Now we know from the figure α is Δ/l and β is Δ/2l
let us substitute that here. So, W*Δ is going to be 2Mp(Δ/l + Δ/2l) , so which will become 3
Mp /l. So, now Δ goes away W c is 3 Mp /l .

(Refer Slide Time: 25:25)

In the first iteration we got 2 Mp /l now, we have 3 Mp /l let us also do a combined

mechanism; Let us say this is my combined mechanism I am going to assume hinges at this
place and this place here and here. Let us say this is α and this is 𝜃, this is 2 𝜃 because this is
also 𝜃 and this is β and this is β.

So, the loads are this is W and here there is a vertical load of 2W. So, the external virtual was
going to be 2 W*𝛿 + W*Δ. internal virtual work is going to be, Mp*α this is at A, plus Mp*2
𝜃 that is at E, plus Mp*𝜃 that is at C, plus Mp*β that is at C, plus Mp*β that is at D.

So, one can say Mp(α+3𝜃+2β). Now, keeping this angle as 90, for sway to happen α should
be 𝜃. So, therefore, internal virtual work will become Mp(4α+2β) and we already know the
value of α and β, let us substitute and get the true collapse load.

437
(Refer Slide Time: 28:38)

So, now I can say external virtual work should be equal to internal virtual work. Now see
here α is equal to 𝜃. So, where α is Δ /l whereas, 𝜃 is also 𝛿 /l. So, I can say these two are
identical. So, the external virtual work is WΔ + 2W𝛿 which can be written as 3 W Δ that is
external virtual work. internal virtual work is going to be Mp*(4α+2β) and α is 𝛿 /l and this is
𝛿 /2l.

So, Mp *(4Δ /l + 2Δ /2l) Which makes this as 5 Mp Δ/l. So, equating 3 W Δ is 5 Mp Δ/l. So,
I get Wc as 5 Mp/3l. So, now, this is the third one I have which is about 1.67. So, now, I have
this is 3, this is 2 and this is 1.67 So, the true collapse load is the least of all which is 5 Mp/3l
and for a given cross section we already know Mp, I can find Wc.

So, friends these are some examples where we have obtained very easily the collapse load for
different mechanisms combine them then we got a test of how to get the iterative mechanism
about this and get the collapse load.

438
(Refer Slide Time: 31:17)

So, as a summary in this lecture we learnt more examples of plastic analysis, we will also do
one design application very quickly now.

(Refer Slide Time: 31:41)

Let us do a design application, say I have a simply supported beam of size 230 by 450 in
cross section. The beam is fully loaded till the central section is completely plasticized the
beam is subjected to a central concentrated load. What is want is demarcate the yield region.

(Refer Slide Time: 32:59)

439
So, the beam is like this the simply supported beam. Central load for a span of 5 meters the
cross section is a rectangle of size 230 by 450. Take yield value as 250 MPa. we know Mp is
going to be equal to Z p*𝛔y which will be shape factor*Zy*𝛔y which is (1.5*230*4502
/6)*250 which becomes 2.9*103 kNm for this problem, by the way Wc the collapse load will
be 4 Mp /l.

So, if I know Mp, I can easily find the collapse load and so, on which will be 2.32 MN that is
the load at which the beam will collapse. Let us also find out what is the yield moment of
this.

(Refer Slide Time: 34:53)

440
So, step number ii let us try to find out the yield moment which will be 𝛔y *z y which will be
250*1/6 of 230*4502 which will be 1.95*103 kNm. Let us draw the demarcated region. This
is my beam this is my bending moment diagram this is my deflected profile. Here the
moment is 0, at this point it is Mp, at some section this value is My from a similar triangle we
can find this.

The span of the beam is 5 meters. So, we can say for 2.5 meter this is Mp, therefore, for x it is
My, I know Mp and My, I can find x which will be My by Mp of 2.5, M y is 1.95 and Mp is
2.9*2.5. I get x as 1.67 meters this is 1.67 meters. So, only this portion is trying to get yielded
remaining this portion is still elastic, because up to this point it is My.

(Refer Slide Time: 36:59)

So, now let us say this is my beam in a deflected profile is the bending moment diagram as
well this, this is Mp here, this is My here and this distance is 1.67 is also 1.67 by symmetry.
So, I can say M at 1.67 is My which is 1.95*103kNm. Can I find M at 1.8 from the similar
triangle principle? Which will be (1.8/2.5)Mp.

I can find this 1.8 /2.5*2.9x103. So, I can know this value as 2.088 x 103 kilo newton meter.
So, can I this way find M at 1.9, M at 2.0, M at 2.1, 2.2 and so, on at M at 2.5 will be equal to
Mp, which will be 2.9x103 kilonewton and meter. So, I have moment at different sections I
have moment at different sections I have computed them.

(Refer Slide Time: 39:09)

441
So, let us also find the depth of elastic core. We know the equation M is equal to Mp
*(1-e2/3h2). So, I can say M /Mp is *(1-e2/3h2) by this logic e2 will become (1- M/Mp)*3h2.
So, it will have a parabolic variation as e2.

(Refer Slide Time: 39:58)

So, can you now find what is the depth of elastic core at 1.67? How do you do that? It is 1
minus M at 1.67 by Mp*3 h2, h is the depth of the section. So, it will be 1 minus M at 1.67 is
1.95*103 you can see here 1.95 x103 by Mp is 2.9x103 3 h2 which is 4502. So, I can find very
easily this will be 450mm which is completely elastic at 1.67, see here at this section it is
fully elastic is it not?

442
That is what we are getting similarly can we find e at 1.8, e at 1.9, e at 2, how can you do
that? Because at all these sections I have moment at 1.8 moment at 1.9 I have here seen here
1.8 1.92 I have all the moments you can substitute here and I can find the depth of elastic
section here.

(Refer Slide Time: 41:09)

So, if you try to plot that the plot will look like this if this is my depth of the section if this is
the center and this is 1.67. So, it is 1.67 from here and it is 1.67 from here and you see the
hinge formation will be like this. So, this is how it is plasticized and at the mid span it is fully
plastic it is Mp.

(Refer Slide Time: 41:54)

443
So, friends this is how we have demarcated the yield region. So, we have also learnt how to
demarcate the yield region and we learnt how to find out Wc by iteration. So, friends, in this
lecture we learnt about the analysis and design in this module we learnt about plastic analysis
and design in detail now we will move on to the second module the third module which is on
stability of members.

Thank you very much have a good day.

444
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 26
Structural Stability

(Refer Slide Time: 00:22)

So, friends welcome to the 26th Lecture which will be on Stability of Structural systems. It is
a new module stability lecture 1 we are talking about stability of systems. Stability of a
system in general and to steel structures in particular is extremely important because most of
the complaint offshore structures encounter the loads based upon the geometric form and not
by strength.

So, now when we talk about complaint systems which are innovative in geometry they have a
special characteristic they counteract the loads not by material strength alone, a major a major
contribution comes from the geometry. So, form dominant, but we have got to impose a very
important condition that under any circumstance the geometric form should be stable.

So, the stability of the system cannot be challenged you can use a form dominant design, you
can try an innovative geometric form, you can derive the benefit of indeterminacy and
arrangement of members to support the system and resist the load all are fine. Provided one

445
condition that the geometric form what you design should remain stable. So, the stability
becomes a very important criteria.

(Refer Slide Time: 02:35)

So, we say here that stability is an important criteria in design, this is much more important in
form dominant design let us put it this way, because there are challenges. We also know that
the structural form is subjected to variety of challenges in given loads and boundary
conditions. The structural configuration or the geometric form should satisfy the conditions
of static equilibrium, compatibility relationship and force displacement relationships. It
should also remain stable were nicely said by Presley et al in 2007 now these are conditions
which are to be met.

(Refer Slide Time: 04:34)

446
Therefore, friends an unstable condition is referred to as a failed state. So, unstable condition
is a failure therefore, stability understands conditions are failure. Furthermore, stability is
affected significantly under compressive forces to be very precise under axial compressive
forces. Therefore friends the geometric form or the geometric stability is more important than
the material strength, you can use any material of a higher order in terms of its strength.

If the geometry is not able to sustain a stable configuration equilibrium condition the
geometry cannot be used to resist the applied loads. Therefore, geometric stability challenges
the functional requirements of any member.

(Refer Slide Time: 06:52)

447
Therefore stability is always referred as stable state of equilibrium; How do you define
stability? Stability is defined as the ability to remain stable. stability the ability of any
structural system to remain or continue to remain in its stable geometric form which can
perform the intended function even if the geometric position is disturbed by external forces.
So, we can say it is ability to remain stable-stability.

By this definition friends complaint offshore structures should remain stable because they
perform their functions of disbursing the loads only by counteracting the loads with the
novelty of the geometry. Say we have studied about an example of a complaint platform
which is tension like platform; a tension like platform in it is disturbed position tries to reach
recentering.

So, as long as the change of the geometric position does not affect its loads dispersing
capacity we can say the structure is instable equilibrium. Therefore, structures do not need to
remain in their original geometry to classify them as stable that is a very important statement.

(Refer Slide Time: 09:43)

Friends, structures do not require to stay in their original geometric form to classify them as
stable. They can remain stable even under the displaced position that is very important. So,
form dominant structures can remain to stay in the displaced position and as long as they are
able to disperse the loads and perform intended function they are considered to be stable.

448
So, please do not get confused that stability refers to an static equilibrium position. No, it can
be a deformed position of geometry I am saying of geometry can be a deformed position. As
long as under the deformed position if the structure is able to perform intended functions
successfully we can call that geometric form or that structure as a stable system. So, stability
is always referred to performance index.

(Refer Slide Time: 11:30)

Stability refers to performance or functional characteristics. You may ask me a question, what
are the functions of any given structural system? The foremost function is the structural
system should disperse the loads efficiently without failure is it not, that is the intended
function first. Then comes the question of it is going to stay in equilibrium, it is going to
cover the roof, it is going to cover the walls, what are may be the functional properties come
later.

The foremost requirement of structural system is to disperse the loads acting on the structure
efficiently without failure. Having said this there are three criteria of checking stability.

(Refer Slide Time: 12:37)

449
(1) Euler’s static criteria, (2) Liapunov’s dynamic criteria and the last one is (3) the potential
energy stability criteria. Let us see one by one very quickly what are these?

(Refer Slide Time: 13:31)

Let us talk about Euler’s static criteria. Euler’s static criteria is applicable under non trivial
equilibrium state. It evaluates the stability of a structural system by examining the optimum
geometric configuration of the system other than the original configuration. So, it examines
the geometric configuration other than the original configuration. So, what is the new system
it is examining?

450
It is examining the geometric configuration or it examines the changed geometric
configuration and check whether the structural system can disburse the load, the applied load
without failure. The applied load in this case is called 𝑃𝑐𝑟 it is called as a critical load we will

talk about that later.

(Refer Slide Time: 15:35)

While examining this, what are the factors which are important; (1) boundary conditions, (2)
initially perfect, straight geometry. Remember; when no load is applied when the structure is
not enabling itself to disperse the load by its form dominant characteristic structure has to
remain initially straight and perfectly elastic is it not, that is the basic design requirement we
have. So, initially the system should be perfectly straight geometry and if it is not so that will
affect the stability criteria as per Euler’s condition.

In addition to that the boundary conditions applied on the system will also be examined under
Euler’s criteria. So, therefore, friends Euler’s criterion evaluates the structure evaluates
whether the structure is capable of carrying load (P), remember this P can be even less than or
equal to. Let us put it this way capable of carrying a load (𝑃 > 𝑃𝑐𝑟) let us put it like this.

So, let us say we are looking at the failure condition instead of remaining in a state where it is
unable to carry any load at all. Friends, it is very interesting to see that the load carrying
capacity of the member is reduced from 𝑃 to 𝑃𝑐𝑟.

451
(Refer Slide Time: 18:30)

Therefore, the load capacity of the member is reduced from P to 𝑃𝑐𝑟. So, the condition is P

exceeds 𝑃𝑐𝑟 that is the original capacity, but because of the change in geometry because of the

incapability of the member we arrive at a load which is 𝑃𝑐𝑟 and this condition holds good. So,

there is a reduction in load capacity which is estimated which is computed using Euler’s
criteria. But still, it is an important fact that the member should at least carry 𝑃𝑐𝑟 even at the

disturbed geometric condition which is quite weak from the original form.

So, please understand friends, the load capacity at the changed or disturbed geometric form is
not 0. The load capacity at the changed disturbed geometric form is 𝑃𝑐𝑟 and we believe that

𝑃𝑐𝑟 is much lower than P the original capacity. But still the load capacity is not 0, it is having

some critical load carrying capacity that is what is estimated using Euler’s criteria.

(Refer Slide Time: 20:49)

452
So, this value of the load 𝑃𝑐𝑟 which is reduced magnitude in comparison to the original

capacity of load is termed as critical load or buckling to be very precise is called as Euler’s
buckling load.

(Refer Slide Time: 21:51)

So, it can be easily seen that 𝑃𝑐𝑟 is dependent on boundary conditions cross sectional

properties of the member and slenderness ratio. I believe all of us know how to estimate a
slenderness ratio for a given structural member. Furthermore, instability occurs when two or
more adjacent equilibrium positions correspond to different mode shapes. That is why friends

453
you note; in well-designed structural systems there exist distinctly different well separated
mode shapes.

On the other hand the frequencies of the system will not be closer they will be well separated.
So, the mode shapes do not overlap very closely with each other because if they have very
closely related condition there can be an overlap and that can cause instability. So, when you
talk about stability of offshore structures to be more interesting because we have form
dominant design well practice in offshore structures therefore, we have to learn about this in
detail.

We should at least know about them the advanced steel design is not meant only for offshore
structural applications, form dominance in existing for many years for example, arches,
domes, cable stay bridges these are all statically indeterminate systems they are all form
dominant.

(Refer Slide Time: 24:33)

But when we assess stability of offshore complaint platforms it disagrees with Euler’s
criterion. So, Euler’s criterion cannot be applied to a complaint platform, maybe why? Let us
ask this question why? Why because, the boundary conditions of the member do not permit
examination of Euler criterion. So, the problem it is the boundary conditions, because we are
talking about floating systems, we are talking about floating structures anchored to the seabed
using pre tensioned axial tethers.

454
So, such boundary conditions cannot be implemented in Euler’s criterion, because they
completely float under hydrodynamic stability what we call as compliancy. So, in that case
stability is largely dependent on the pre-axial tension in their tethers. Then how do you ensure
stability in such cases when Euler’s criteria is not applicable? We ensure stability in such
cases through recentering.

Friends, please recollect we discussed this property of recentering, recentering is an ability of

the structural system to return back or regain its original geometric position inherently in the
presence of load. So, that is exactly the condition what we discuss in stabilities, stability
analysis is the capability of the structural system to assess its condition of disbursing loads
and performing the intended function at the displaced geometric configuration.

It is well understood, well established, it is very well clear that the load capacity of disturbed
geometric form will be far lesser than the original load capacity there is no question about
that. The disturbed geometric form can always carry only a lesser load than its original
geometric form which is initially straight perfectly elastic etcetera. But what we want to
emphasize here is if that load capacity at the change geometric form is 𝑃𝑐𝑟 we call that as a

stable condition that geometric capacity or the load capacity at the intended geometric form is
not 0, it has some capacity.

We all agree that their capacity is much lesser than the original load carrying capacity of the
structural system in its original geometric shape and size and position agreed. But the new
geometric shape size and position is having some load capacity which is not 0 it is that
condition which is called as the stable condition.

455
(Refer Slide Time: 28:44)

Let us talk about Lyapunov stability; Lyapunov’s condition examines the stability of the
structural system under dynamic excitations. If a member is subjected to continuously
varying disturbing force, it is necessary to examine whether equilibrium under the dynamic
force is satisfied. So, when P is function of time, it is necessary to check its equilibrium or
stability condition during the excitation process, as seen in the literature stability can be
influenced by two ways.

(Refer Slide Time: 30:21)

456
Stability of systems under dynamic loads can be influenced by two factors one is the
amplitude I should say varying amplitude, second is the period of excitation. These two
criteria can influence the stability assessment of a structural system under dynamic
conditions. So, the amplitude can influence the load capacity directly. So, load capacity is
strength dependent parameter we all know that whereas, the period of excitation can
influence an unconditional response at near resonance state.

Hence, Lyapunov’s assessment of stability is focused more towards dynamic response

behavior of the system.

(Refer Slide Time: 32:57)

So, let us quickly compare Euler and Lyapunov, Lyapunov system is stability assessment
under dynamic loads. So, it is focused more towards the response behavior whereas, Euler’s
condition is purely to assess the load capacity. So, that is the quick comparison here.
Therefore, friends Lyapunov’s condition is more significant for structures designed to
perform intended function under varying or time varying external forces.

The most challenging part of stability of such structures is a geometric form assumed at any
instant of time. So, the greater challenge here geometric form at any instant of time that is we
can say displaced configuration. Structural systems like ships, offshore compliant platforms,
floating production platforms, fall under this category of stability check.

457
The stability calculations of ships for example, focus on estimating the center of gravity,
buoyancy and meta center vessels and their interaction. In case of structures which are form
dominantly designed like TLPs they are permitted to undergo large displacements. So, one
should ensure that under such large displaced condition the geometric design ensures proper
recentering and therefore, the load carrying capacity is not towards close to 0, that is very
important condition.

(Refer Slide Time: 35:45)

The third condition which potential energy stability criteria is applicable to structural systems
for which the systems potential energy ceases to be minimum, this is more applicable to
conservative systems.

458
(Refer Slide Time: 37:18)

So, friends having said and discussed different stability criteria as from the literature, let us
pay more attention towards estimating Euler’s critical load. Euler’s critical load is based on
application to an ideal column. So, here we are talking about the compression member which
is axially loaded and the column is ideal let us say how an ideal column looks like. Let us say
this is my positive x and anticlockwise 90 is positive y anticlockwise 90 is positive y, the
deflector profile is shown here. This is the initial position of the column or initial geometry
this is deflected geometry under load.

Let us quickly see what are the ideal conditions of the column? We say ideal column know,
what are the conditions? The conditions are (1) column is pinned at both the ends, (2) the
column is uniformly slender, (3) the column is assumed to be laterally restrain in position at
both the supports, (4) it has got a negligible weight and (5) it is perfectly elastic. What does it
mean upon removal of this load P the reflected profile of the column will reach its original
geometry, it is not recentering. Please understand, I am talking about the material here.

459
(Refer Slide Time: 40:53)

What does it mean is to be very specific, stresses developed by the axial force are assumed to
be within the proportional limit of the material. So, this is purely material property it is not a
recentering property. Now let us imagine let us consider that this column has a load capacity
𝑃𝑐𝑟. Now, if the applied force or the applied load is lesser than the critical value, the column

will continue to remain straight and undergo only axial compression, under this state the
column is said to be in static equilibrium.

(Refer Slide Time: 43:08)

460
What is this condition? P lesser than equal to 𝑃𝑐𝑟, the column is set to remind in static

equilibrium condition this is called stable equilibrium condition. Therefore friends, under
such condition if the lateral load applied at any point. Please understand, now I am talking
about two loads one is axial load which is causing these ideal conditions and I say we have
worked out a critical load and we have declared that the column is in stable equilibrium.

(Refer Slide Time: 44:22)

Under such condition, if a lateral load is applied at any point, it will cause a large lateral
deflection at the mid height of the column. However, importantly the column will regain and
return to the original position in terms of geometry shape and size of cross section. Upon a
continuous increase of axial load P the straight form of equilibrium tends to gradually
become unstable.

Let us say P is applied to the column, the column has a capacity of 𝑃𝑐𝑟 which is known

imagine that P applied is far lesser than 𝑃𝑐𝑟, the column will be stable no disturbance to the

column. Let us say P is gradually reaching 𝑃𝑐𝑟 column will still be stable and no change.

Remember, upon removal of load column will return to its original geometry. That is very
important that is the condition we are assessing at.

461
(Refer Slide Time: 46:50)

If you continue to increase the load beyond this imagine P has reached 𝑃𝑐𝑟 column is still

stable no worry, but if P is gradually increased more than 𝑃𝑐𝑟, then the column will become

unstable, this is the point of worry. Now, under this condition, even a lateral load of small
magnitude can cause large lateral deflections which will not disappear upon removal of load.

(Refer Slide Time: 48:12)

So, this will be a permanent deformation. Remember, this was not true under these
conditions, under these conditions this was not true, this was true only when this P exceeds

462
the critical capacity. Therefore, friends it is very important for a given column member to
compute this 𝑃𝑐𝑟 and we should be very clear if this critical load capacity of the column is

exceeded the column will become unstable. And we already said unstability or instability is a
form of failure.

(Refer Slide Time: 49:15)

So, friends in this lecture we learnt about stability criteria; Euler’s, Lyapunov’s and the third
one which is based on potential energy. We started learning more about Euler’s critical load
and we have understood that when the load capacity or the load applied load exceeds the
critical capacity column becomes unstable and instability is a form of failure.

Friends, we will close the lecture here thank you very much for your patient learning and you
will be keep on referring to the notes and the reference material which has been available in
the website of this course for additional learning. Stability is a very difficult topic to
understand I am going very gradually and slowly and I wish that all of you should pick up
this completely at single pace and try to learn this thoroughly undoubtedly. And then
disseminate it to your students to your fellow researchers and use it effectively in consultancy
and design offices.

So, that the area which is gray in structural engineering domain which is stability studies
becomes very simple and user-friendly treatment of this subject in the advanced steel design
domain for structural steel designs.

463
Thank you very much and have a good day bye.

464
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 27
Euler’s load

Welcome to the 27th lecture on Advanced Steel Design which is now going to focus on
estimating Euler’s critical load. So, in the last lecture we discussed about the conditions for
stability. We defined stability, then we have also defined what is a critical load. Let us rewind
that slightly and understand what do we mean by a critical load.

(Refer Slide Time: 01:08)

We can say the critical load is the axial load necessary to maintain the member in its initial
straight position. It is a classical definition which is given by Timoshenko and Gere in 1961.
The immediate question comes how do you estimate this critical load? Critical load is
computed based on the elastic curve equation. So, the elastic curve equation is classical
2
𝑑𝑦 𝑀
theory, we know that 2 = 𝐸𝐼
.
𝑑𝑥

We call equation number 1, where M is the moment, to be very clear bending moment, I is
the moment of inertia of the cross section and E is the modulus of elasticity of the material.
And, y and x are defined according to this figure which I am going to draw.

465
(Refer Slide Time: 03:48)

Let us say I have the initial portion of the column this way. We apply a load here, then at any
section let us say this is my deflector profile and the load is now shifted here. The distance of
this shift is y whereas, this is my origin and this becomes my x axis and of course, this is my
y axis where I am measuring y in that axis. This is a free body diagram of the column
member.

2
𝑑𝑦 𝑀
So, now we understand 2 = 𝐸𝐼
, let us go ahead. With reference to this figure, let us make
𝑑𝑥
2 2
𝑑𝑦 𝑀 𝑑𝑦
the following statements 2 = 𝐸𝐼
which I can say, this is equation number 1 𝐸𝐼 2 = 𝑀
𝑑𝑥 𝑑𝑥
2
𝑑𝑦
which in my case is going to 𝐸𝐼 2 = 𝑀=-Py. Why negative? Because, this is going to open
𝑑𝑥
2
𝑑𝑦
up the curvature, going to open up the curvature. So, 𝐸𝐼 2 + 𝑃𝑦 = 0.
𝑑𝑥

2
So, we can now say
𝑑𝑦
𝑑𝑥
2 + ( )𝑦 = 0 is a classical second order ordinary differential
𝑃
𝐸𝐼

equation, whose solution has got two components. The complementary function and the
particular solution. Since, RHS is 0 particular solution will be 0 in this case.

466
(Refer Slide Time: 06:51)

We can write the complementary solution like y is let us say some variable

𝑦 = 𝐴 sin 𝑠𝑖𝑛 ( ) + 𝐵 cos 𝑐𝑜𝑠 ( ) ;

α𝑥
𝐿
α𝑥
𝐿
equation number 4. Where the above equation

𝑃
alpha α = 𝐿 𝐸𝐼
, we call this as 4a. Say equation 4 has got two unknowns A and B, we need

to find out them. I can apply the standard boundary conditions and evaluate these constants A
and B.

So, the boundary conditions at x equal 0, y is 0. See here at x equals 0 y is 0 which implies
applying in this equation 4a, B becomes 0. Therefore, equation 4a is now written as

𝑦 = 𝐴 sin 𝑠𝑖𝑛 ( ).
α𝑥
𝐿

467
(Refer Slide Time: 08:40)

Let us supply next boundary condition. There is one more boundary condition at x equal to L,
y will be again 0. I have a column both ends position restrain, x is measured from here and
this is my y axis and the length of the column is L and the deflector profile of the column is
this under the load P. So, at x is equal to L, y is again 0 which implies 𝐴 sin 𝑠𝑖𝑛 (α) = 0 .

So, which means either A should be 0 or sin 𝑠𝑖𝑛 (α) should be 0. If A equal 0 look at the
original equation, there will be no lateral deflection, is it not. Hence, this is not applicable
condition. Therefore, setting sin 𝑠𝑖𝑛 (α) = 0 , we know alpha α = 𝑛π for n equal 0, 1, 2, 3
𝑃
etcetera. So, now we can say we already know that alpha α = 𝐿 𝐸𝐼
. So, let us say

𝑃
𝑛π = 𝐿 𝐸𝐼
. Call this as equation number 5.

468
(Refer Slide Time: 11:18)

2 2 2
2 2 𝐿𝑃 𝑛 π 𝐸𝐼
This says 𝑛 π = 𝐸𝐼
which says 𝑃 = 2 for n equals 1, 2, 3 and so on. n equals 0 is
𝐿

meaningless as this will cause no load, that is P will become 0. So, when there is no axial
load, the condition applicable is not effective. We call this equation as equation number 6. So,
equation 6 is termed as Euler’s critical.

(Refer Slide Time: 12:28)

469
 2 2
𝑛 π 𝐸𝐼
So, we can now say 𝑃𝐸 = 2 for n equals 1, 2, 3 etcetera. So, we have now estimated the
𝐿

Euler’s load. Now, let us try to find out the stability functions considering a standard beam
element.

(Refer Slide Time: 13:06)

So, let us now consider a standard beam element and we neglect the axial deformation. Let us
consider the standard beam element which is fixed at both the ends. The standard beam
element. This is x axis, this is y axis, I put m indicating is for the member. We take a
𝑡ℎ
prismatic section so, EI is constant and this becomes my span of the member of the 𝑖
𝑡ℎ
member 𝐿𝑖. This is my 𝑖 member. So, there are some sign conventions which we have to

follow.

So, let us see what are the sign conventions. (1), the end moment, joint rotation and joint
moments which are anti-clockwise or positive. The convention (2), upward force or
displacement is considered positive. (3) Force or axial displacement towards right direction is
considered positive.

470
(Refer Slide Time: 16:03)

(4), upward end shear is positive. (5), right direction force acting on the beam at the ends is
positive. So, these are some sign conventions which we will be following. So, let us mark
them. So, we say at any joint anti-clockwise, upward towards right or positive.

(Refer Slide Time: 16:54)

Now, let us consider a fixed beam undergoing deformation due to bending. We neglect the
axial deformation in the member, that is deformation along the axial of the member is
neglected. Now, the standard beam is what is shown here, the standard beam. Let us draw this
figure again and mark it here. This is my origin, this is my x axis of the member,

471
 𝑡ℎ
anticlockwise 90 degree is y axis, the 𝑖 member. This member has got uniform cross section
𝑡ℎ
EI, span of the member is 𝐿𝑖 and this is considered as a 𝑗 end of the member. This is
𝑡ℎ
considered as a 𝑘 end of the member.

So, there is a order by which we mark this x axis and y axis, identify the origin, mark the x
direction or x axis along the length of the member, y axis is anticlockwise 90 to x axis and
𝑡ℎ 𝑡ℎ
mark on 𝑗 and 𝑘 end of the member. And, very interestingly 𝑥𝑚𝑦𝑚 plane defines the plane

of bending of this beam element. Neglecting axial deformation; now, we have to identify the
degrees of freedom for this beam. Let us mark them.

(Refer Slide Time: 19:40)

𝑡ℎ
So, I am marking the beam again. This is my 𝑥𝑚, I know how to mark 𝑦𝑚. This is my 𝑗 end,
𝑡ℎ
𝑘 end, prismatic cross section, length of the member is 𝐿𝑖 and this member is actually i. So,

let us mark rotation at j as θ𝑝, rotation at k as θ𝑞, vertical displacement at j along positive m is
𝑡ℎ
δ𝑟 and vertical displacement at 𝑘 along positive 𝑦𝑚 is δ𝑠.

So, we are just marking these dimensions. I want to derive the stiffness matrix of this. So, to
do that I have to apply unit rotation. So, let us quickly revise and understand what is stiffness
coefficient; 𝑘𝑖𝑗 is the stiffness coefficient of the stiffness matrix k.

472
 𝑡ℎ 𝑡ℎ
This is defined as force in the 𝑖 degree of freedom for unit displacement in 𝑘 degree of
freedom, keeping all other degrees of freedom restrained. This is very important condition.
So, I must give unit rotation at all the degrees of freedom 1 by 1 and try to find out the forces,
that becomes a stiffness matrix. Let us draw those figures.

(Refer Slide Time: 21:54)

𝑡ℎ
So, we will draw 4 figures. Let us mark the unit rotation at the 𝑗 end and draw a tangent,
measure the angle from the axis. I call this as θ𝑝 and let the θ𝑝 be unity. So, now this will

invoke forces. I call that as 𝑘𝑝𝑝, this as 𝑘𝑞𝑝 and this reaction as 𝑘𝑟𝑝 and this as 𝑘𝑠𝑝. So, friends

let us pay attention to the subscripts of this.

The first subscript indicates where do you measure, the force the second subscript indicates
where do you give the unit displacement. Like look at this figure. This is p, this is q, r and s.
𝑡ℎ 𝑡ℎ
So, I am measuring force at the 𝑝 end. For giving displacement the 𝑝 end so, it is 𝑘𝑝𝑝. I am
𝑡ℎ 𝑡ℎ
measuring force at the 𝑞 end for displacement at the 𝑝 end so, 𝑘𝑞𝑝. Similarly, I get 𝑘𝑟𝑝 and

𝑘𝑠𝑝. Let us call this figure number (i).

𝑡ℎ
Let us go to figure number (ii), I want to give unit rotation at the 𝑘 end. So, draw a tangent
from the axis measure in the anti-clockwise direction call this as θ𝑞 that is unity. So, the

corresponding forces will be 𝑘𝑞𝑞 𝑘𝑝𝑞 𝑘𝑟𝑞 and 𝑘𝑠𝑞. You will obviously, see the second subscript

473
indicates the degree of freedom label where unit rotation is given. The first subscript
indicates the degree of freedom, where the force is measured. I call this figure number (ii).

𝑡ℎ
So, figure (i) and (ii) in fact, figure (i) refers to unit rotation at 𝑗 end correct, figure (ii)
𝑡ℎ 𝑡ℎ
refers to unit rotation at 𝑘 end, is it not; can also give unit displacement at the 𝑗 end. Let us
say I want to move this end by unit displacement that is here. So, I want to give this
displacement delta as unity which will be nothing, but δ𝑟 is unity. So, now, I do not want to

give displacement any other end.

Therefore, the displaced position will be this. It is a straight line. Let me rub this and redraw
1
again, it is a straight line. Let me draw an axis normal to this. Let me call this angle as α = 𝐿

, because this is L and this is unity. Now, the angle between the normals will be equal.
1
Therefore, friends this angle will also be 𝐿
, same as this angle.

𝑡ℎ 𝑡ℎ
So, now this will invoke forces in the 𝑝 degree, I call this as 𝑘𝑝𝑟. In the 𝑞 degree I call this
𝑡ℎ
as 𝑘𝑞𝑟, in the 𝑟 degree I call this as 𝑘𝑟𝑟 and this as 𝑘𝑠𝑟. The second subscript refers to the

place where we have given unit displacement. Similarly, let us say this unit displacement is
𝑡ℎ 𝑡ℎ
figure (iii) at 𝑗 end. Let us draw figure (iv) by giving unit displacement at the 𝑘 end.

𝑡ℎ
Let us do that. So, let us give unit displacement at the 𝑘 end. This delta equals 1 which is
actually equal to δ𝑠 which is unity. And, this becomes my new axis of the member, let me
1
draw normal to this. So, if I say this is α = 𝐿
and we all agree that this angle will also be 1

by L as well as this angle. This has invoked forces as 𝑘𝑞𝑠 𝑘𝑝𝑠 𝑘𝑟𝑠 and 𝑘𝑠𝑠. Friends you must

wonder how am I marking the arrow directions of the moment and then the axial reactions.

So, please note here I have given an unit anticlockwise rotation to this. So, the moment is
applied on the same direction, same direction is transferred here. Similarly, I have given
𝑡ℎ
anti-clockwise rotation at the 𝑘 end, in the same direction moment is applied to cross this
rotation, the same is transferred here. Whereas, in figure (iii) I have given upward
𝑡ℎ
displacement towards positive y at the 𝑗 end.

474
So, I want to bring this position back to normal. So, this is calling it back anti-clockwise.
𝑡ℎ
Same is applied here to 𝑘𝑞𝑟. When I come to figure (iv) 𝑘 end has moved up. So, I want to

bring it back. So, calling it back 𝑘𝑞𝑠 will now become clockwise. The same will be applied to

ps and I am marking these reactions rp, sp etcetera similar to the degrees of freedom which is
originally in the fixed beam.

𝑡ℎ 𝑡ℎ 𝑡ℎ
So, now we have given unit displacements at 𝑗 end, 𝑘 end and unit rotations at the 𝑗 end
𝑡ℎ
and 𝑘 end, is it not. And, we have also marked the corresponding forces at both the ends j
and k in all the 4 cases. Now, looking at this figure, we can write a statement.

(Refer Slide Time: 31:00)

𝑡ℎ
The unit rotation and displacements invoked forces and moments at j and 𝑘 ends of the
beam. Now, to maintain equilibrium, the governing equation should be applied. What is that
𝑡ℎ 𝑡ℎ
equation to be applied? I should say moment at 𝑝 end should be some of force at 𝑝 end
𝑡ℎ 𝑡ℎ 𝑡ℎ
due to θ𝑝, force at the 𝑝 end due to θ𝑞, force at the 𝑝 end due to δ𝑟 and force at the 𝑝 end
𝑡ℎ
due to δ𝑠, is it not. You can see here 𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑝𝑟 and 𝑘𝑝𝑠 or the net forces acting at the 𝑗 end

due to arbitrary displacements θ𝑝𝑞 δ𝑟 and δ𝑠 given in cycle, is it not.

𝑚𝑝 = 𝑘𝑝𝑝θ𝑝 + 𝑘𝑝𝑞θ𝑞 + 𝑘𝑝𝑟δ𝑟 + 𝑘𝑝𝑠δ𝑠

475
 𝑚𝑞 = 𝑘𝑞𝑝θ𝑝 + 𝑘𝑞𝑞θ𝑞 + 𝑘𝑞𝑟δ𝑟 + 𝑘𝑞𝑠δ𝑠

𝑝𝑟 = 𝑘𝑟𝑝θ𝑝 + 𝑘𝑟𝑞θ𝑞 + 𝑘𝑟𝑟δ𝑟 + 𝑘𝑟𝑠δ𝑠

𝑝𝑠 = 𝑘𝑠𝑝θ𝑝 + 𝑘𝑠𝑞θ𝑞 + 𝑘𝑠𝑟δ𝑟 + 𝑘𝑠𝑠δ𝑠

So, this equation should be valid. We call the equation number for example, we will continue
with the new numbering. We call this equation number 1. Similarly, I can write the equation
𝑡ℎ 𝑡ℎ
for 𝑚𝑞 which is now at the 𝑞 end. It should be the force at the 𝑞 end because of θ𝑝, force
𝑡ℎ 𝑡ℎ 𝑡ℎ
at the 𝑞 end because of θ𝑞, force at the 𝑞 end because of δ𝑟 and force at the 𝑞 end

because of δ𝑠. Then, can also find the reaction.

𝑡ℎ 𝑡ℎ
Let us call this as p r is the force at the 𝑞 end, force due to θ𝑝, force at 𝑞 end due to θ𝑞,
𝑡ℎ 𝑡ℎ 𝑡ℎ
force at the 𝑟 end due to δ𝑟 and 𝑠 end due to δ𝑠. Similarly, p s will be force at the 𝑞 end
𝑡ℎ 𝑡ℎ
due to θ𝑝, force at the 𝑘 end due to θ𝑞, force at the 𝑟 end due to δ𝑟 plus four force at the
𝑡ℎ
𝑠 end due to δ𝑠.

(Refer Slide Time: 34:40)

{𝑚𝑖} = [𝑘𝑖]{δ𝑖}

476
 {𝑚𝑝 𝑚𝑞 𝑝𝑟 𝑝𝑠 } = [𝑘]{θ𝑝 θ𝑞 δ𝑟 δ𝑠 }
Now, we can write this in a matrix form. We know that
{𝑚𝑖} = [𝑘𝑖]{δ𝑖}, {𝑚𝑖} is a vector, [𝑘𝑖] is a matrix, {δ𝑖}is a vector. I call this equation number 5.
So, where {𝑚𝑖} is 𝑚𝑝 𝑚𝑞𝑝𝑟𝑎𝑛𝑑 𝑝𝑠 , where [𝑘𝑖] is a full matrix and {δ𝑖} is θ𝑝 θ𝑞 δ𝑟 and δ𝑠.

(Refer Slide Time: 35:28)

𝑡ℎ 𝑡ℎ 𝑡ℎ
So, let us see what is this matrix which will be a 4 by 4 matrix. This will be the 𝑝 , 𝑞 , 𝑟
𝑡ℎ 𝑡ℎ 𝑡ℎ 𝑡ℎ 𝑡ℎ 𝑡ℎ
and 𝑠 columns; 𝑝 , 𝑞 , 𝑟 and 𝑠 rows. So, this is 𝑘𝑝𝑝𝑘𝑞𝑝𝑘𝑟𝑝𝑘𝑠𝑝 it is a 𝑝 column. So,
𝑡ℎ 𝑡ℎ 𝑡ℎ
𝑘𝑝𝑞𝑘𝑞𝑞𝑘𝑟𝑞𝑘𝑠𝑞 the 𝑞 column, 𝑘𝑝𝑟𝑘𝑞𝑟𝑘𝑟𝑟𝑘𝑠𝑟, the 𝑟 column and 𝑘𝑝𝑠𝑘𝑞𝑠𝑘𝑟𝑠𝑘𝑠𝑠, the 𝑠 column.

{𝑚𝑖} = [𝑘𝑖]{δ𝑖}
{𝑚𝑝 𝑚𝑞 𝑝𝑟 𝑝𝑠 } = [𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑝𝑟 𝑘𝑝𝑠 𝑘𝑞𝑝 𝑘𝑞𝑞 𝑘𝑞𝑟 𝑘𝑞𝑠 𝑘𝑟𝑝 𝑘𝑟𝑞 𝑘𝑟𝑟 𝑘𝑟𝑠 𝑘𝑠𝑝 𝑘𝑠𝑞 𝑘𝑠𝑟 𝑘𝑠𝑠 ]{θ𝑝 θ𝑞 δ𝑟 δ𝑠 }
So, our job is to find out this matrix, which will be the elemental property based on the
geometric and material characteristics of the beam member. Now, let us pick up the figures
𝑡ℎ
back again. So, I will copy this page, put it here. Now, friends for unit rotation given at the 𝑗
end, I develop 𝑘𝑝𝑝 and 𝑘𝑞𝑝. I can say this 𝑘𝑟𝑝 is actually equal to because the net moment now

generated is 𝑘𝑝𝑝 plus 𝑘𝑞𝑝.

477
So, 𝑘𝑟𝑝 can be simply said as 𝑘𝑝𝑝 plus 𝑘𝑞𝑝 by l, where this is my l and 𝑘𝑠𝑝 will be opposite to

𝑘𝑟𝑝. So, I can write here, this is equal to minus of 𝑘𝑟𝑝. So, this direction will be reversed

because, this is now going to cause a couple which will counteract this moment. So, 𝑘𝑟𝑝 and

𝑘𝑠𝑝 depends on 𝑘𝑝𝑝 and 𝑘𝑞𝑝 only. If I know 𝑘𝑝𝑝 and 𝑘𝑞𝑝, I can find the end reactions 𝑘𝑟𝑝 in

𝑘𝑠𝑝. Similarly, let us go to this case (ii).

So, now, this is invoking an anti-clockwise moment of 𝑘𝑝𝑞 and 𝑘𝑞𝑞, this will be counteracted

by a couple which is 𝑘𝑟𝑞 and 𝑘𝑠𝑞. So, 𝑘𝑟𝑞 will be now 𝑘𝑝𝑞 plus 𝑘𝑞𝑞 by l and 𝑘𝑠𝑞 with a minus

of 𝑘𝑟𝑞. Now, this will be get reversed. We have now expressed these reactions in figures (i)

and (ii) as a function of the moments pq and qq. Similarly, I can also find or express these
reactions as a sum of 𝑘𝑝𝑟 and 𝑘𝑞𝑟.

Friends, please understand we are trying to express the reactions in terms of the end
moments; we have done with figure (i), we have done with figure (ii). We will do the figure
(iii) and figure (iv) plus the moment. We will do that. So, we will do it in the next lecture.

(Refer Slide Time: 40:16)

We will write a summary here. In this lecture, we learned how to estimate Euler’s critical
load which is called as 𝑃𝐸𝑢𝑙𝑒𝑟'𝑠. We have also started understanding the derivation for

478
standard fixed beam and forming the stiffness matrix for this beam from 1st principles. We
will continue in the next lecture and attend to this in the next lecture.

Thank you very much and have a good day. Bye.

479
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 28
Rotation coefficients for stability functions

(Refer Slide Time: 00:21)

Welcome to the 28 Lecture on Advanced Steel Design course, we are in the process of
deriving the stability functions. Now in this lecture we will continue to learn the rotation
coefficients and then the stiffness matrix of a standard beam.

480
(Refer Slide Time: 00:54)

Let us rewind back this is what we discussed in the last lecture, we learnt how to obtain 𝑘𝑟𝑝

and 𝑘𝑠𝑝 in terms of p p and q p with reference to figure 1 and figure 2.

𝑘𝑟𝑝 = ( 𝑘𝑝𝑝+𝑘𝑞𝑝
𝑙 )
𝑘𝑠𝑝 =− 𝑘𝑟𝑝

𝑘𝑟𝑞 = ( 𝑘𝑝𝑞+𝑘𝑞𝑞
𝑙 )
𝑘𝑠𝑞 =− 𝑘𝑟𝑞

481
(Refer Slide Time: 01:22)

Let us focus on figure 3 now, friends the standard beam looks like this. So, if I give a unit
rotation of θ𝑝 as unity this invokes a moment which is 𝑘𝑝𝑝 and 𝑘𝑞𝑝. If this is true then suppose

if I have a beam where the rotation is 1 by 𝐿𝑖 where this is L. So, this is figure a lets have one
𝑡ℎ
more figure and the 𝑘 here.

𝑘𝑝𝑟 = ( 𝑘𝑝𝑝+𝑘𝑝𝑞
𝐿 )
=( )
𝑘𝑞𝑝+𝑘𝑞𝑞
𝑘𝑞𝑟 𝐿

If I give a rotation here then this gives me a moment of 𝑘𝑞𝑞 and 𝑘𝑝𝑞. So, this is equivalent to a

member where I give unit displacement here and the member rotates by 1 by L. Now
considering these two figures and comparing this with figure 3 and 4 let us say 𝑘𝑝𝑟 is a
𝑡ℎ
moment which is invoked by giving unit displacement at the 𝑟 end. But can I write 𝑘𝑝𝑟 as an

effect of these two sum for unit rotation.

So, I am writing 𝑘𝑝𝑟 as 𝑘𝑝𝑝 plus 𝑘𝑝𝑞 into 1 by L why 1 by L 𝑘𝑝𝑝 and 𝑘𝑝𝑞 are the moments for

unit rotation whereas, 𝑘𝑝𝑟 is the moment for 1 by L rotation. So, I am multiplying this with 1

by L. So, I can write 𝑘𝑝𝑟 is it not similarly can I also write 𝑘𝑞𝑟 similarly 𝑘𝑞𝑟 can also be a sum

482
of 𝑘𝑞𝑝 plus 𝑘𝑞𝑞 by L, because again these are moments because of unit rotation, but I want

moment because of o1 by L.

So, therefore, I can multiply this with 1 by L. So therefore, friends once in this figure I know
𝑘𝑝𝑟 and 𝑘𝑞𝑟 let me copy this I will put it here.

(Refer Slide Time: 05:50)

𝑘𝑝𝑟 = ( 𝑘𝑝𝑝+𝑘𝑝𝑞
𝐿 )
=( )
𝑘𝑞𝑝+𝑘𝑞𝑞
𝑘𝑞𝑟 𝐿

=( )
𝑘𝑝𝑟+𝑘𝑞𝑟
𝑘𝑟𝑟 𝐿

𝑘𝑟𝑟 = ( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 )
𝑘𝑠𝑟 =− 𝑘𝑟𝑟

So, I already know 𝑘𝑝𝑟 as 𝑘𝑝𝑝 plus 𝑘𝑝𝑞 by L. see here, I also know 𝑘𝑞𝑟 as 𝑘𝑞𝑝 plus 𝑘𝑞𝑞 by L.

So now, I can write the net moment here is 𝑘𝑝𝑟 plus 𝑘𝑞𝑟 which is here. So, now I can say 𝑘𝑟𝑟

is actually 𝑘𝑝𝑟 plus 𝑘𝑞𝑟 by L because there is a couple, can you now write 𝑘𝑟𝑟 as

( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 ). So now, can I also write 𝑘 𝑠𝑟
as minus of 𝑘𝑟𝑟 because that is a couple. So,

483
this will be acting downward this will act up this becomes a couple the counter act this
moment.

(Refer Slide Time: 08:24)

(
𝑘𝑝𝑠 =−
𝑘𝑝𝑝+𝑘𝑝𝑞
𝐿 )
𝑘 =− ( )
𝑘𝑞𝑝+𝑘𝑞𝑞
𝑞𝑠 𝐿

=− ( )
𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞
𝑘𝑟𝑠 2
𝐿

=+ ( )
𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞
𝑘𝑠𝑠 2
𝐿

Similarly, let us do this for the other end let us copy this figure and put it here and then work
on this similarly same order I am just trying to do for the other end. You know 𝑘𝑝𝑠 will be

actually 𝑘𝑝𝑝 plus 𝑘𝑝𝑞 by L. 𝑘𝑞𝑠 will be 𝑘𝑞𝑝 plus 𝑘𝑞𝑞 by L, but there is a small change here

originally 𝑘𝑝𝑝 look at here originally 𝑘𝑝𝑝 and 𝑘𝑝𝑞 were anticlockwise, but I am here looking

for a clockwise point.

So, shall I say these two are minus is it not they are opposite to the original one is it not. So,
if I know the net moment is 𝑘𝑝𝑠 plus 𝑘𝑞𝑠 then this will have a moment which is a couple like

this. So, this is going to be 𝑘𝑟𝑠 going to be 𝑘𝑠𝑠. So, I can now write 𝑘𝑟𝑠 as ( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 ),
484
but with the minus sign. Whereas, 𝑘𝑠𝑠will be ( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 ), but with the plus sign because
this sign matches with this. So, now I have all the coefficients let me transfer these
coefficients to the matrix.

(Refer Slide Time: 10:52)

[
[𝐾] = 𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑝𝑟 𝑘𝑝𝑠 𝑘𝑞𝑝 𝑘𝑞𝑞 𝑘𝑞𝑟 𝑘𝑞𝑠 𝑘𝑟𝑝 𝑘𝑟𝑞 𝑘𝑟𝑟 𝑘𝑟𝑠 𝑘𝑠𝑝 𝑘𝑠𝑞 𝑘𝑠𝑟 𝑘𝑠𝑠 ]
So, what is the original matrix k the original matrix k is if this is p if this is q , r ,and s this is
p, q, r and s are the labels. So, this is 𝑘𝑝𝑝, 𝑘𝑝𝑞, 𝑘𝑝𝑟, 𝑘𝑝𝑠, 𝑘𝑞𝑝, 𝑘𝑞𝑞, 𝑘𝑞𝑟, 𝑘𝑞𝑠, 𝑘𝑟𝑝, 𝑘𝑟𝑞, 𝑘𝑟𝑟, 𝑘𝑟𝑠,

𝑘𝑠𝑝, 𝑘𝑠𝑞, 𝑘𝑠𝑟, 𝑘𝑠𝑠. So now, I have all the values with me please look at the notes they have all

the values.

485
(Refer Slide Time: 12:04)

Let me write down these coefficients here I will tell you very simple way to remember this. It
is a very simple way the k matrix can be easily remembered please watch here. So, it is easy
for you to recollect. So, it s a 4 by 4 matrix please watch here p,q,r,s & p,q,r and s are the
labels. this is 𝑘𝑝𝑝 and this is 𝑘𝑞𝑝 this is 𝑘𝑝𝑞, this is 𝑘𝑞𝑞 and this value is 𝑘𝑝𝑝 plus 𝑘𝑞𝑝 by L it is

nothing but sum of these two by L and this is minus of this.

𝑓𝑜𝑟 𝑚𝑎𝑡𝑟𝑖𝑥 [𝐾] = ⎡⎢𝑘𝑝𝑝 𝑘𝑝𝑞

⎣
𝑘𝑝𝑝+𝑘𝑝𝑞
𝐿
− ( 𝑘𝑝𝑝+𝑘𝑞𝑝
𝐿 )𝑘 𝑘
𝑞𝑝 𝑞𝑞
𝑘𝑞𝑝+𝑘𝑞𝑞
𝐿
− ( 𝑘𝑞𝑝+𝑘𝑞𝑞
𝐿 ) 𝑘𝑝𝑝+𝑘𝑝𝑞
𝐿
𝑘𝑝𝑞+𝑘𝑞𝑞
𝐿
𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞

𝐿
2

Similarly, sum of these two by L will be here this is minus of this. The third column is very
easy sum of these two by L second element sum of these two by L, third element sum of
these two further by L. So, that becomes 𝑘𝑝𝑝 𝑘𝑞𝑝𝑘𝑝𝑞. 𝑘𝑞𝑝 𝑝𝑙𝑢𝑠 𝑘𝑞𝑞 by L square is it not and

this will be − ( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 ).
Now, once you written the 3rd column the 4th column is very easy the 4th column is minus of

3rd column. So, minus of 𝑘𝑝𝑝 𝑘𝑞𝑝 by L minus of 𝑘𝑞𝑝𝑘𝑞𝑞 by L. − ( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

2
𝐿 ). then
( 𝑘𝑝𝑝+𝑘𝑝𝑞+𝑘𝑞𝑝+𝑘𝑞𝑞

𝐿
2 ). Friends you can always note that the elements along leading diagonal are
positive they are positive they can never be negative. So, please check that when I do the
derivation and understand very carefully. So, in this matrix these 4 values 1 2 3 and 4 these
are called rotation coefficients.

486
(Refer Slide Time: 15:47)

Let us say this is my matrix this is p q r and s are the degrees of freedom, these are called
rotational coefficients and these 4 are called translational coefficients. Friends carefully look
at the derivation back look here in this equation if we know this value and this value or these
4. I can fill up the remaining matrix very fast. So, I need to know only the rotational
coefficients if you know the rotational coefficients I can fill up this matrix very fast.

(Refer Slide Time: 17:16)

487
So, our job is to now derive the rotational coefficients. So, the rotational coefficients will be
derived based on the flexibility method. So, we want to consider a simply supported beam as
shown in the figure. So, I will draw the figure.

1
𝑓α [𝑘]

Let us consider simply supported beam, we will consider 2 figures in one figure we will give
𝑡ℎ 𝑡ℎ 𝑡ℎ
unit rotation at 𝑗 end this is my 𝑗 end this is my 𝑘 end is it not, we will give unit rotation.
So, I am going to now use flexibility coefficients, I know flexibility and stiffness are
inversely proportional we know that. So, I will find the flexibility coefficient matrix and
invert it to get the stiffness matrix which is rotational matrix to do that I want to give.

Let us say unit rotation, so let me give unit notation here. So, let us say I give θ𝑗 is unit. So, I

draw a tangent to this I call this as δ𝑗𝑗 and I call this as δ𝑘𝑗, where these are flexibility
𝑡ℎ
coefficients. Similarly I give unit rotation at the 𝑘 end and I call this as δ𝑘𝑘 and I call this as

δ𝑗𝑘. So now, imagine we will compare this.

(Refer Slide Time: 21:14)

I will copy this I will compare this with our standard member which is initially normal and
this is 𝑘𝑝𝑝 we know that and this was 𝑘𝑞𝑝 and this was θ𝑝 which is unity. I am comparing

488
these two figures, similarly an equivalent figure for this would be comparison of we can even
draw this straight θ𝑞 is unity and this was 𝑘𝑞𝑞 and this was 𝑘𝑝𝑞 and these are equivalent.

[δ𝑗𝑗 δ𝑗𝑘 δ𝑘𝑗 δ𝑘𝑘 ][𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑞𝑝 𝑘𝑞𝑞 ] = [𝐼]

[𝑓] [𝑘] = [𝐼]

Now comparing these two figures I can write the control equation like this δ𝑗𝑗δ𝑗𝑘δ𝑘𝑗δ𝑘𝑘 which

are flexibility coefficients multiplied by 𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑞𝑝 𝑘𝑞𝑞, will give me identity matrix.

This is 1 into 1 match, 1 is to 1 match given identity matrix. So, this my flexibility matrix,
this my stiffness matrix, my anti matrix. Now my question is to find these values of flexibility
matrix, if I know this, I can invert this matrix to get my rotational coefficient matrix you see
here this is what I want this is what I want this is 𝑘𝑝𝑝 this is 𝑘𝑝𝑞 this is 𝑘𝑞𝑝 this 𝑘𝑞𝑞 this is what

I want this is here.

So, to get this I will first find the flexibility matrix and invert it I will get this matrix.

(Refer Slide Time: 24:24)

𝑡ℎ
To get that flexibility is inverse of stiffness let us define what is stiffness first 𝑘𝑖𝑗 is force in 𝑖
𝑡ℎ
degree for unit displacement in the 𝑘 degree by keeping all other degrees of freedom
restrained. Now I want to write the flexibility definition, flexibility coefficient will be

489
 𝑡ℎ 𝑡ℎ
displacement in 𝑖 degree for unit force in 𝑘 degree by keeping all other degrees of
freedom restrained is it not just a reverse of this.

So, displacements can be of 2 rotational translational, if I give a rotational displacement I will

get a moment if I give a translational displacement I get a force. So now, I must give unit
𝑡ℎ 𝑡ℎ
rotation at 𝑖 degree I mean I have to find the rotation 𝑖 degree by giving unit moment at
𝑡ℎ
the 𝑘 degree, that is what I am going to do now let us do that.

(Refer Slide Time: 26:15)

So, let me draw the figure of the simply supported beam again is having length L this my
displaced configuration.

( 1
2
𝑙
1
𝐸𝐼 )( ) = 𝑉
𝑙
3 𝐵

𝑙
6𝐸𝐼
= 𝑉𝐵

𝑉𝐴 = ( 1
2
(𝑙)
1
𝐸𝐼 )− 𝑙
6𝐸𝐼

𝑙
𝑉𝐴 = 3𝐸𝐼

I want to give unit moment and the displacement cost is rotation which is δ𝑗𝑗 and here this is

δ𝑘𝑗. So, let me draw the bending moment diagram of this. So, I have given unit moment here

and this is 0. So, by this logic tension will be the top and compression will be the bottom is it

490
not I given moment like anti clockwise here. So, let me apply this as an M by EA diagram to
my beam. So, this is my conjugate beam I am loading this beam with my M by EA diagram
this is my M by EA diagram. This is how this diagram is done.

So, this is going to be 1 by EI why one because this is actually unity that is what is here right.
So, this will have a CG somewhere here this will be at the distance l by 3. Now this will
invoke reaction here and here let us see we call this as 𝑉𝐵 this is 𝑉𝐴. Now let us take moment

about A. So, half into base into height that gives the area of the triangle into the distance of
the CG from a this is l by 3 should be equal to 𝑉𝐵 into l provided 𝑉𝐵 is now opposite.

So, I can now find 𝑉𝐵 as so l, l goes away. So, it is going to be l by 6 EI am I right its acting

downward. Now I can say 𝑉𝐴 will be equal to half into base into height the total force minus l

by 6 EI that is what my 𝑉𝐴 is? So, this will give me 𝑉𝐴 as l by 3 EI acting upward. So, this is

l by 3 EI this is l by 6 EI is acting downward this acting upward. So, I have 𝑉𝐴 and I have 𝑉𝐵

Now, as per the conjugate beam effect this will be the displacements. So, I can straight away
say they are going to be I can write here this is going to be δ𝑗𝑗 and this is going to be δ𝑘𝑗. I

can do the same way for the other end.

(Refer Slide Time: 31:26)

491
 𝑡ℎ
Similarly, one can do this process at 𝑘 end. So, if you do that you will get the same
equation.

𝑙
δ𝑗𝑗 = 3𝐸𝐼

𝑙
δ𝑘𝑗 =− 6𝐸𝐼

𝑙
δ𝑗𝑘 =− 6𝐸𝐼

𝑙
δ𝑘𝑘 = 3𝐸𝐼

𝑙 𝑙 𝑙 𝑙
𝑓 = ⎡ 3𝐸𝐼 − − ⎤
⎣ 6𝐸𝐼 6𝐸𝐼 3𝐸𝐼 ⎦

We do that you will find that we wanted to find δ𝑘𝑘δ𝑗𝑘. δ𝑗𝑘 will be minus 6 l by 6 EI and this

will be l by 3 EI. So, we have now all the 4 values we have δ𝑗𝑗 as l by 3 EI δ𝑘𝑗 as l by 6 EI

then δ𝑗𝑘 as minus l by 6 EI and δ𝑘𝑘 as l by 3 EI can write this in a matrix form as l by 3 EI

minus l by 6 EI minus l by 6 EI and l by 3 EI this my flexibility matrix.

(Refer Slide Time: 34:04)

I want to find the stiffness matrix of rotational coefficients which will be inverse of this
matrix. So, that will be 𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑞𝑝 𝑘𝑞𝑞 which will be conduct inverse of this which is l by 3

EI minus l by 6 EI minus l by 6 EI and l by 3 EI inverse, you will find this matrix as 4 EI by l

492
2 EI by l 2 EI by l 4 EI by l. So, the rotational coefficients can be obtained like this let us go
back to the original matrix.

−1
[𝑘]𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑑𝑜𝑓 = [𝑓]

−1
[𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑞𝑝 𝑘𝑞𝑞 ] = ⎡⎣ 3𝐸𝐼𝑙 −
𝑙
6𝐸𝐼
−
𝑙
6𝐸𝐼
𝑙
3𝐸𝐼
⎤
⎦
4𝐸𝐼 2𝐸𝐼 2𝐸𝐼 4𝐸𝐼
[𝑘] = ⎡ ⎤
⎣ 𝑙 𝑙 𝑙 𝑙 ⎦

So, we are talking about the rotational coefficients which are these 4, if I know these fours
𝑘𝑝𝑝 𝑘𝑝𝑞𝑘𝑞𝑝 𝑘𝑞𝑞 is what we have just found out. So, let me say if I know this I can write the

remaining, let us see how? Let us copy this I want the full stiffness matrix, let us rub the
remaining one let us see how do you fill up this we know this.

(Refer Slide Time: 36:06)

[𝑘] = ⎡⎢ 𝑙
4𝐸𝐼 2𝐸𝐼 6𝐸𝐼 6𝐸𝐼 2𝐸𝐼 4𝐸𝐼 6𝐸𝐼 6𝐸𝐼 6𝐸𝐼 6𝐸𝐼 12𝐸𝐼 12𝐸𝐼 6𝐸𝐼 6𝐸𝐼 12𝐸𝐼
𝑙 2 − 2 𝑙 𝑙 2 − 2 2 2 3 − 3 − 2 − 2 − 3 +
⎣ 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙

Let us rub this this is 4 EI by l and this is 2 EI by l and this is 2 EI by l and this is 4 EI by l
just now we computed there see here just now we computed. Once you know this sum of
these two by l is 6 EI by l square minus of this. Similarly sum of these two by l 6 EI by l
square then minus of this, then sum of these two by l sum of these two by l then sum of these
two by l, then minus of that the last column is negative of the third column. That is how we
have the stiffness matrix of the beam element is ready.

493
(Refer Slide Time: 37:51)

So, friends in this lecture we learnt how to derive the rotational coefficients of stiffness
matrix. We also learnt how to formulate the complete stiffness matrix neglecting axial
deformation and we learnt that k is 4 by 4 is very easy to write. If you know this label p q r
and s if I know this value I can keep on writing the remaining all very easily based on this.
So, in the next lecture we are going to continue discussing stability functions and axial
compression. But based on the same derivation technique what we used for a fixed beam.

So, keep on revising what we discussed in the previous lectures that will be very helpful for
you and the reference already slated for you. Please go through the additional references
given in the website of this course and improve your additional learning skills in parallel. So,
that the derivation becomes very clear for you.

Thank you very much and have a good day bye.

494
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 29
Stability functions - 1

(Refer Slide Time: 00:21)

Friends, welcome to the 29th Lecture on Advanced Steel Design course, in this lecture we are
going to learn how to derive the Stability Functions under axial compression. Friends, in the
last lecture we discussed about the derivation of a stiffness method using stiffness method of
a standard fixed beam and we understood that a standard fixed beam has got 4 kinematic
degrees of freedom as marked on the screen.

We have used a specific sign convention that anticlockwise moments and rotations are
positive, displacements upward shear and displacements upward are positive. Similarly, along
x axis displacement and forces will be positive, we have used this n convention and we have
derived a stiffness matrix of 4 by 4; obviously, we neglected axial deformation.

495
(Refer Slide Time: 02:07)

[
[𝐾] = 𝑘𝑝𝑝 𝑘𝑝𝑞 𝑘𝑝𝑟 𝑘𝑝𝑠 𝑘𝑞𝑝 𝑘𝑞𝑞 𝑘𝑞𝑟 𝑘𝑞𝑠 𝑘𝑟𝑝 𝑘𝑟𝑞 𝑘𝑟𝑟 𝑘𝑟𝑠 𝑘𝑠𝑝 𝑘𝑠𝑞 𝑘𝑠𝑟 𝑘𝑠𝑠 ]
And we derive the stiffness matrix as you see on the screen, we have the labels p q r s and p q
𝑡ℎ
r s these labels have a specific order. Please see this order rotation, if this is my 𝑗 end of the
𝑡ℎ
member this is my 𝑘 end of the member along the length of the member is my x axis y axis
is anticlockwise 90 to x axis.

𝑡ℎ
Therefore, the degrees of freedom are labelled in such a manner rotation at 𝑗 end, rotation at
𝑡ℎ 𝑡ℎ 𝑡ℎ
𝑘 end, displacement along positive y at 𝑗 end and displacement along positive y at 𝑘 end.
So, p q r s this is the order.

So, we have written this order here and we know this is going to be 𝑘𝑝𝑝, k represents the

element of this matrix, 𝑘𝑝𝑝, 𝑘𝑝𝑞, 𝑘𝑝𝑟, 𝑘𝑝𝑠, 𝑘𝑞𝑝, 𝑘𝑞𝑞, 𝑘𝑞𝑟, 𝑘𝑞𝑠, 𝑘𝑟𝑝, 𝑘𝑟𝑞, 𝑘𝑟𝑟, 𝑘𝑟𝑠, 𝑘𝑠𝑝, 𝑘𝑠𝑞, 𝑘𝑠𝑟, 𝑘𝑠𝑠.

While writing this subscripts of these elements of the stiffness matrix the first subscript refers
the row and second subscript refers the column, that is a standard practice what we do in
representing the equations or mathematical formulae in a matrix form.

So, this is a 4 by 4 matrix, yesterday we learnt that these 4 coefficients are named as rotation
coefficients and these 4 coefficients are termed as translational coefficients. And we derived

496
this the remaining coefficients as a function of this is not. For example, I am just revising 𝑘𝑟𝑝

is sum of these two by l. So, I can say 𝑘𝑟𝑝 = ( 𝑘𝑝𝑝+𝑘𝑞𝑝

𝑙 )and so on and so forth.
So, we have got all the 16 coefficients learnt and understood for a standard fixed beam
neglecting axial deformation is it not? Having said this, we will follow the same analogy, but
now we will derive the stability functions under axial compression, it means I am going to
apply an axial compressive force to the same fixed beam module and derive the matrix in the
same order as we have derived earlier.

Let us see how we are going to do that.

(Refer Slide Time: 05:41)

So, now, I am going to say consider a beam element both ends fixed. Let us consider the
element this is my element and let us mark the axis we know this is my x axis, my y axis
measured anticlockwise ninety and this A and E is considered to be constant along the length
of the member it is a prismatic section.

So, let us mark the degrees of freedom for this they are all restrained let us mark them in red
colour. So, this is θ𝑝, this is θ𝑞 and this we know is δ𝑟 and this we know is δ𝑠.

In addition to this I am going to apply an axial force 𝑝𝑎 at both the ends. Let me mark this 𝑥𝑚

slightly in a different colour. So, that it does not get confused this is 𝑦𝑚 let me also mark this

497
𝑦𝑚 here this is 𝑥𝑚, m represents the member and this is my force which is axial. So, there is
𝑡ℎ 𝑡ℎ
no confusion in this and the length of the member is 𝑙𝑖 let say is an 𝑖 member, it is an 𝑖
𝑡ℎ 𝑡ℎ
member and we know that this is my 𝑗 end, this is my 𝑘 end that is how we have marked.

Now, this is the figure of a fixed beam under axial compressive load. Friends, please
understand when you derive the stiffness matrix there was no load applied, stiffness matrix is
of course, independent of any load, but now in this case I am going to derive the stability
function. Therefore, we already learnt stability is that function or that capacity or that ability
of the structure to perform its intended function up to the critical load for which it is
designed.

So, I am applying the axial load which is in compression, I am going to compress this beam
from either ends. Please note very carefully here; the element or the module does not have
axial degree of freedom, the axial degree of freedom is neglect. Please note that I have only 4
degrees of freedom p q r s. Similarly, what we had in the fixing there is no change in that.

(Refer Slide Time: 10:02)

Having said this let us now derive what is called rotation functions, as earlier we derived
what is called rotational coefficients, now I am talking about stability function therefore, I
have rotation functions.

498
So, now to get this rotation function I must apply a unit rotation let us see that, I am taking
the beam fix both the ends the beam is now subjected to and unit rotation θ𝑝. So, it is the

same algorithm what we did in the beam earlier that is why I explained the beam analysis
first.

Now, let us mark the axial force present in the section, is it not? This was new in the earlier
beam analysis this was not there, let us also mark the degrees of freedom, let us say. I will not
mark the degrees I will mark the forces. So, what I am going to say this is my force 𝑘𝑝𝑝, this

is my force 𝑘𝑞𝑝 and this is 𝑘𝑟𝑝 and this is k. Let us write this 𝑘𝑞𝑝 here.

(Refer Slide Time: 11:41)

Let us write this 𝑘𝑝𝑝 here, let us mark exactly this here so, 𝑘𝑟𝑝 and 𝑘𝑠𝑝. What are these 𝑘𝑟𝑝 𝑘𝑠𝑝

etcetera? They are all forces when the system is subjected to unit displacement. So, they are
actually similar to stiffness coefficients can also put a superscript i, saying that this is for the
𝑡ℎ
𝑖 member, but that is redundant when it will it may cause additional confusion. So, I am
avoiding that and of course, we know this is the length of the member 𝐿𝑖.

(𝑘𝑝𝑝+𝑘𝑞𝑝)
𝑘𝑟𝑝 = 𝐿𝑖

𝑘𝑠𝑝 =− 𝑘𝑟𝑝

499
 2
𝑑𝑦
𝐸𝐼 2 = 𝑀
𝑑𝑥

=− 𝑃𝑎(𝑦) − 𝑘𝑝𝑝 + 𝑘𝑟𝑝(𝑥)

2
𝑑𝑦
𝐸𝐼 2 =− 𝑃𝑎𝑦 − 𝑘𝑝𝑝 + 𝑘𝑟𝑝(𝑥)
𝑑𝑥

( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦
𝐸𝐼 2 =− 𝑃𝑎𝑦 − 𝑘𝑝𝑝 + 𝐿𝑖
𝑥
𝑑𝑥

𝑡ℎ
So, now this figure indicates unit rotation at 𝑗 end of the fixed beam, subjected to axial
combustion correct. So, now, when you do this, this is start developing moments and shear at
𝑡ℎ 𝑡ℎ
both ends of the member both 𝑗 and 𝑘 ends. So, now, this has an unbalanced moment
anticlockwise which will be 𝑘𝑝𝑝 plus 𝑘𝑞𝑝.

So, can I now say 𝑘𝑟𝑝 will be actually equal to this unbalanced moment by l or l I, because

there is going to form a couple when I say this is a couple. So, this will be upward this will be
downward. So, can I say now 𝑘𝑠𝑝 is same as 𝑘𝑟𝑝, but with a negative sign. So, I will call this

equation as 1.

I hope there is no confusion at this stage, is it not same identical analysis what we discussed
for the fixed beam, I am just following the same algorithm. So, it is easy for us to do it. Now,
what I do I draw a free body diagram under the influence of this by cutting a section
somewhere here let us see and draw free body diagram let me do that.

I draw a free body diagram fixed beam one end fixed I have applied a rotation an donut
engine and this is unity it is subjected to an axial force 𝑃𝑎 it is having end moment 𝑘𝑝𝑝

because free body diagram is a true representation of all the internal external forces acting on
the segment considered. So, this is going to be 𝑘𝑟𝑝.

Now, at this stage I have a balance force 𝑃𝑎 to counteract this let us say that is at a distance y

from the axis of the member and it also has a moment which is yeah, I should say this figure
𝑡ℎ
represents free body diagram under axial compression and unit rotation at 𝑗 end.

Now, we also have the classical differential equation to represent this which we discussed.
So, I will copy this figure or write it I write it here itself. So, now, I say with reference to this

500
 2
𝑑𝑦
figure free body diagram I can say 𝐸𝐼 2 is yeah and this
𝑑𝑥

𝑀 =− 𝑃𝑎(𝑦) − 𝑘𝑝𝑝 + 𝑘𝑟𝑝(𝑥). I am considering a section at a distance x from here.

2
𝑑𝑦
𝐸𝐼 2 =− 𝑃𝑎𝑦 − 𝑘𝑝𝑝 + 𝑘𝑟𝑝(𝑥)
𝑑𝑥

( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦
𝐸𝐼 2 =− 𝑃𝑎𝑦 − 𝑘𝑝𝑝 + 𝐿𝑖
𝑥
𝑑𝑥

(Refer Slide Time: 18:56)

𝑃𝑎 = ϕ𝑖𝑃𝐸

2
π 𝐸𝐼
𝑃𝐸 = 2 𝑓𝑜𝑟 𝑛 = 1
𝐿

2
π Φ𝑖𝐸𝐼
𝑃𝑎 = 2 𝑓𝑜𝑟 𝑛 = 1
𝐿𝑖

In the above equation, let us express the axial load that is 𝑃𝑎 as a function of Euler load 𝑃𝐸,

you may ask me a question why are you doing this? Because answer is we are aiming to
derive the stability function, stability function is related to Euler load. So, therefore, let 𝑃𝑎 be

expressed as some function of 𝑃𝐸 call this equation number 4.

501
 2
2
π 𝐸𝐼 π Φ𝑖𝐸𝐼
We already know that 𝑃𝐸 = 2 for n equals 1, hence 𝑃𝑎 will be 𝐿𝑖
for n equals 1 can I
𝐿

have this as equation number 5. So, friends, it is a very important observation you want to
make here. Please note that, buckling is occurring in the plane where unit rotation is applied,
you may ask me a question, suddenly how I have introduced the term buckling? Friends, look
at this figure this is a beam or a column element subjected to axial compression.

Under axial compression column will buckle and that is one of the failure mode, we already
discussed that in the previous lecture in detail is not it. So, I will not no more call this as
bending this is not happening because of any lateral load or gravity load, the profile what is
shown on the screen is happening purely because of buckling and it is happening on the same
plane where unit rotation is given, is that.

Then you may ask me a question, sir, why we are estimating buckling when rotation is given?
That is very good; we are estimating that load which is going to cause failure. So, we are
expressing that load of 𝑃𝑎 as a function of Euler’s load, if that load 𝑃𝑎 exceeds Euler’s load it

will fail and that failure will be called as buckling failure, that is what Euler stated, is it not?
So, I am using the same logic I am expressing it.

So, therefore, I am calling this as buckling and let us remember that this is happening on the
same plane where unit rotation is applied. What is this plane to be very clear; it is 𝑥𝑚 𝑦𝑚

plane, is it not see here, it is this is 𝑥𝑚 and this is 𝑦𝑚 that is the board, is it not? It is

happening in the same plane. Therefore, let us modify equation 3 in this understanding,
because we have substituted, we have got 𝑃𝑎 in equation 3. 𝑃𝑎 is expressed in a different

form in equation 5. So, let us substitute equation 5 and update equation 3.

502
(Refer Slide Time: 24:34)

2
π Φ𝑖𝐸𝐼
( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦
𝐸𝐼 2 =− 2 (𝑦) − 𝑘𝑝𝑝 + 𝐿𝑖
(𝑥)
𝑑𝑥 𝐿𝑖

2
π Φ𝑖
( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦 1 1
2 =− 2 (𝑦) − 𝐸𝐼
𝑘𝑝𝑝 + 𝐸𝐼 𝐿𝑖
𝑥
𝑑𝑥 𝐿𝑖

2
π Φ𝑖
( )
2
𝑑𝑦
+ (𝑦) =
1 ⎡𝑘 + 𝑘 𝑥
− 𝑘𝑝𝑝⎤⎥
2 2 𝐸𝐼 ⎢ 𝑝𝑝 𝑞𝑝 𝐿𝑖
𝑑𝑥 𝐿𝑖 ⎣ ⎦

So, let us update equation 3 after substitution. So, that becomes

2
π Φ𝑖𝐸𝐼
( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦
𝐸𝐼 2 =− 2 (𝑦) − 𝑘𝑝𝑝 + 𝐿𝑖
(𝑥),we call this equation as 6.
𝑑𝑥 𝐿𝑖

Now, I will divide this equation 6 by E I and rearrange the terms, divide equation 6 by E I and
rearrange the terms let us do that. So, if you do that I will get this equation now which is
2
π Φ𝑖
( 𝑘𝑝𝑝+𝑘𝑞𝑝
)
2
𝑑𝑦 1 1
2 =− 2 (𝑦) − 𝐸𝐼
𝑘𝑝𝑝 + 𝐸𝐼 𝐿𝑖
𝑥.
𝑑𝑥 𝐿𝑖

2
π Φ𝑖
( )
2
Let me rearrange it further
𝑑𝑦
+ (𝑦) =
1 ⎡𝑘 + 𝑘 𝑥
− 𝑘𝑝𝑝⎤⎥ we will call this as
2 2 𝐸𝐼 ⎢ 𝑝𝑝 𝑞𝑝 𝐿𝑖
𝑑𝑥 𝐿𝑖 ⎣ ⎦
equation 7 a and this as 7 b.

503
(Refer Slide Time: 27:38)

𝑦 = 𝐴 sin 𝑠𝑖𝑛 ( )
α𝑖𝑥
𝐿𝑖
+ 𝐵 cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
+
𝐿𝑖
2
α𝑖 (( 𝑘𝑝𝑝 + 𝑘𝑞𝑝 ) 𝐿𝑥 𝑖
)
− 𝑘𝑝𝑝

α𝑖 = π ϕ𝑖

2
𝐿𝑖
𝐼)𝐵𝐶: @𝑥 = 0, 𝑦 = 0, 𝑤𝑒 𝑔𝑒𝑡 𝐵 = 2 𝑘𝑞𝑝
α𝑖 𝐸𝐼

2
𝐿𝑖
𝐼𝐼)𝐵𝐶: @𝑥 = 𝐿𝑖, 𝑦 = 0, 𝑤𝑒 𝑔𝑒𝑡 0 = 𝐴 sin 𝑠𝑖𝑛 α𝑖 + 𝐵 cos 𝑐𝑜𝑠 α𝑖 + 2 𝑘𝑞𝑝
α𝑖 𝐸𝐼

Now, look at this equation 7 b it is a second order differential equation where right hand side
is not equal to 0. So, this will have two solutions one is a complementary function are there is
a particular integral. So, the solution of equation 7 b will have two components.

Namely; complementary function and particular solution is a standard procedure which you
will have, described well in many mathematic books on differential equations. So, we will
not cover that discussion at this moment in this lecture, but I request you to go through
additionally some material to learn this.

So, therefore, now the solution y will be equal to 𝐴 sin 𝑠𝑖𝑛 ( ) α𝑖𝑥
𝐿𝑖
+ 𝐵 cos 𝑐𝑜𝑠 ( ) α𝑖𝑥
𝐿𝑖
this is
2

the complementary function then the particular solution comes +

𝐿𝑖
2
α𝑖 (( 𝑘𝑝𝑝 + 𝑘𝑞𝑝 ) 𝐿𝑥 𝑖
)
− 𝑘𝑝𝑝 ,

504
we call equation number 8 the standard solution which can be easily understood from
ordinary differential equation textbooks.

We will also say where α𝑖 which is introduced here in the above equation is π ϕ𝑖, where ϕ𝑖

already is here we have derived we have defined that. Here is actually the number or the ratio
between the axial force supply and the Euler’s load that is what we have said in equation 4.

We will call this as 8 a for example, this is 8 b for our understanding. Now, this equation 8 a
has got 2 differential equation coefficients that is a and b we need to find so; obviously, we
need to apply the boundary conditions. Let us apply the boundary condition first boundary
condition at x equals 0, y is 0, that is what it is? See, this figure at x equals 0, y 0, is it not,
that is this point.

So, if you apply that condition, we get B because this term go away this term becomes unity,
this term goes away and you get this which is negative. So, can I say B is going to be equal to
2
𝐿𝑖
2 𝑘𝑞𝑝, can I say that. Let us apply the second boundary condition, what is that condition,
α𝑖 𝐸𝐼

at x is equal to l i again y is 0 see the original beam at x is equal to l that is here, again y is 0,
is it not?

So, let us apply that condition here. So, if you apply that condition, we will get 0 equals
2
𝐿𝑖
𝐴 sin 𝑠𝑖𝑛 α𝑖 + 𝐵 cos 𝑐𝑜𝑠 α𝑖 + 2 𝑘𝑞𝑝, because 𝐿𝑖 and 𝑥𝑖 gets cancelled because x is equal
α𝑖 𝐸𝐼

to l is a condition this is what I get. We already have the value for B, let us substitute that
here.

505
(Refer Slide Time: 33:10)

2
𝐿𝑖
− 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) = 𝐴 sin 𝑠𝑖𝑛 (α𝑖) + 𝐵 cos 𝑐𝑜𝑠 α𝑖( )
2
𝐿𝑖
− 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) − 𝐵 cos 𝑐𝑜𝑠 (α𝑖) = 𝐴 sin 𝑠𝑖𝑛 α𝑖( )
2 2
𝐿𝑖 𝐿𝑖
− 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) − α𝑖 𝐸𝐼
2 cos 𝑘𝑝𝑝𝑐𝑜𝑠 α𝑖 ( ) ( )
= 𝐴 sin 𝑠𝑖𝑛 α𝑖

−
𝐿𝑖
⎡ 𝑘𝑞𝑝+cos𝑘𝑝𝑝𝑐𝑜𝑠 (α𝑖) ⎤ = 𝐴
2 ⎢
α𝑖 𝐸𝐼 ⎣ sin𝑠𝑖𝑛 (α𝑖) ⎥
⎦
2
𝐿𝑖
Now, substituting for B and simplifying,− 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) = 𝐴 sin 𝑠𝑖𝑛 (α𝑖) ( )
+ 𝐵 cos 𝑐𝑜𝑠 α𝑖

2
𝐿𝑖
.Let us do this. So, − 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) − 𝐵 cos 𝑐𝑜𝑠 (α𝑖) ( )
= 𝐴 sin 𝑠𝑖𝑛 α𝑖 . we will name these

equations we will call this as 9 a, this is 9 b, this is 9 c, this is 9 d. So, now I can say
2 2
𝐿𝑖 𝐿𝑖
− 2
α𝑖 𝐸𝐼
(𝑘𝑞𝑝) − 2
α𝑖 𝐸𝐼
cos 𝑘𝑝𝑝𝑐𝑜𝑠 α𝑖 ( ) = 𝐴 sin 𝑠𝑖𝑛 α𝑖 , ( ) can I say that?

𝐴 =−
𝐿𝑖
⎡ 𝑘𝑞𝑝+cos𝑘𝑝𝑝𝑐𝑜𝑠 (α𝑖) ⎤;
2 ⎢
α𝑖 𝐸𝐼 ⎣ sin𝑠𝑖𝑛 (α𝑖) ⎥
⎦

506
(Refer Slide Time: 35:53)

2
𝐿𝑖
− 2
α𝑖 𝐸𝐼
[𝑘𝑞𝑝 csc 𝑐𝑠𝑐 (α𝑖) + 𝑘𝑝𝑝 cot 𝑐𝑜𝑡 (α𝑖) ] = 𝐴
2

𝑦 = 𝐴 sin 𝑠𝑖𝑛 ( ) α𝑖𝑥

𝐿𝑖
+ 𝐵 cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
+
𝐿𝑖
2
α𝑖 𝐸𝐼 (( 𝑘𝑝𝑝 + 𝑘𝑞𝑝 ) 𝐿𝑥𝑖
− 𝑘𝑝𝑝)
2
α𝑖 𝐸𝐼

𝐿𝑖
2 [
(𝑦) =− 𝑘𝑝𝑝 cot 𝑐𝑜𝑡 (α𝑖) + 𝑘𝑞𝑝 csc 𝑐𝑠𝑐 (α𝑖) sin 𝑠𝑖𝑛 ] ( )
α𝑖𝑥
𝐿𝑖
+ 𝑘𝑝𝑝 cos 𝑐𝑜𝑠 ( ) ((
α𝑖𝑥
𝐿𝑖
+ ) 𝐿𝑥
𝑘𝑝𝑝 + 𝑘𝑞𝑝
𝑖
−

2
𝐿𝑖
So, which means− 2
α𝑖 𝐸𝐼
[𝑘𝑞𝑝 csc 𝑐𝑠𝑐 (α𝑖) + 𝑘𝑝𝑝 cot 𝑐𝑜𝑡 (α𝑖) ] = 𝐴 . So, let me call this as 9
f. So, I have the value for A I have the value for B also see B is available here, A is available
here. So, substituting for A and B in the original solution we get, what do we get? Please
check that, what do we get? Where is the original solution? This is the original solution I
have, this is the original solution I have. So, let us rearrange the terms, let us I think you can
even copy this let us try to copy this.

Let us remove this marking you know the number also which is equal to α𝑖 because you know
2
A has a multiplier α𝑖 , if you look at B it also has a multiplier and this term also has a

multiplier. So, if I keep all this common out and take it to the left hand side this will become
2

𝑦 = 𝐴 sin 𝑠𝑖𝑛 ( )
α𝑖𝑥
𝐿𝑖
+ 𝐵 cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
+
𝐿𝑖
2
α𝑖 𝐸𝐼 (( 𝑘𝑝𝑝 + 𝑘𝑞𝑝 ) 𝐿𝑥 𝑖
)
− 𝑘𝑝𝑝 .

507
(Refer Slide Time: 40:31)

2
α𝑖 𝐸𝐼

𝐿𝑖
2 ( ) =− [𝑘
𝑑𝑦
𝑑𝑥
α𝑖
𝐿𝑖 𝑝𝑝
cot 𝑐𝑜𝑡 (α𝑖) + 𝑘𝑞𝑝 csc 𝑐𝑠𝑐 (α𝑖) − ]
α𝑖
𝐿𝑖
𝑘𝑝𝑝 sin 𝑠𝑖𝑛 ( ) ((
α𝑖𝑥
𝐿𝑖
+ 𝑘𝑝𝑝 + 𝑘𝑞𝑝 ) 𝐿1 )
𝑖

2
α𝑖 𝐸𝐼

𝐿𝑖
2 ( )= 𝑘
𝑑𝑦
𝑑𝑥
⎡1 − α sin 𝑠𝑖𝑛
𝑝𝑝 ⎢
⎣ 𝑖 ( )
α𝑖𝑥
𝐿𝑖
− α𝑖 cot 𝑐𝑜𝑡 (α𝑖) cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
⎤ + 𝑘 ⎡1 − csc 𝑐𝑠𝑐 (α ) cos 𝑐𝑜𝑠
⎥
⎦ 𝑞𝑝⎢
⎣ 𝑖 ( α

Differentiating with respect to x equation 10 we get equation 11.

So, we have 2 equations now for the displacement and for the slope ( )is slope is it not. So,
𝑑𝑦
𝑑𝑥
2
α𝑖 𝐸𝐼
let us rearrange this term this equation in a closed form which will now become
𝐿𝑖
2 ( )
𝑑𝑦
𝑑𝑥

will be equal to because there are many 𝑘𝑝𝑝 terms let us group them all.

So, 𝑘𝑝𝑝 term here, there is a 𝑘𝑝𝑝 term here, let us group them which will become 𝑘𝑝𝑝 times of

⎡1 − α sin 𝑠𝑖𝑛
⎢
⎣ 𝑖 ( ) α𝑖𝑥
𝐿𝑖
− α𝑖 cot 𝑐𝑜𝑡 (α𝑖) cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
⎤.
⎥
⎦

Let us talk about 𝑘𝑞𝑝 of⎡⎢1 − csc 𝑐𝑠𝑐 (α𝑖) cos 𝑐𝑜𝑠
⎣ ( )
α𝑖𝑥
𝐿𝑖
⎤, you may wonder why we are
⎥
⎦
arranging the equation 11 in this order. Please understand, I want to obtain 𝑘𝑝𝑝 and 𝑘𝑞𝑝

ultimately that is how we already derived the rotation coefficients for the fixed beam when
we do the stiffness analysis, is it not? So, you want to group it in that order, we are trying to

508
do that let us call this as equation 11 a because equation 11 and 11 a are only rearranging of
terms with some simplification.

(Refer Slide Time: 45:54)

So, friends we look at the summary what we discussed in this lecture. In this lecture we are
learning to derive the stability function for beam under axial compression. We have followed
identically the same procedure with which we derive the stiffness matrix of a fixed beam.

But in this case, we applied axial compressive force 𝑃𝑎 and deriving the stability function.

So, while doing so, we made a simplification of expressing this particular as a function of
Euler load. Why we did that, because we wanted to assess stability and Euler’s load will help
you to give stability because I can quickly compare 𝑃𝑎 with 𝑃𝐸 and say whether the member

is stable or unstable. We already learnt it in the previous lecture how to define stability in
Euler’s criterion. Therefore, we are trying to use this relationship and derive the stability
functions.

We will continue this derivation and do this stability functions in the coming lectures.

Thank you have a good day.

509
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 30
Stability functions - 2

(Refer Slide Time: 00:21)

Friends, welcome to the lecture 30 on Advanced Steel Design. We will continue to discuss
Stability Functions under axial compression, we will call selection number II on the same
perspective.

510
(Refer Slide Time: 00:52)

( ) =− ( )[𝑘 ( ) ] cos 𝑐𝑜𝑠 ( 𝐿 ) ( ) ( )

α𝑖 𝐸𝐼 α𝑖 α𝑖𝑥 α𝑖 α𝑖𝑥

𝐿𝑖
2
𝑑𝑦
𝑑𝑥 𝐿𝑖 𝑝𝑝 ( )
cot 𝑐𝑜𝑡 α𝑖 + 𝑘𝑞𝑝 csc 𝑐𝑠𝑐 α𝑖
𝑖
− 𝐿𝑖
𝑘𝑝𝑝 sin 𝑠𝑖𝑛 𝐿𝑖 (
+ 𝑘𝑝𝑝 +

2
α𝑖 𝐸𝐼
𝐿𝑖 ( )= 𝑘
𝑑𝑦
𝑑𝑥
⎡1 − α sin 𝑠𝑖𝑛
𝑝𝑝⎢
⎣ 𝑖 ( )
α𝑖𝑥
𝐿𝑖 ( )
− α𝑖 cot 𝑐𝑜𝑡 α𝑖 cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
⎤ + 𝑘 ⎡1 − α csc 𝑐𝑠𝑐 α cos 𝑐𝑜𝑠
⎥
⎦ 𝑞𝑝⎢
⎣ 𝑖 𝑖 ( ) (
In the last lecture, we derived equation 11 (a) which is expressed in terms of 𝑘𝑝𝑝 and 𝑘𝑞𝑝. So,

here on the left-hand side you will see there is a term representing the slope ( ) is the slope
𝑑𝑦
𝑑𝑥

of the v, is it not which is given as a function of 𝑘𝑝𝑝 and 𝑘𝑞𝑝 which is given by the equation 11

(a).

511
(Refer Slide Time: 02:27)

2
α𝑖 𝐸𝐼
𝐿𝑖 ( )= 𝑘
𝑑𝑦
𝑑𝑥
⎡1 − α sin 𝑠𝑖𝑛
𝑝𝑝⎢
⎣ 𝑖 ( )
α𝑖𝑥
𝐿𝑖 ( )
− α𝑖 cot 𝑐𝑜𝑡 α𝑖 cos 𝑐𝑜𝑠 ( )
α𝑖𝑥
𝐿𝑖
⎤ + 𝑘 ⎡1 − α csc 𝑐𝑠𝑐 α cos 𝑐𝑜𝑠
⎥
⎦ 𝑞𝑝⎢
⎣ 𝑖 𝑖( ) (
We will copy this equation to the next screen and we anyway replace this with black color.
So, we will do this, 11 (a).

𝑑𝑦
@x=0, 𝑑𝑥
= θ𝑝 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑒𝑞𝑢𝑎𝑙 𝑡𝑜 𝑢𝑛𝑖𝑡𝑦.

𝑑𝑦
@x=L, 𝑑𝑥
= Zero.

Substitute the above condition,

[ ( ) ( )
0 = 𝑘𝑝𝑝 1 − α𝑖 sin 𝑠𝑖𝑛 α𝑖 − α𝑖 cot 𝑐𝑜𝑡 α𝑖 cos 𝑐𝑜𝑠 α𝑖 ( ) ] + 𝑘𝑞𝑝[1 − α𝑖 csc 𝑐𝑠𝑐 (α𝑖) cos 𝑐𝑜𝑠 (α𝑖) ]
𝑑𝑦
Now, we know at x equal 0, 𝑑𝑥
= θ𝑝 because that is how we gave the rotation. This was the

beam under axial compression 𝑃𝑎 and we gave a rotation which is unity and this was θ𝑝 we

gave.

That is how we generated 𝑘𝑝𝑝 and we say 𝑘𝑞𝑝 δ𝑟𝑝 or 𝑘𝑟𝑝 and 𝑘𝑠𝑝, that is how we did, is it not.

So, we are given θ𝑝 which is equal to of course, unity, but we have θ𝑝 . So, let us say θ𝑝

512
 𝑑𝑦
which is equal to it whereas, at x is equal to L the slope is 0, see here. @x=L, 𝑑𝑥
= Zero. So,

let us substitute this condition and see what happens; substituting the above conditions.

So, let us say slope is 0. So, that is going to happen at x is equal to L. So, let us write that.
[ ( ) ( )
0 = 𝑘𝑝𝑝 1 − α𝑖 sin 𝑠𝑖𝑛 α𝑖 − α𝑖 cot 𝑐𝑜𝑡 α𝑖 cos 𝑐𝑜𝑠 α𝑖 ( ) ] + 𝑘𝑞𝑝[1 − α𝑖 csc 𝑐𝑠𝑐 (α𝑖) cos 𝑐𝑜𝑠 (α𝑖) ]
, Equation 12a.

(Refer Slide Time: 05:04)

( ) ⎤
cos𝑐𝑜𝑠 α𝑖
0 = 𝑘𝑝𝑝 − 𝑘𝑝𝑝α𝑖⎡⎢sin 𝑠𝑖𝑛 α𝑖 ( ) + cos 𝑐𝑜𝑠 α𝑖 ( ) sin𝑠𝑖𝑛 (α ) ⎥ + 𝑘𝑞𝑝 − 𝑘𝑞𝑝α𝑖 cot 𝑐𝑜𝑡 (α𝑖)
⎣ 𝑖 ⎦

This was the equation, 12a. By simplifying,

(
0 = 𝑘𝑝𝑝 1 − α𝑖 ( 1
( )
sin𝑠𝑖𝑛 α𝑖 )) ( )
+ 𝑘𝑞𝑝(1 − α𝑖 cot 𝑐𝑜𝑡 α𝑖 )

α −sin𝑠𝑖𝑛 α𝑖
𝑘𝑞𝑝 = ⎡⎢ sin𝑠𝑖𝑛 α𝑖 −α cos𝑐𝑜𝑠 ⎤𝑘
⎥ 𝑝𝑝
⎣ ( 𝑖) 𝑖 (α 𝑖) ⎦

This was equation number 13; is very simple because you know this is expressed as cos by
sin. There is sin here, take a common denominator, multiply that with 0, the denominator
goes away.

So, rearranging you will get 𝑘𝑞𝑝. So, 𝑘𝑞𝑝 is given by equation 13, that is one of the coefficient

we have, rotation coefficient.

513
(Refer Slide Time: 09:43)

𝑑𝑦
@𝑥 = 0, 𝑑𝑥
= 1

𝑘𝑝𝑝 =
𝐸𝐼 ⎡ α𝑖sin𝑠𝑖𝑛 (α𝑖)−α𝑖cos𝑐𝑜𝑠 (α𝑖) ⎤
𝐿𝑖 ⎢ 2(1−cos𝑐𝑜𝑠 (α ))−α sin𝑠𝑖𝑛 (α ) ⎥
⎣ 𝑖 𝑖 𝑖 ⎦

Similarly, next condition at x is equal to 0, the slope is 1. So, substitute this condition in the
original equation and you will find now 𝑘𝑝𝑝. Call this is equation number 14. So, now, I can

express the rotation functions in terms of stiffness coefficients. Let us do that.

(Refer Slide Time: 11:24)

514
 α𝑖sin𝑠𝑖𝑛 (α𝑖)−α𝑖cos𝑐𝑜𝑠 (α𝑖)
𝑘𝑝𝑝 = 𝑟𝑖 = ⎡⎢ 2(1−cos𝑐𝑜𝑠 ⎤
α𝑖))−α𝑖sin𝑠𝑖𝑛 (α𝑖) ⎥
⎣ ( ⎦
α −sin𝑠𝑖𝑛 (α𝑖)
𝑘𝑞𝑝 = 𝑐𝑖 = ⎡⎢ sin𝑠𝑖𝑛 𝑖α −α cos𝑐𝑜𝑠 ⎤
α𝑖) ⎥
⎣ ( 𝑖) 𝑖 ( ⎦

One can express the stiffness coefficients, that is 𝑘𝑝𝑝 and 𝑘𝑞𝑝. These are rotation coefficients;

as a function of rotation functions 𝑟𝑖 and 𝑐𝑖.

Friends, if you remember 𝑘𝑝𝑝 this was some function of EI by L and 𝑘𝑞𝑝 is also some function

of EI by L. Here also you look at this equation 𝑘𝑝𝑝 is a function of EI by L and 𝑘𝑞𝑝 is a

function of 𝑘𝑝𝑝 which is again EI by L. So, I am just trying to express this, only these 𝑘𝑝𝑝 as

𝑟𝑖 and 𝑘𝑞𝑝 as 𝑐𝑖. We will call this equation as 15 (a) and 15 (b). These are called stability

functions, that is rotation functions for stability.

They are called as 𝑟𝑖 and 𝑐𝑖 are called rotation functions for compressive axial load case. I can

take a special case of axial load zero and see what happens to these functions. Let us take a
special case of axial load zero and see what happens.

(Refer Slide Time: 14:58)

Special case of axial load zero, that is 𝑃𝑎 is set to 0. So, what does it mean? Immediately, it

implies a fact that α𝑖 will become zero. So, when α𝑖 becomes zero, look at the rotation

515
coefficients which we just now derived. There is a possibility of 0 by 0. So, one need to apply
L’ Hospital’s rule, let us explain the L’ Hospital’s rule very briefly for our learning, though
we have learnt it in mathematics, but still, it is important.

(Refer Slide Time: 16:08)

Suppose, f(x) and g (x) are differential functions and g dash of x is not equal to 0 on an open
interval which contains that is except at a following conditions apply. Suppose, limit x tends
to a, f(x)= 0, limit x tends to a, g(x) is 0, or limit x tends to a, f(x) is plus or minus infinity,
limit x tends to a g(x) is plus or minus infinity. Then, it may reduce to a form 0 by 0 or
infinity by infinity, under such conditions following equation is good.

516
(Refer Slide Time: 18:32)

Limit x tends to 0, f(x) by g(x) is equal to limit x tends to a f dash of x g dash of x, if the right
hand side of the equation exists. Let us say for example, limit x tends to 0 e power x minus 1
by x square plus x is our 0 by 0 form, but the denominator derivative is it not 0, is it not it
exists. So, I can say is equal to limit x tends to 0 d by dx of the numerator d by dx of the
denominator which is going to be limit x tends to 0 e power x by 2 x plus 1. Now, the answer
will be 1, for which the condition is the right-hand side should exist.

So, L’ Hospital rule uses derivatives to evaluate the limits involving indeterminate forms. It
states that for indeterminate functions, where the unity tends to form 0 by 0 or infinity by
infinity, the limit of that form is equal to the limited derivative itself.

For 0 by 0 form or this form, limit of the form is equal to limit of the derivative. L’ Hospital
rule can be applied any number of times until the function does not reduce to a condition
back again to 0 by 0 or infinity by infinity. So, we have a problem which is to be used in this
specific case. So, let us consider the rotation functions 15 (a) and 15 (b).

517
(Refer Slide Time: 22:07)

α𝑖sin𝑠𝑖𝑛 (α𝑖)−α𝑖cos𝑐𝑜𝑠 (α𝑖)

𝑘𝑝𝑝 = 𝑟𝑖 = ⎡⎢ 2(1−cos𝑐𝑜𝑠 ⎤
α𝑖))−α𝑖sin𝑠𝑖𝑛 (α𝑖) ⎥
⎣ ( ⎦
α −sin𝑠𝑖𝑛 (α𝑖)
𝑘𝑞𝑝 = 𝑐𝑖 = ⎡⎢ sin𝑠𝑖𝑛 𝑖α −α cos𝑐𝑜𝑠 ⎤
α𝑖) ⎥
⎣ ( 𝑖) 𝑖 ( ⎦

So, let us copy those functions here. Let us rub this. As α𝑖 reduces to zero or approaches zero

because we are considering no axial load case, is not that; both f(α𝑖) by g(α𝑖) approaches

zero. So, this will turn to a 0 by 0 form, is it not. So, I can apply L’ Hospital rule which will
give me the function value as 𝑟𝑖 will now become that is 𝑟𝑖 at ϕ𝑖 equal 0 will become 4 and 𝑐𝑖

at ϕ𝑖 becomes 0 is 0.5.

(Refer Slide Time: 24:13)

518
So, at 𝑃𝑎 equals 0, the stiffness coefficients will reduce to the conventional carryover factors.

𝐸𝐼
𝑘𝑝𝑝 = 𝑟𝑖 𝐿𝑖

𝑘𝑞𝑝 = 𝑐𝑖𝑘𝑝𝑝 = 𝑐𝑖𝑟𝑖 ( )

𝐸𝐼
𝐿𝑖

𝑘𝑝𝑝+𝑘𝑞𝑝 𝑟𝑖𝐸𝐼
𝑘𝑟𝑝 = 𝐿𝑖
=
𝐿𝑖
2 ( 1 + 𝑐 𝑖)
𝑘𝑠𝑝 =− 𝑘𝑟𝑝

Let us call this equation as equation 16. Similarly, friends by applying any rotation the kth
end, another set of stiffness coefficients can be derived.

519
(Refer Slide Time: 25:58)

𝐸𝐼
𝑘𝑝𝑞 = 𝑐𝑖𝑟𝑖 𝐿𝑖

𝐸𝐼
𝑘𝑞𝑞 = 𝑟𝑖 𝐿𝑖

𝑡ℎ
Now, by applying unit rotation of the 𝑘 end, another set of stiffness coefficients can be
derived. Let us draw the figure and understand that I am trying to apply unit rotation at the
𝑡ℎ
𝑘 end. So, this is θ𝑞 which is unity and we know the length of the beam is 𝐿𝑖 and we have

already applied the axial load 𝑃𝑎. And, this is 𝑘𝑝𝑞, this is 𝑘𝑞𝑞 and this force of coefficient is

𝑘𝑟𝑞 and this is 𝑘𝑠𝑞.

So, I can again cut a section here, draw the free body diagram. 𝑃𝑎 is applied here and 𝑃𝑎 is

reciprocated here. And, there is a moment 𝑘𝑞𝑞 and there is force 𝑘𝑠𝑞 which is actually equal to
𝑘𝑝𝑞+𝑘𝑟𝑞
− 𝐿𝑖
𝑘𝑞𝑞. Why it is minus? I think you realize it there is a net moment of 𝑘𝑝𝑞 + 𝑘𝑞𝑞.

Here, the counteract that this will be the couple and this is opposite to 𝑘𝑠𝑞. Therefore, there is

a minus sign here.

So, you can follow the same logic and we can write the coefficients directly as here. I am
leaving it for learning, you easily do that.

520
(Refer Slide Time: 29:49)

So, friends we have learnt the derivation of stability function, that is rotation coefficients
which are 𝑐𝑖 and 𝑟𝑖 for two cases. Case i, under axial compressive load 𝑃𝑎, ii when 𝑃𝑎 is tend

to 0 and we realize that in both the cases 𝑃𝑎 is expressed as a function of PE. Why? We are

looking for stability functions. So, please look at the derivation back again and try to
understand that is very simple.

It has got involvement of differential equation understanding, have a parallel reading on some
text book on engineering mathematics. And, learn this and try to get a hold of the derivation.
The next lecture we are going to discuss about the rotation function and stability functions of
a beam under axial tensile load.

Thank you very much friends. Have a good day.

521
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 31
Stability functions - 3

(Refer Slide Time: 00:22)

Friends, welcome to the lecture 31 on Advanced Steel Design course, we are going to discuss
Stability functions. In the last lecture there is a small correction I apologize for that, let us
rewrite that equation for kpp and kqp, please turn back your notes I am rewriting this equation
there is a small correction. So, kpp is given by

So, the error what we had is that this was missing. So, it is written as αi sin αi - αi cos αi,
divided by twice of 1 - cos αi - αi sin α this was written.

So, and of course, kqp there is no change in that,

522
there is no change in this. The error was I missed out this bracket when we did the derivation,
I mean you could have observed it, but; however, does not matter.

(Refer Slide Time: 03:20)

So, then subsequently we said the rotation coefficients can be expressed as ri and ci. So, ri
again becomes

And of course, ci is

We applied this for zero axial load is a special case and we derived it using L’Hospital’s rule.
And we got for zero axial load, we got ri at ϕi 0 special case is 4 and ci at ϕ i 0 special case is
0.5 I think we derived it.

523
(Refer Slide Time: 05:15)

Furthermore, friends we made a very interesting and comprehensive statement saying that
kpp is ri of EI by Li kqp is ci of kpp and krp is kpp + kqp by Li which becomes ri EI by Li
square of 1 + ci. And we already know ksp is - of krp, this is actually the standard case where
this is kpp, this is kqp, this is krp, and this is ksp, and these are the values what we have
derived. Similarly, we extended this concept for unit rotation at the kth end.

(Refer Slide Time: 06:22)

Then we also got we also obtained the following relationships, we got kpq as ci, ri, EI by Li
and kqq as ri EI by Li and so on right. So, this is simple I am giving you know rotation here,

524
this was kqq, this is kpq, this is krq, and this is ksq, the second subscript stands for the degree
of freedom where we have given unit rotation or displacement. The first subscript stands for
the degree of freedom where you are measuring the force.

I mean there is a standard logic we have in the stiffness derivation; we follow the same thing
here there is even not a millimetre change in the whole concept of derivation. That is the idea
why we presented the derivation of the fixed beam first, then we started extending that same
algorithm and logic for deriving the stability functions for Euler’s load, that is the idea we
have here. Having said this; now, let us start focusing on deriving rotation functions under
axial tensile load, let us go ahead.

(Refer Slide Time: 08:03)

Now, we are going to do the 3rd part where I am deriving the rotation functions, they are
actually stability functions for axial tensile load. If a beam is subjected to axial tension, then;
obviously, ϕ i becomes negative is not that I have a beam, I am subjecting this axial tensile
load of P a, I am looking for the stability of this. So, if the beam is under axial tension, then
we can say my ϕ i is negative.

So, therefore, I now derive or describe a new variable β i which is now going to be 𝜋times of
root of - ϕ i, which I can write as i 𝜋 ϕ I, because you know i square is - 1. We already know
friends this 𝜋 ϕ i is αis it not we already know αi is 𝜋 square root of ϕ i, we have used it in the
earlier derivation as well. So, can i now say β i as i αi; let us have this with us, we call the

525
equation number 1, we also know we also know that sin βi, because there is an imaginary
function here right.

(Refer Slide Time: 10:30)

So, we should say this is equal to e i β i - e - i β i by 2 i, and cos β i is taken as e i β i + e - i β

i by 2, we call equation number 2 a and 2 b. Let us substitute this 2 a and 2 b in this equation
1, we know that β i’s αi. So, let us substitute substituting 1 in equations 2, what do we get?
We get the following sin β i yes e i square αi - e - i square αi by 2i, which makes this as e - αi
- e αi by 2i we call this as equation 3 a now.

Similarly, cos β i will be e i square αi + e - i square αi by 2 which now become e - αi + e αi by

2 which is 3 b. So, the equations 2 a and 2 b got refined to 3 a and 3 b by substituting
equation 1 into them. I mean there is no confusion in this it is a straight away clear derivation,
I hope you understand and follow the steps clearly.

Now, what I do is, we want to give rotation constants at the jth here let us work out the
rotation constants we already have them with us we just now wrote them also right. So, we
have rotation constants, we have rotation constants ri and ci; so, let us copy this copy these
two equations write it here we already have.

526
(Refer Slide Time: 13:44)

So, we know the rotation constants which are useful in stability functions . So, now, what I
am going to do is, I am going to replace them as β i instead of αi am going to replace in β i
right; so, I am going to do that. So, α β i will be i αi you can see that β i will be i αi; so, I am
going to just replace all αs into β.

So, replace all αi to β i that is what I am trying to do. So, for example, this will become i β i
mean i αi and so on. And this will become sin β i; so, i must get this equation extended there,
I mean this equation extended there and so on is not that. So, αi to β i I keep on doing this
right; so, let us do that please do this I will do it parallelly here.

527
(Refer Slide Time: 15:16)

So, my ri will now become i αi e to the power of - αi - e αi by 2 i, this is 1 - you can see here
as the first term second term here - i αi, because i have replaced as β i here i αi of. So, I
should have cos β i which is here, which is here, this is here. So, let us do that, e to the power
of - αi + e to the power of αi by 2 am I right, let me close this bracket divided by twice of you
know divided by twice of this.

So, let us do that 2 of 1 - e to the power of - αi + e to the power of αi by 2 - see this is again β
i which i say yes, i αi of then this is sin I have the value which is e to the power of - αi - e αi
by 2 i. Similarly, let us write down for ci also we have ci also here; so, again the same story i
αi see here my αi becomes β i.

So, - sin e to the power of - αi - αi by 2i divided by this one, e to the power of - αi - e αi by 2 i

- this value i αi of cos. So, e to the power of - αi + e αi by 2, we call this as equation numbers
4 a and 4 b, 4 a and 4 b which are now the rotation constants, for axial tensile load 4 a and 4 b
are rotation constants for stability function for axial tensile load.

Now, I can simplify these two equations further I can simplify, because there are I can just
show you a simple manipulation. Let us say if you do this, I will copy this equation to the
next sheet then I will do that I will copy this equation to the next sheet then I will do this.

528
(Refer Slide Time: 19:52)

Let us do this a simplification how let us say this i and this i gets cancelled and this i with this
becomes - 1 and that with this becomes +; so, these are the changes what I am going to
expect when I write. So, I am writing this new ri as, αi times of e - αi - e αi by 2 + αi times of
e - αi + e αi by 2, am I right divided by 2 times of 1 - e - αi + e αi by 2 -, this i again goes
away.

So, I can straight away say αi times of, let me write it here - αi times of e - αi - αi by 2 am I
right, that is my ri. Let us write down the function for ci; so, I can now take a constant and
say αi + e to the power of - αi - αi by 2; so, I make changes to +.

So, what happens is if you multiply this i with this i get - here there is already a - here, I am
putting that - the denominator there is a reason. See this mathematical simplification is very
important, because this is how the standard expression will be reached as you see in the
literature; so, let us carefully look at this.

So, I am putting the - here then I am saying this equation is I think incomplete it should be
there is a 2i here there is a 2i here, I think I have not copied that - of e to the power of - αi - αi
by 2, I am removing this i . When you do that, this becomes - i gets cancelled, I am taking i
constant out right - αi times of e - α + e α by 2 that is my ci call this as equation number 5 a
and 5 b friends.

529
(Refer Slide Time: 24:01)

Now, I can rewrite this in hyperbolic forms, now rewriting 5 a and 5 b in hyperbolic forms.
What is an hyperbolic form? We know that sin hyperbolic αi is e αi - αi by 2, cos hyperbolic
αi is e αi + αi by 2. So, I am using this relationship, I am writing ri and ci like this.

So, ri now becomes αi times of αi cos hyperbolic αi - sin hyperbolic αi by twice of 1 - cos
hyperbolic αi + αi sin hyperbolic α, you can look at this simplification. And ci now becomes
αi - sin hyperbolic αi by sin hyperbolic αi - αi cos hyperbolic α. So, we call this is equation
number 6a and 6b, these are expression of rotation coefficients in terms of hyperbolic
functions.

Now, interestingly friends we can plot this rotation coefficients, we can plot these rotation
coefficients under axial tensile and axial compressive loads for a wide range of ϕ is a variable
ϕ is a variable. These typical tables are available in the literature of stability books, that is
called stability tables I can say stability constants they are also called as stability charts.

530
(Refer Slide Time: 27:01)

I will show you a typical chart here; so, for different values of ϕ r and c are plotted, I mean
given the table for closer intervals. You can also do that I will also show you a figure and
then I show you a MATLAB program, which can plots this, but let us say it is a continuous
table which starts from a specific value. So, negative in the table indicates tensile load,
positive ϕ in the table that is this value; if it is negative tensile, positive in the table indicates
compressive load.

Please understand both of these loads are axial, they are axial loads we are talking about
stability functions. Now where are they used these charts are useful to estimate the critical
buckling load you may wonder, why I am talking about buckling load estimate very
interesting question, buckling load will help you to estimate Pa. Now, Pa will be compared
with buckling load and if Pa exceeds buckling load then it is a failure is not that.

So, to ascertain whether the structure will have a stability failure under the given load you
must know, what is the capacity of the structure to sustain its stability? So, that is available
from Pcr we already discussed that in the stability theory earlier Euler’s criteria. So, now, I
can estimate this ϕ I mean Pa b tensile or compressive from this table are charts compare it
with critical load, then I will know critical load will help you to estimate from the charts axial
load is the one which is applied to the system.

We can now compare and identify whether the structure will have destabilized mode under
this Pa. So, remember stability starts will not help you to give Pa, Pa is a known input,

531
stability charge will help you to calculate critical buckling load; so, we have done that; so,
now, interestingly we can also plot these charts.

(Refer Slide Time: 30:16)

(Refer Slide Time: 30:33)

But before that let us also do the translation coefficients, let us also do the translation
coefficients. So, let us do the translation coefficients, let us say under axial compressive load.
So, how do you do a translation coefficient or translation function? Initially I will have a
beam both ends fixed; I will give unit translation to this end the beam will move here, this
will be unit translation I call this as del r unity.

532
Now, the beam will take this position; so, if this angle is 1 by Li this is a Li, we will agree
upon this fact that this angle is also 1 by Li, this angle is also 1 by Li there is no dispute in
this . Let us now mark the moments and shear causing this; so, this will be k pr, this will be
kqr, and this shear will be krr and this shear will be ksr is it not.

Now, let us subject this to an axial load compressive Pa; so, what we have done? We have
given unit translation at jth end under axial compressive load that is what you have done right
that is the fuel. Now, we will take moments about the kth end, taking moment about the kth
end, where is the kth end? This is my jth end, this is my kth the end of the beam right we get.

Let us see what do we get Pa into 1 that is clockwise - k pr antic clockwise + krr into Li
clockwise - kqr should be 0 sigma m . So, from this I can write krr as P a into 1 - of kpr + kqr
by Li, what is this actually very clear this is the unbalanced moment in the section right. So,
let us remember this we call this equation number I think 5 a 5 b let us call the 6 let us call
the 6 a and 6 b.

(Refer Slide Time: 34:36)

Now, when have a member with unit rotation and this was kpp and this is kqp this we know,
and this was krp, and this was ksp, I think we know this . We also know that k pp that kpp is
actually ri times of EI by Li and so on we remember that is it not. We wrote it in the
beginning itself see k p a pp is ri times of this this is ca times of kpp and so on we already
wrote that I am going to make use of this relationship now.

533
Now, friends if this rotation is unity this is kpp, if this rotation is not unity if this rotation is 1
by Li what would be this value. Let us say I know this is going to be k pr and this is going to
be kqr, this is only at the jth end I am talking about. So, can I write k pr as ri times of EI by Li
of 1 by Li, because this is for unity this value for 1 by Li this value, is not it? This is only for
the jth end.

Can I write kqr as ri ci EI by Li times of 1 by Li, where kqr we already have here kqr or kqp
is c i times of kpp; so, I am using that ci times of kpp I am using that.

So, you know kqp is ca times of kpp, I just multiplied this ci you can see that here, this is
only for the jth end; so, I also have a similar contribution if I give unit rotation at the kth end
and so on now. So, now, I can add that contribution here as this is just swapped ri ci EI by Li
times of 1 by Li and this is swapped ri EI by Li of 1 by Li, we already derived this in the last
lecture you can verify that. So, what we did is kpr and kqr we obtained, kpr and kqr we
obtained; so, we have realized that kpr and kqr are equal.

So, we realize that hence kpr is kqr is equal to ri times of 1 + ci times of EI by Li square can I
say that can I say that. So, we call this as equation we said 6 a and 6 b we call this as 7 a, 7 b,
7 c is that agreed. Now, it is very simple for me to find krr, can I say krr is sum of this by the
span can I say that; so, I can say krr is k pr + kqr by Li, can I say that. So, substituting I will
get this as 2 times of ri of 1 + ci of EI by Li cube can I say that.

Now, we should say this is also available to me here; so, comparing these two can I now say,
because 6 b should be equal to this because I am going to equate both of them. So, can I say
that krr is replaced as - Pa by Li, this factor is coming here. And we also know this will be
negative of krr; so, can I say this as - ksr .

534
(Refer Slide Time: 41:04)

So, friends for zero axial load let me name this equation as 7 d, for 0 axial load krr is - ksr is
2 r 1 + c i 2 r i sorry. So, all for ith beam EI by Li cube can I say that, because Pa goes away
that is what I will get equation number 8. Now, we want to make a general form, now to
make a general form both in the presence and absence of axial load krr is - ksr is actually I
am putting a new function ti of 2 ri 1 + ci of EI by Li cube. I am just introducing a new term
this ti is called this equation 9 is called translation function, translation function.

(Refer Slide Time: 42:57)

535
So, now they can say Pa is 𝜋 square ϕ i EI by Li square for n equals 1 . Now, I want to equate
the k rr that is equate 9 equation 9 to or with compare equation 9 with 7 d we will compare
with 7 d . So, I can now say ti times of 2 ri 1 + ci EI by Li cube is equal to 2 ri 1 + ci EI by Li
cube - pa by Li this is what I am writing. So, this gives me ti as 1 - pa by 2 ri of 1 + ci of Li
square by EI equation number 10 substituting for pa, we get ti as 1 - 𝜋 square ϕ i by 2 ri 1 + ci
10 b.

(Refer Slide Time: 45:46)

So, I have now ri, ci, ti equations with me for axial compression, axial tension am I right is it
not. Similarly, I can induce unit translation at kth end and coefficients kps, kqs, krs, kss can
be obtained. What I am trying to do is I am introducing translation here; this is 1 by Li this is
del s which is unity. So, this is going to be 1 by Li 1 by Li and the moments created are kqs,
kps, krs, and kss, this is what.

536
(Refer Slide Time: 47:06)

So, I can compute them I leave it this small exercise for you I am writing the equations for
you. So, kps is given as equal to krs is - ri of 1 + ci EI by Li cube sorry Li square k rs. So, this
is kps, kqs, krs is - of kss which is - 2 ti ri 1 + ci EI by Li cube, we call this equation number
11 a and b, 11 a, 11 friends please derive this get it satisfied before you proceed further.

(Refer Slide Time: 48:01)

So, friends I can now write the stiffness matrix by taking EI constant as this is ri by Li this is
ci ri by Li. This is 1 + ci of ri by Li square this is - of you can easily remember this, friends. If
you remember this and this sum of these 2 by l will be this and - of this is this is as same as

537
stiffness matrix of the steam beam member. Now, to get the second column swap this; so, this
is ci ri by Li, this is simply ri by Li and this is same as that of the previous one previous
column.

Now, to get this sum of these 2 by L we did the same thing in beams also is not that; so, 1 + ci
into ri 𝜋 Li square. Similarly sum of these 2 by L 1 + ci into ri by Li square to get this we
have to add these two and divide by L. So, there I am going to do the transfer function 2 ti ri
1 + ci by Li cube, this ti is new here.

So, this is - 2 ti ri 1 + ci by Li cube, the fourth column is negative of the third column. So, we
have got now the transfer functions which can be quickly plotted using MATLAB, I will
show you the MATLAB program which can be used for this I will show you the MATLAB
program.

(Refer Slide Time: 50:42)

So, you can see the MATLAB program on the screen for plotting the stability indices. So, the
ϕ is varied from this range and the degree is converted to radiance and the now r c t are
computed from the equation.

538
(Refer Slide Time: 51:03)

Then for 0 for tension and for compression for compression all the three are computed then
the chart is prepared.

(Refer Slide Time: 51:12)

This program is available in a reference book download the program run in MATLAB,
MATLAB supports you for NPTEL courses through IIT Madras.

539
(Refer Slide Time: 51:27)

So, please download MATLAB software, use it for this particular course intensively this is
what it is; so, this is the MATLAB program we have. So, if I do that my plot for the rotation
function r c and t appears like this, I wish that you should plot this and see yourself how they
are appeared please see that friends.

(Refer Slide Time: 51:58)

So, in this lecture we learnt the stability functions, we also use the MATLAB program to plot
the stability function, we also prepared the chart for stability functions right. We learnt it for
zero axial load, axial load compression, axial load tension we have developed a generic case.

540
And very interestingly this compares variable with the standard fixed beam derivation
without axial load same algorithm.

So, this is also 4 by 4 and the stability function is also 4 by 4; so, there is an absolute
similarity between these two in the derivation. So, that is the beauty of the whole exercise can
easily remember this and try to teach this in the class in the same order. So, the stability
derivations become very simple like a stiffness analysis of a fixed beam.

Thank you very much friends, have a good day bye.

541
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 32
Buckling and stability

(Refer Slide Time: 00:21)

Friends, welcome to the 32nd lecture on Advanced Steel Design. We are discussing about
derivation of stability functions in this lecture we will continue discuss that with some more
interesting details about, difference between Buckling and stability.

So, friends while discussing the stability functions we already said we will compare a
standard fixed beam which is used in stiffness method of matrix analysis. So, the basic
module which is used to analyze a statically indeterminate beam or a frame or a structure is a
fixed beam with some prismatic cross section and A. So, we have used this as a standard
nomenclature.

542
(Refer Slide Time: 01:50)

And we said that I can give unit rotation at different ends and displacements and try to derive
the stiffness matrix. So, the standard fixed beam has got four degrees of freedom
kinematically that is theta p theta q and del r and del s.

The member is prismatic with its EI, L and A. So now, this is a jth end of the member and kth
end of the member. So, from j running towards k is xm and ym is anti clockwise 90 to xm
this is a standard convention which we have said. Now I want to derive the stiffness
coefficient for this beam which is statically indeterminate it is a fixed beam we know that. So,
we know kij the stiffness coefficient is a force in ith degree of freedom.

For unit displacement in jth degree of freedom keeping all other degrees of freedom
restrained we know this. So, what we did is we gave unit displacement rotation in pth degree,
and we marked these forces as kpp kqp krp and ksp is it not. So, the in kpp for example, here
the first subscript stands for the place or the degree of freedom we are measuring the force
which is p and the second subscript refers to the degree of freedom where you are giving unit
rotation.

So, theta p is unity therefore, its p you will find in figure one the second subscript of all you
know p you know is p. So, unit rotation at jth end unit rotation at kth end unit displacement at
jth end and unit displacement at kth end. So, standard convention we derive the stiffness
matrix for this.

543
(Refer Slide Time: 04:25)

And we said the stiffness matrix for a fixed beam neglecting axial deformation is given by
this. So, 4 by 4 matrix.

So, standard derivation you can refer to my book on analysis of structures using MATLAB
written by for CRC. This book is a very popular book and this gives you the MATLAB
program for doing analysis of indeterminate structures using stiffness method please refer to
this book and MATLAB programs can be downloaded free. I can use this book as an
important reference for a teaching and research as well.

So, this book refers the derivation of this in detail, but let us say this is my p q r and s labels
and we know this is kpp this is kqp and this is sum of these two. So, I should say is kpp + kqp
by l and the fourth one is negative of this. The second column is just the cross of this. So, this
becomes kpq and this is kqq and sum of these two by l will be here, so kpq + kqq by l. Now
fourth one is negative of this.

Coming to the third column of this matrix sum of these two by l will be here incidentally
friends this value and this value are same because stiffness matrix is symmetric ok, but do not
go by that there are possibilities where the stiffness matrix can become asymmetric as well,
so let us not bother about that.

So, sum of these two by l is here which is kqp + kqq by l. Then sum of these two by l further
is here which is kpp + kpq + kqp + kqq by l square, because there is already an l here I am

544
dividing it further by l and the fourth one is negative of this. The fourth column is very
straightforward negative of the third column.

This is standard derivation we have not done this derivation in this course, but it is a standard
derivation available in all standard reference books which are meant for analysis of
indeterminate structures using stiffness method.

So, I am going to borrow the same algorithm and we explain that in the previous lectures I
am just summarizing it now. So, that stability functions can be easily remembered friends that
is very important, because generally stability is an area or is a topic of discussion which is
having a lot of confusion. People do not feel very comfortable to teach in the PG course or
definitely not in UG course UG level and in research it is handled with most complicated
manner.

So, the fundamental reason for this is the stability functions and the approach of stability
problem is not understood properly. So, let us deprive of this particular feeling get out of this
and try to remember this from the basic understanding of a simple statically indeterminate
structure which we already know. So, I am following the same algorithm how we did the
derivation for this. So, what we did is instead of giving the rotations of this.

We started finding out this using flexibility method and we also wrote this equation for us in
the previous lecture.

(Refer Slide Time: 10:00)

545
And we said that my K matrix for a simple fixed beam neglecting axial deformation is given
by this. Let us have the labels as p q r and s this is q ok, let us say we know this is 4 2 6 by l 4
by l 4 by l 6 by l - 6 by l square. So, 2 by l 4 by l 6 by l - 6 by l square, 6 by l square 6 by l
square 12 by l cube - 12 by l cube, - 6 by l square - 6 by l square - 12 by l cube + 12 by l.

I am following the same logic as we have explained here to this. We have used the same form
I want you to remember this form and I will correlate this form to the stability functions now
that is what we are going to do. So, that if you remember this form you can always remember
the stability functions form that is what it is.

(Refer Slide Time: 11:40)

So, what we did is we took the same fixed beam we took the same fixed beam of theta p theta
q del r and del s of A E L I etcetera, but we added one new concept here we said i am
applying an axial force; because I am studying stability, we applied axial force right.

So, we gave unit rotation at jth end then we wrote the free body the equation for the free body
diagram at the displaced position under the presence of axial load we started with axial
compressive load. And we derived we derived the rotational functions that is what we did we
derived rotation function, using second order ordinary differential equation from the three
here flexure and we found out the rotation coefficients from the solution.

546
And we said r i is α i times of sin α i - α i cos α i by 2 times of 1 - cos α i - α i sin α. This was
the r i and c i we derived as α i - sin α i by sin α i - α i cos α we derived this they are called as
rotation functions.

Why rotation functions we derive this by giving unit rotation we derived this. We also proved
at axial 4 0 at axial load 0 we showed that I write it here r i at phi i 0 axial load 0 is 4 and c i
at phi i 0 is 0.5 is it not we use l hospitals rule then we derived this.

(Refer Slide Time: 14:34)

Then subsequently friends if we recollect, if we recollect, we rotation at the kth end and
derived the coefficients simultaneously.

And we found out that for this case kpp was found to be ri times of EI by Li kqp was found to
be c i times of kpp which is ci times of ri of E I by Li. And we said krp is ri (1 + ci) EI by Li
square and ksp which is here was -krp. We derived this is it not then we gave unit rotation to
the kth end derive the stiffness matrix in the same.

Then we said kpq which is here is, now ci ri EI by Li whereas, kqq is ri E I by Li and krq is 1 +
ci of ri E I by Li square and ksq is - of krq we got this. Then we also did for tensile load we
also did for tensile load, and we expressed beta as a function of α.

547
(Refer Slide Time: 17:03)

And we worked out we should say axial tensile load we got r i as αi of α i coshαi - sinhαi by
twice of 1 - coshαi + αi sinhαi. And ci was found to be αi - sinhαi by sinhαi - αi coshα.

(Refer Slide Time: 18:20)

Then we started working on translational coefficients, so we worked on translational

coefficients functions let us say under axial compressive load. So, we gave unit displacement
at r we also did same way in s as well and these forces generated kpr kqr krr and ksr were
determined for this free body diagram is it not. We found that kpr is same as kqr which is
same as ri into 1 + ci of EI by Li2.

548
And krr is 2 r i 1 + c i of E I by L i q and ksr is - of kr this is what we derived am I right. We
wanted to express this as a function of translational constant, so we said also let krr be
expressed as t i times of 2 ri 1 + ci E I by Liq where ti is called as translation function we also
say this is - of ksr.

(Refer Slide Time: 20:35)

So, we should write this way ksr is - of kr we follow the same logic here. Where t i is a
translation function which was given by t i is equal to 1 - pi square phi i 2 r i 1 + c i.
Remember phi is a multiplier or the ratio between the axial load and the Euler’s buckling load
that is a ratio, I think we already know this equation how phi is connecting p and p e because
we derived this.

After deriving this we summarize these constants ok, we summarize them and we wrote it in
a matrix form which is quite convenient and this resembles the same I am writing the matrix
form here.

549
(Refer Slide Time: 21:39)

The stiffness matrix for stability function k E I times ofi i r i by L i c i r i by L i let us.

Then I am using the same logic see here how we wrote this sum of these two by l was this
and this is negative of this is it not that is what we did I am using the same algorithm here.
Some of these two by l that is 1 + c i of r i by L i square this is - of r i by L i square 1 + c i am
using the same logic. If you look at this is crossed, is it not see here this is crossed.

So, I will cross it here I will cross this here. So, I am writing this as c i r i by L i this is simply
r i by L i. In fact, that is what you will get here see here, kpp kqp this is the first column the
second column will be kqq is r i E I by L i, I am doing the same thing this is p this is q r and
s.

So, I am looking for let us say kpq I am looking for kpqkpq let us see here kpq is c i r i E I by
L i c i r i E I by L i. So, I am doing the same thing I am just so, I am making you to easily
remember this equation that is what I am trying to do. Similar to what we wrote in the beam
analysis. So, once we do this some of these two will be here by l.

So, that is 1 + c i of r i by L i square then - of this r i by L i square 1 + c. Now if you look at

the third value here sum of these two by L i. So, r i times of 1 + c i by L i square sum of these
two by L i.

So, that is r i 1 + c i times of by L i square this is what we did in the beam also is it not see
here. Sum of these two by L is this, sum of these two by L is this ok, sum of these two by L is

550
this same way here. Sum of these two by L will be this which will be 2 times of r i 1 + c i by
L i cube.

Since it is a translation function, I put t i here let me do it in different colour I am putting t i

here. So, therefore, this will be - two t i of r i of 1 + c i L i cube. The fourth column as you
see here is negative of the third column see here, I am doing the same thing here negative of
the third column. So, - of r i 1 + c i by L i square - of r i 1 + c i by L i square - of 2 t i r i 1 + c
i by L i cube + of 2 t i r i 1 + c i by L i cube.

Friends it is very easy to remember this matrix if you remember this matrix. So, I have
conveniently mapped for your convenience of at least remembering a standard beam
derivation which is a fixed beam, to a stability function which is much more complicated in
derivation see. The derivation is complicated, but you can remember this equation very well
without any mistake.

And whereas, this r i and c i values are already given we already derived them you can
remember them right.

(Refer Slide Time: 27:29)

Now having said this, now we have established a perfect correlation between a standard fixed
beam used for static indeterminate analysis with that of a problem with axial load to derive
stability functions there is a perfect correlation is it not. So, this has got to be clearly

551
understood and one can also plot this r i c i and t i values for different phi this is called
stability chart which I showed you in the last lecture.

So, we have a program available in the textbook written by me which is referred in this
course it has got a MATLAB program the MATLAB program will help you to generate this
chart in excel sheet. Can also plot which I am going to use it for solving the problems in this
coming lectures. Now the question comes friends people like to know, what is the difference
between buckling and stability.

(Refer Slide Time: 29:14)

Friends, instability can be assessed based on the following 1 material property, 2 geometric
conditions, 3 nature and magnitude of the loads acting on the structure the standard thing.

552
(Refer Slide Time: 30:25)

Therefore, friends having said this assessment based on material strength highlights whether
the applied load exceeds the acceptable strength level of the material this classifies, or this
assessment classifies whether the structure is stable or not.

(Refer Slide Time: 31:40)

Now, assessment based on geometric configuration explores the functional success of the
member, it explores a success even under that is a very important condition here friends even
under the displaced configuration this is what we say as dynamic stability.

553
(Refer Slide Time: 32:46)

Therefore, friends in general in general loss of stability is assumed or is assessed in terms of

load deformation relationship. So, loss of stability arising either from change in geometry or
change in structural configuration, lead to further classifications of instability what are they.

(Refer Slide Time: 34:12)

They are; flexural buckling, torsional buckling, flexural-torsional buckling, lateral-torsional

buckling. So, friends we are slowly connecting stability to buckling. So, very clearly loss of
stability will lead to classifications of buckling.

554
Instability leads to classification of buckling that is what how they are connected to.

(Refer Slide Time: 35:15)

Furthermore, it is important to note that structural system remains in rest before instability
occurs also remains in rest immediately after the failure is initiated by buckling. So,
instability will initiate a failure and buckling is a failure instability is a geometric
displacement, it is a geometric imbalance.

The consequence of instability is buckling. There are four methods to solve buckling
problems; 1 equilibrium state approach, 2 work approach, 3 energy approach, 4 dynamic
approach. So, we have connected bridged the understanding between stability and buckling is
it not.

555
(Refer Slide Time: 37:33)

So, let us see what we learned from this lecture.

We have learnt how a standard fixed beam derivation is connected to stability function
derivation. We have remembered the matrix with respect to the stability function and we
realize that it is very easy, and it is having an one to one mapping. We have also learnt the
difference between stability and buckling and we realize that buckling is a failure mode is
consequence of instability.

So, in the next lecture we will take up some problems use MATLAB to find out or to check
stability of a given structural system we will do some examples.

Thank you very much and have a good day bye.

556
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 33
Critical buckling load - Numerical examples

(Refer Slide Time: 00:23)

Friends, welcome to the 33rd lecture of the course Advanced Steel Design. In this lecture we
are going to do some numerical examples to determine Critical Buckling load we are going to
use MATLAB programs for doing this.

(Refer Slide Time: 01:12)

557
So, let us first see the basis of doing this analysis. Friends, one of the basic assumptions
which we made to obtain the buckling load of a structural system is the deformation is
sufficiently small compared to its initial condition that is it undergoes small deformation. And
hence critical buckling load can be estimated using linear theory. Furthermore, friends it is
also important that buckling loads are estimated under the assumption that structures only
transfer axial forces.

In case they encounter transverse loads then these loads will cause additional moments which
will alter the stiffness significantly. Therefore, in such situation one must use continuous P-M
interaction curve to estimate the buckling load we must remember this. So, we are under the
assumption the structure will undergo only or will encounter only axial forces.

(Refer Slide Time: 04:33)

558
Now, let us see what the basis fundamental logic is to obtain the buckling load. Friends, we
already said we are taking a basic module we are neglecting the axial deformations.
Therefore, joint loads related to the unrestrained joint displacements are set to 0. You may be
wondering how to identify an unrestrained joint displacement we will give some examples.

Let us say I take here three span continuous beam like this let me mark the unrestrained
displacements for all the type of supports first. So, if it is a roller support, the roller can
undergo rotation, but roller cannot undergo vertical displacement. So, I will make a table
unrestrained and restrained let me call this as 𝜃 and this as 𝛿v. So, I showed a table
unrestrained as 1 which is 𝜃 and restrained as 1 which is 𝛿 v.

If I have a hinged support this can undergo rotation, but again vertical displacement is
restrained. So, one rotation unrestrained one vertical displacement restrained. If you have a
fixed support rotation as well as restraints alright, so this 0 this is 2 1 is 𝜃1 is 𝛿ta. So, these
are unrestrained displacements of varieties of supports learning this we can now mark the
unrestrained displacements in green.

So, as 𝜃1, 𝜃2 and restrained degrees of freedom in red 𝜃3, 𝛿4, 𝛿5, 𝛿6, 𝜃7 𝛿8 by neglecting
axial deformation. I can say this beam has got unrestrained joint displacements as 2 which is
𝜃1 and 𝜃2 restrained displacements are 1, 2, 3, 4, 5, 6 which is 𝜃3 𝛿4 and so, on. So, what we
are now saying is unrestrained joint displacements are set to 0.

(Refer Slide Time: 08:48)

559
So, mathematically the joint load unrestrained should be set to 0. We know for a given matrix
k if we partition this matrix as unrestrained and restrained unrestrained and restrained. So,
this is k uu, this is k ur, this is k ru, this is k rr we also know the joint load can also be
partitioned as J Lu and J Lr u stands for unrestrained and the displacements can also be
partitioned unrestrained displacements and restrained displacements restrained displacements
anyway will be equal to 0.

We have an equation now force is stiffness multiplied by displacement is it not. So, we

already said this should be set to 0. So, look at this equation now.

(Refer Slide Time: 10:33)

560
This equation left hand side is 0 which implies the fact. Now this implies a fact that either 𝛿 u
must be 0 or kuu should be set to 0. Now friends you please see 𝛿 u cannot be 0 is under it
undergoes displacement is it not? Because unrestrained joint displacements are not zero. So,
the nontrivial solution could be determinant of this k u u matrix should be set to 0. The
equation 2 in stability analysis is called not stability I will say buckling analysis is called
characteristic determinant. Now when you expand the determinant to obtain the buckling
failure condition.

(Refer Slide Time: 12:37)

Friends, we should also remember if the axial load if the axial load is lesser than the critical
buckling load that is if axial load is lesser than critical buckling load then the joint
displacements of unrestrained joints will be 0. Because for buckling failure you must impose
this displacement hence determinant kuu will be positive which corresponds to a stable
condition.

On the other hand, if the applied load exceeds buckling load critical buckling load this will
refer to unstable condition because we already said buckling is a mode of consequence of
instability. So, in this case kuu determinant will be negative that is an important point which
we have to remember.

561
(Refer Slide Time: 14:45)

Furthermore friends, in a structural system comprising more than one member under axial
load then one need to establish a relationship between the axial load and the buckling load.
This relationship is referred as ϕi which we have used in the derivation. Therefore, the critical
buckling load can be given by the relationship P c r is ϕi of 𝜋square EI by l I square.

So, if you know ϕ if you know ϕi know E I know cross sectional moment of inertia I know
the length of the member I can find the critical buckling load and if P a is lesser than P c r
applied load we do not have to bother if P a exceeds P c r we can say it is failing by buckling
which is causing instability. So, with this background let us start doing a numerical example
of estimating the critical load we will see this example 1.

(Refer Slide Time: 17:22)

562
So, on the screen you see a column subjected to axial compressive load the degrees of
freedom are marked. So, unrestrained degrees are marked in green that is displacements.
Displacements include both rotation and translation remember that are marked in green color
and restrained displacements are marked in red color. So, now we can see very well here this
problem has got 2 unrestrained degrees displacements that is 𝜃 1 and 𝛿 2 and this has got 6
well 4 restrained degrees namely 𝜃3 𝛿4, 5 and 6.

Suppose if we neglect axial deformation then restrained degree will be 3 which will be 𝜃 3, 𝛿
4 and 𝛿 ϕ because this will also become 𝛿 ϕ this will also become 𝛿 ϕ I am neglecting axial
deformation. So, now we have chosen j th end at here which is marked as A and this is kth
end which is B at A and kth end at B. So, this is my x m and y m. Now I can write the
stiffness matrix for stability functions for this problem we can recollect that equation which
we wrote.

(Refer Slide Time: 19:57)

563
So, the stability function which will be EI times of let us mark the labels of this problem. So,
rotation at j. So, label is going to be rotation at j now rotation at k then translation along
positive m positive y at j translation along positive y at k. So, 1, 3, 2, 4 are the labels. How do
you do that? I repeat again rotation at j th end rotation at k th end translation at j th end along
positive y translation at k th end along positive y that is standard know see look at the fixed
beam this is x m this is y m 𝜃 p 𝜃 q 𝛿 r and 𝛿 s is it not.

So, p corresponds to rotation at j th end q corresponds rotation at k th end r corresponds to

translation at j th end along positive y and s is at the k th end, so same logic here. So, let me
write down this matrix. We already know this is r by l this is C r by l this is 1 plus c of r by l
is minus of this in the next column I just interchange this. So, this is r by l this is c by l and
sum of these two this is r c by l.

So, 1 plus c r by l minus 1 plus c r by l and the third column sum of these two by l. So, 1 plus
c r by l sum of these two 1 plus c r by l then sum of these two by l, but I have to use my l
square right this is l square by l square again. So, this is l square this is l square this is l square
because it is sum of these two by l is it not? Again, this is l square sum of these two by l. So,
now, this value k 3 3 will be sum of these 2 read by l. So, let me write that.

So, twice of 1 plus c r by l cube and I also say t translation function is it not. So, minus 2 r t 1
plus c by l cube. Fourth column is minus of third. So, minus of 1 plus c r by l square minus of
1 plus c r by l square minus of 2 r t 1 plus c l cube 2 r t 1 plus c l cube. So, this is what my

564
matrix is. Let me enter the labels we already marked the labels 1 3 2 4. So, 1 3 2 4now friends
please note this problem has two unrestrained degrees is it not.

So, therefore, I must partition this matrix at 2 by 2 to get my kuu. So, let me partition this
matrix at 2 by 2 and this becomes my k uu is it not. So, I can now write kuu as r by l, c r by l
sorry 1 2 sorry. So, not partitioning I think we can pick up the first and second. So, I must
pick up k 1 1 k 1 2 and then k 2 1 and k 2. So, let me write kuu as k11 k12 1 plus c r by l
square then k 2 1 1 plus c r by l square then k 2 2 which is 2 times of r t 1 plus c by l cube am
I right. So, I have kuu now. So, let me copy this and put it here. I have kuu.

(Refer Slide Time: 26:13)

Let me that is kuu matrix as we just now discussed I must set determinant of this to 0. So,
setting determinant of kuu to 0. So, if I do that. So, it is going to be r by l into 2 r t 1 plus c by
l cube minus 1 plus c r by l square the whole square is it not? Should be set to 0. Simplifying
you can simplify this very easily because you know l 4 l 4 here1 plus c into r 1 plus c into r
here in the left-hand side. So, if I say r by l square of 2 r t 1 plus c is equal to 1 plus c into r 1
plus c into r by l 4l 4 goes away 1 plus c goes away r goes away and one more r goes away.

So, I can now say 2 t is equal to 1 plus c or 2 t minus c will be 1 this is my control equation.
These are all stability functions t c and r are stability functions friends which we derived in
the previous lectures under axial compression axial tension we have these values. So, the
equation 1 on the screen is called characteristic equation which is 2 t minus c is 1 is called
characteristic equation.

565
(Refer Slide Time: 29:11)

So, now, for different values of ϕi we have the stability charts is it not which indicates r, c and
t values. So, what I should do I should go to the table choose ϕi value which satisfies the
characteristic equation. I will show you typically how the stability chart looks like give me a
minute.

(Refer Slide Time: 30:23)

Friends, I will just enlarge this it is a typical stability chart friends for different values of
ϕminus indicates tension and positive indicates compression.

566
(Refer Slide Time: 30:38)

(Refer Slide Time: 30:42)

It goes to 0 the 0 value is here then it goes positive.

(Refer Slide Time: 30:47)

567
So, positive compression. So, it is varying from minus 10 ϕi value to plus 4we can see that
plus 4.

(Refer Slide Time: 30:58)

(Refer Slide Time: 31:05)

568
(Refer Slide Time: 31:07)

So, the first column refers to r t and c. So, these are the values. So, what I should do for this
control equation; for this control equation to be satisfied I must choose a ϕ value it is very
difficult know it is very difficult to undergo and scan all entire table to satisfy this equation.
So, to solve this problem we use the MATLAB program. So, I will show you. So, we are now
using a MATLAB program, the program is available in the textbook referred in this course
you can download the program easily MATLAB support is also extended for NPTEL students
through IIT Madras link.

569
You can also undo a free tutorial student MATLAB program is given to you and it does not
require a sophisticated laptop or an desktop to run a MATLAB software. A normal laptop
should be able to run the problem what I am discussing about MATLAB has got many
facilities which requires a high-end computational facility. However, the problem what I am
discussing now can be run in an ordinary conventional laptop do not worry about that. So,
now, I will show you the MATLAB program.

(Refer Slide Time: 32:35)

So, friends what you see on the screen is the MATLAB program.

(Refer Slide Time: 32:45)

570
So, this is a program written for compression you can say from 0.01 to 4 we saw the value
just now on the table and for tension it plots up to 10, tension is minus. There is minus right
and then for 0 also we have plotted.

(Refer Slide Time: 33:08)

It is a program it gives you a stability functions r c and t this is an equation what we already
have derived you know this. So, we can easily, and chart is also prepared.

(Refer Slide Time: 33:23)

571
Then further friends we have a control equation which is the characteristic equation for this
problem 2t minus c is 1. So, a program is written to equate and find a specific ϕ value from
the output file which is generated from here.

(Refer Slide Time: 33:57)

Satisfy this equation. So, this program is now run. So, it gives me the ϕ value please see on
the screen 0.25, it also shows me the chart for your learning this is my r value variation, and
this is my c variation with respect to ϕ and this is variation of t with respect to ϕit is a chart.

(Refer Slide Time: 34:13)

572
So, it gives me after searching in the table or the chart gives me a ϕ value as 0.25. So, now, I
got ϕ value as 0.25 from the MATLAB. So, I will go to the screen now. So, I have got ϕ value
as 0.25 now how do we get the buckling load that is the question is it not. So, now, for ϕ
value of 0.25, from the chart or table you can refer r will be 3.6598, t will be 0.7854 and c is
0.5708 which satisfies the characteristic equation.

So, now, what I do, I substitute P c r as ϕ i of 𝜋 square E I by L i square where ϕ is 0.25 if you
know the material constant if you know the second moment of area or moment of inertia of
the cross section if I know the length of the member, I can find the Euler’s critical load. So,
the problem is very simple stability functions help you to obtain the Euler’s critical load
which is going to help you to estimate the buckling load.

So, program is given, derivations are clear, application is being done and I think it is very
easy for us to practice a greater number of problems is it not let us do one more problem.

(Refer Slide Time: 36:23)

So, example 2we will do one more problem. So, this is an example 2 where one member is
having 2 I other member has moment of inertia as I A B and C are the nodes. As the length of
the member is same, I can interpret this angle has 45 degrees right. Now I can also find at
node B; at node B if I have a load P applied here, I can find the component of this load which
I am going to call as P B C and the component of this load which I am going to call as P AB
or P B A as 0.707 P.

573
Now the angle is 45 degree right this angle is 90 and this angle is 45. So, now, I can say the
axial load on the member AB is 0.707 P compressive the axial load on the member BC is
0.707 P compressive. Now, there are two members sharing this load P. So, we must carefully
form the characteristic equation now let us mark the unrestrained degrees one can see here on
the screen 𝜃 1 and 𝜃 2 are the rotations which are unrestrained remaining all are restrained
and look at 𝛿 4 same both places indicating we are neglecting axial deformation.

Similarly, 𝛿 ϕ and 𝛿 ϕ neglect axial deformation along the member BC. So, I can say this
member has unrestrained degrees 2 which are 𝜃 1 and 𝜃 2 restrained degrees as 1 2 3 1 3 4 5
6 and 7. So, 5 which are 𝛿 3, 𝛿 4, 𝛿 5, 𝜃 6 and 𝛿 7. So, the total kinematic degree of freedom
is 7 neglecting axial deformation restrained is 5 and unrestrained is 2as usual I must now
derive the stiffness matrix using rotational stability functions for the member AB and for the
member BC.

(Refer Slide Time: 39:44)

So, for the member AB I will make a table for the member j and k and l I and labels. Let us
say for the member AB I am going to say the jth end is at B and kth end is at A. So, that is
here. So, this is my x m and this is my y m for the member AB length of the member AB is l
and moment of inertia is 2 I is it not let me mark the labels for the member AB. So, rotation
at j rotation at k displacement along positive y at j and at k is it not. So, rotation is 1 rotation
is then 2 then 5 and 3. So, the label should be strictly speaking 1, 2, 5 and 3. So, let us say the
labels are 1, 2, 5 and 3.

574
Now, for the member BC for the member BC we will take the j th end at B and this at C. So,
the j th end is here. So, this becomes my x m and y m for the member BC the labels are going
to be 1, 6, 4 and 7 is it not. So, labels are 1, 64 and 7 right the length of the member is again l,
but the moment of inertia is only I. So, I do not think there is any confusion for you people to
write this table. Once this table is done let us now write the matrix for the member AB and
the matrix for the member BC4 by 4.

(Refer Slide Time: 41:54)

So, E into 2 I for the member AB right. So, I should say it is r 1 by l, c 1 r 1 by l, 1 plus c 1 of
r 1 by l square minus 1 plus c 1 r 1 by l square. So, then I cross this is c 1 r 1 by l, this is r 1
by l, this is 1 plus c 1 of r 1 by l, 1 plus c 1 of r 1 by l. Then we add these two and divide by l.
So, 1 plus c 1 r 1 by l square, 1 plus c 1 r 1 by l square that is adding these two by l then
adding these two by l, but t along with that.

So, that is going to be 2 t 1 r 1 1 plus c 1 by l cube right minus 2 t 1 r 1, 1 plus c 1 by l cube

the fourth column is negative of 3. So, minus of 1 plus c 1 r 1 by l square minus of 1 plus c 1
r 1 by l square minus of 2 t 1 r 1 1 plus c 1 by l cube plus of 2 t 1 r 1 1 plus c 1 by l cube
right. Let us mark the labels for the member AB is 1, 2, 5, 31, 2, 5 and 3 similarly here 1, 2, 5
and 3.

575
(Refer Slide Time: 44:35)

Then for the member BC let me do it to the next page. k BC r 2 by l 2 l that is the length of
the member c 2 r 2 by l 1 plus c 2 r 2 by l minus 1 plus c then it is crossed, so c 2 r 2 by l r 2
by l 1 plus c 2 r 2 by l minus of. Then sum of these two by l this is l square right there will be
l square here the previous case also l square that is right. It l square here then sum of these
two by l then sum of these two by l again. So, 2 t 2 r 2 1 plus c 2 by l cube minus of 2 t 2 r 2 1
plus c 2 by l cube the fourth column is minus of this. We also know the labels of this is 1, 6,
4, 7. So, let us enter the labels 1, 6, 4, 7.

(Refer Slide Time: 46:57)

576
Now, friends, let us form k uu. Let us form k uu how many unrestrained degrees are there?
We have got two unrestrained degrees in this problem see here 1 and 2. So, I must get E I
common 1 and 2. So, I must find k 1 1. Let us look at this k 1 1 is this I am just indicating
with maybe the pink color k 1 1 is this from here. So, k 1 1plus the other one is k 1. So, now,
please note there is a two multiplier here. So, 2 r 1 by l 1 2 r 1 by l l 1 and is same. So, then
for the second member it is r 2 by l then let us go to 1 2 from here 1 2 is this value.

So, 2 c 1 r 1 by l right 2 c 1 r 1 by l why this 2 has come? This 2 has come because of this 2
has come because of this. Now, I must get 2 1 go here 2 1 is this value. So, 2 c 1 r 1 by l
correct then let us go to 2 2. 2 2 is this value we also have 2 2 here, we also have 2 2 here no
we do not have it only this. So, 2 2s r 1 by l there is a 2 here. So, I should say 2 r 1 by l
friends we have got kuu matrix now.

Now, I must set this determinant of k uu to 0. So, this diagonal multiply minus this diagonal
and set it to 0. So, I should say 2 r 1 plus r 2 by l into 2 r 1 by l minus 2 c 1 r 1 by l the whole
square should be 0 am I right.

(Refer Slide Time: 50:14)

This is what I must get set it to 0 when I set this to 0 and expand and simplify we get the
characteristic equation as 2 r 1 of 1 minus c 1 square plus r 2 is u this my characteristic
equation. Further, we know P c r 1 is P c r 2 which is P c r of 0.707 is it not? Well, that is the
load which is being applied therefore, friends we have 1 more relationship ϕ 1 will be P times
of 0.707 𝜋 square E to I by l square ϕ 2 will be P of 0.707 by 𝜋 square E I by l square.

577
So, comparing these two equations we can now also say 2 ϕ 1 is ϕ 2 you can see easily from
here2 ϕ 1 will be 5 2 right. So, I have to satisfy two equations while selecting ϕ. So, I must
find ϕ 1 and ϕ 2 because ϕ 1 and ϕ 2 will be helpful to get P and the characteristics equation
should be satisfied. So, I can refer to the chart or the table and try to get.

(Refer Slide Time: 52:19)

So, I am going to use MATLAB again, so using MATLAB.

(Refer Slide Time: 52:25)

So, again the same plot, but I am going to show you the control equation.

578
(Refer Slide Time: 52:32)

(Refer Slide Time: 52:39)

The control equation what we had is 2 r 1 1 minus e 1 square plus r 2 is 0 is it not this is the
control equation here.

579
(Refer Slide Time: 53:07)

So, we have developed this control equation apply this and wrote an algorithm and then we
run this program. So, forget about these plots, we get ϕ 1 as 1.01 and ϕ 2 as 2.02. So, I will
write it here I will close this minimize this, these are charts being plotted I will minimize this.

(Refer Slide Time: 53:31)

So, now I can write ϕ 1 is 1.01 and ϕ 2 is 2.02 for these values we got r 1 as 2.4493, t 1 as
0.0124 and c 1 as 1.0101 and r 2 as 0.1120, t 2 as 1.7148 and c 2 as 31.6264. So, now I have
got p critical is ϕ 1 times of 𝜋 square E of 2 I by 0.707 of l square. So, ϕ 1 I already have,
which is 1.01. So, if you know E if you know I if you know l I can find.

580
(Refer Slide Time: 54:55)

So, friends in this lecture we learnt the difference between buckling and stability analysis. We
also learnt how to set the characteristic equation to solve a buckling problem for a frame
subject to axial loads. So, we said that set the k uu determinant to 0 for non-trivial solution
and try to get the control equation and then choose the ϕ values from the chart or the table
and try to find the ϕ.

Once I know the ϕ value I can find p critical. So, I wish you should do more examples to
solve more problems and try to learn more about this. So, friends stability analysis is not
difficult it is easy please practice more problems and use the MATLAB program given.

581
(Refer Slide Time: 56:16)

And also this book will be very helpful for you which is being used as a reference. So, please
refer to this book Design Aids for Offshore Structures Topside Platforms under Special Loads
which has got the stability charts and the table and also the MATLAB program. This is a
CRC press release this book is available in public domain please access to this book through
your institute library subscribe this to a library recommend this to your resource scholars and
friends and faculty colleagues and I hope you will find an extremely useful contributions
from this book for your additional learning on this course.

Thank you very much and have a good day bye.

582
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 34
Stability problems - numerical examples

(Refer Slide Time: 00:20)

Friends welcome to the lecture 34 of the course Advanced Steel Design. In this lecture we are
going to learn more about the stability problems numerical example. We will deal with more
numerical examples in this lecture and learn couple of more with application to MATLAB
programs.

583
(Refer Slide Time: 01:21)

Before that let us revise quickly, that my stiffness matrix of the member is given by; the
stiffness matrix is given by the basic equation which we should remember like that of a beam,
which is a 4 by 4 matrix, which is ri by li, this is ci ri by li and these two by li will be this.

So, 1 + ci of ri by li square and negative of this. The 2nd column is swapping of this and sum
of these two by li and the last one is negative of the previous one. The 3rd column is sum of
these two by li is also sum of these two by li and the 4th one is the next one is sum of these
two by li again. So, I should say 1 + ci ri twice, but introduce a new coefficient ti ok and this
is minus of that. The 4th one of course, is negative of the 3rd one this is what we should,
remember which is similar to what we have in the beam implication problem.

584
(Refer Slide Time: 04:02)

Having said this let us go to the example 3, as you see on the screen here. So, this example is
about an orthogonal frame has got two members. One member is having lengths 2L the other
one is L. We have marked the degrees of freedom as you see here.

So, the degrees of freedom are marked as theta L and theta two unrestrained ok and
remaining all are restrained, we are neglecting axial deformation. So, restrained degrees, so it
has got unrestrained degrees 2, which are theta 1 and theta 2 and restrained degrees are let us
say theta 3, del 4, del 5, del 6 and del 7 which is about 5. So, the total kinematic degrees of
freedom is 7. So, each matrix will be 4 by 4 and we can try to find out this very easily.

So, now for the member k a b or k 1 ok, for the member k 1 I should say r 1 ok by 2l, 1 + c 1
c 1 r 1 by 2l. So, 1 + c 1 of r 1 by 2l square sorry by li square. So, that is 2l the whole square
ok so, the negative of this 1 + c 1 r 1 by 2l the whole square. The 2nd column will be c 1 r 1
by l, r 1 by l, 1 + c 1 r 1 by 2l the whole square - of 1 + c 1 r 1 by 2l the whole square with a
negative sign.

The 3rd column will be 1 + c 1 r 1 by 2l or 2l the whole square. So, 1 + c 1 r 1 by 2l the

whole square the 3rd one is 2 r 1 t 1 1 + c 1 by l cube. So, I should say 2l the whole cube and
negative of this. The 4th column of course, negative of the 3rd column. So, we have this, let
us mark the labels. We will take the j th end here and k th end here for this member. So, we
should say this is 1, 3, 7 and 5; 1, 3, 7 and 5.

585
So, rotation at j th end, rotation at k th here, translation along positive y if I say this is my x
m, anti-clockwise 90 is my y m. So, this becomes my degree of 7 and this is my label 5. So, k
1 for the member ab is established of course, we have an EI multiplier outside. So, we can do
this. The next one is for the 2nd member k 2.

(Refer Slide Time: 08:50)

So, let us do it for the member k 2 which is again EI times of again a 4 by 4 matrix. So, this is
going to be r 2 by l, because you see the length of the member 2 is only l. The length of the
member 2 is only l. So, r 2 by l, then c 2 r 2 by l, 1 + c 2 of r 2 by l square.

So, let us mark the labels, look at this figure I am calling this as my j th end and this as my k
th end, this becomes my x m and this becomes my y m for the member 2. So, I should say the
labels are going to be 1, 2, 4 and 6. Let us mark them 1, 2, 4 and 6. Now, we can assemble
these two matrices and try to get k u u, the k u u will be partitioned at 2 by 2.

586
(Refer Slide Time: 11:34)

So, now I am getting try to get k u u because I have set determinant to 0 of this. So, that is k u
u which is 2 by 2. Let us do this to be 2 by 2. Let us try to get this. So, we should look into
this figure 1 1 is I have here, so r 1 by 2 l. So, let we take, r 1 by 2l, then do we have 1 1 here.
So, we have r 2 by l as well.

This is 1 one ok let us say 1 2 which is not in the 1st matrix, but 1 2 can be in the 2nd matrix.
So, 1 2 is c 2 r 2 by l, c 2 r 2 by l. Then let us say 2 1, 2 1 is here, 2 1 is the 2nd matrix here c
2 r 2 by l. So, c 2 r 2 by l and 2 2 is here which is r 2 by l ok right. So, I want to set this
determinant to 0 to get my characteristic equation. So, if I do that I am writing it here. So, r 1
+ 2 r 2 by 2l square - c 2 r 2 by l the whole square should be 0.

So, this gives me the expression r 1 + 2 r 2 by 2l square into r 2 ok, into r 2 because we have
to product this r 1 by r 1 + 2 r 2 by 2l square into r 2 - this is set to 0, which gives me the
characteristic equation as r 1 equals 2 r 2 times of C 2 square - 1, this is what we call as my
characteristic equation, ok right. Now, I want to also see what my axial load is coming on the
section.

587
(Refer Slide Time: 14:51)

So, if I say P critical of the 1st member ok look at this figure. For the 1st member this is 2P,
for the 2nd member this is P, is it not. So, 1st member this is 2P and for the 2nd member ok,
let us I think use square brackets ok which will be P. So, by this logic ϕ1 should be 2P,
because ϕ is the ratio between the applied load and the Euler’s critical load which is pi square
EI by l square.

So, in this case it is 2l the whole square ok, which will become 8P by pi square EI by l
square, am I right? Similarly, for ϕ2 which will be P by pi square EI by l square. So, by this
logic comparing equation 1 a and 1 b we can get the following relationship, which will beϕ1
is 8 times of ϕ2: equation 2. So, we got relationship we got k uu determinant as well.

588
(Refer Slide Time: 16:38)

(Refer Slide Time: 16:52)

So, I am going to now satisfy these two conditions. What are the conditions? So, I have to
satisfy the characteristic equation which is given by r 1 is equal to 2 r 2 c 2 square - 1 the
other condition is I must satisfyϕ1 is 8ϕ2, is what I have to satisfy. So, let us run the
MATLAB program to do this. Let us now run the MATLAB program and try to obtain the
values ok, let us do that.

589
(Refer Slide Time: 17:34)

(Refer Slide Time: 17:51)

590
(Refer Slide Time: 18:00)

So, friends this is the program which you are going to run. You can see here the control
equation is this r1 square ok r, let us say r1 is equal to or r 1 - 2 r2 times of c2 square - 1
should be set to close to 0. We simply say exactly 0, we may not be able to pick up the correct
value of r1 r2 and c2 and c1 from the chart so, closely.

(Refer Slide Time: 18:34)

So, let us closely pick up this and let us run this program. So, I get these charts, I think we
already have this charts with us, we know right. So, we know these charts of r i and t. So, the

591
answers areϕ1 is 2.96 andϕ2 is 0.37. So, I will go back here. So, I getϕ1 asϕ1 as 2.96 andϕ2
as 0.37.

So, for these values if you look at the chart, I get r 1 as - 4.6727, I get t 1 as - 5.7540 and I get
c 1 as - 1.4628 and I get further r 2 as 3.4878 from the chart, t 2 as 0.6754 and c 2 as 0.6127.
So, now, I can interpret P critical asϕ1 times of pi square EI by 8 l square ok, 8 l square. So,
we will be able to get the critical load because I knowϕ1. If I know if I know EI and length, I
can get critical value (Refer Time: 20:29) critical ok, it is very simple and straightforward.

(Refer Slide Time: 20:35)

(Refer Slide Time: 20:47)

592
Let us go to example 4; we will go to example 4. So, let us draw this figure here, one end is
fixed other end is hinged and this is also fixed. Let us call this as A, B, C and D, this is of two
length l, this is l and this is l and this is member 1 and this is member 2 and this is member 3.
This is subjected to in horizontal load of 2P here of the whole problem.

Let us know mark the unrestrained degrees of freedom. So, I get rotation here theta 1, I am
marking them in green rotation here theta 2, these are my two unrestrained degrees of
freedom. Let us mark restrained degrees of freedom theta 3 del 4 if I say del 4 and I neglect
axial deformation, I have del 4 here and I have del 4 here also. And this is del 5 ok, this is del
6 therefore, this is also del 6 because I am neglecting axial deformation ok and this is del 7
and this is theta 8 and this is del 9.

So, I have 9 degrees of freedom, the unrestrained degrees of freedom are 2, which is theta 1
and theta 2. The restrained degrees of freedom r theta 3, del 4, 5, 6, 7, theta 8 and then 9, so
which is 7. So, the total kinematic degrees of freedom are 9 and k uu will be of size 2 by 2,
because there are 2 unrestrained degrees and remember we are anyway neglecting the axial
deformation.

So, I can easily write k AB, I can write k BC, I can write k CD ok, assemble them and k k u
that is what I am going to do now. So, let us do that exercise, very clearly it is simple and we
can do that.

So, let us look at the labels for the member AB, let us look at what are the labels, we will take
the j th end here ok we will take the j th end here and mark it; j th end here and k th end here.
This is my j th end and this is my k th end. So, let us mark a table quickly, j th end is at B, and
k th end is at A and the labels are rotation at j rotation at k deformation or displacement along
positive y at j and positive y at k ok, let us do it for the member BC.

So, for the member BC, we will take the j th end here. So, j th end here and k th end here. So,
I should say the j th end is at C and k th end is at B and the labels are rotation at the j th end
rotation at the k th end displacement along positive y at the j th end and at the k th end. Let us
do for the member say BD ok, for the member BD. So, for the member BD which is
horizontal the length is of course, 2l we should remember that and we will take the j th end
here and the k th end here for this member.

593
So, the j th end is at B, k th end is at D therefore, the labels are 1 rotation, then 8, then 4 and
9. These are the labels; I do not think we have any difficulty in writing the stiffness matrices.
For all the 3 members let us quickly do that.

(Refer Slide Time: 25:42)

So, I should say k AB is EI times of, similarly k BC is EI times of, similarly k BD EI times of
remember this is of 2 length l. So, be careful about the length. We also know the labels let us
write the labels 1, 3, 6, 5, that is for the member AB, for the member BC, BC 2, 1, 7, 6; 2, 1,
7, 6 and for the member BD we know the labels are 1, 8, 4, 9, is it not; 1, 8, 4, 9.

And we know the 1st value here is r 1 by l, this is c 1 r 1 by l, 1 + c 1 of r 1 by l and - 1. This

is again c 1 r 1 by l, r 1 by l, 1 + c 1 of r 1 by l - 1 + c 1 of r 1 by l. 1 + c 1 r 1 by l, 1 + c 1 r 1
by l, 2 times of t 1 r 1 1 + c 1 by l ok, these are all l square friends, I am sorry. This is l cube -
2 t 1 r 1 1 + c 1 by l cube and the 4th column is of course negative of the 3rd column.

Let us go to member BC. So, it is going to be r 2 by l, c 2 r 2 by l, 1 + c 2 r 2 by l square; so,

c 2 r 2 by l r 2 by l 1 + c 2 r 2 by l square. Similarly, 1 + c 2 r 2 by l square, 2 times of t 2 r 2
1 + c 2 by l cube and - of this the 4th one of course, is negative of the 3rd column, I am not
entering it ok it is very simple to remember.

Let us go to the member k 3. So, it is going to be, I will do it here r 3 by 2l, 1 c 3 r 3 by 2 l.

So, 1 + r 3 of sorry, 1 + c 3 r 3 by 2l the whole square and negative of this, c 3 r 3 by l, r 3 by
2 l. So, 1 + c 3 of r 3 by 2l the whole square negative of this. So, this will be 1 + c 3 r 3 by 2l,

594
1 + c 3 r 3 by 2l this is going to be 2 times of t 3 r 3 1 + c 3 by 2l the whole cube. This is
square is it not and this is negative of this. The last column of course, is negative of the 3rd
column we can know this. So, now, if you assemble k and partition a 2 2 this is going to be k
uu sub matrix.

So, now I should write k uu here, which is EI common of let me write this k BD somewhere
here. So, 1 2 and 1 2, I am writing k uu. So, I should say r 1 by l that is 1, then I get r 2 by l
then I also get r 3 by 2l ok, this one. Let us now do 1 2. So, 1 2 I can pick up from this which
is c 2 r 2 by l, then I should say 2 1 2 1 I should pick up from here which is c 2 r 2 by l, then I
should say 2 2, 2 2 I will pick up from here, which is r 2 by l, right. I should set this
determinant to 0. So, let us multiply this and - of this let us do that. So, I can write that which
will be, I will copy this matrix.

(Refer Slide Time: 32:45)

Let me put it here, let me rub this. So, I should say r 1 by l + r 2 by l + r 3 by 2l times of r 2
by l - c 2 r 2 by l the whole square should be set to 0, am I right. Which will be equal to 2 r 1
+ 2 r 2 + r 3 by 2l square of r 2 - c 2 square r 2 square by l square is 0, which will become 2 r
1 + 2 r 2 + r 3 times of r 2 - 2 c 2 square r 2 square should be 0, am I right. r 2 times r 2 r 1 +
2 r 2 + r 3 is equal to 2 r 2 square c 2 square.

So, let us say r 2 goes away ok, r 2 goes away. We can say 2 r 1 + 2 r 2 + r 3 is equal to 2 r 2 c
2 square. So, that is my characteristic equation, the control equation to develop this.
Furthermore, friends please see this figure, I should say the axial load on the member 1, the

595
axial load on the member 1 is 0, the axial load for the member 2 is also 0, the axial load for
member 3 is 2P compressive; am I right?

(Refer Slide Time: 35:30)

So, by this logic can I sayϕ1 andϕ2 will be 0? Because P1 and P2 are 0, am I right. So, let us
sayϕ3 yes 2P times of pi square EI by 2l the whole square, which will be 8P by pi square EI
by l square, let us keep this information with us to compute my critical load. So, I have got 2
control equations now, one is the characteristic equation, other is this relationship. I must now
select ϕ in such a manner from the chart to satisfy equation 1 and equation 2, this is equation
2 sorry and equation 1 is this.

596
(Refer Slide Time: 36:37)

(Refer Slide Time: 36:50)

Let us run the MATLAB program and get the answer. So, now, I am running the MATLAB
program. So, this is a program friends. So, the control equation is this.

597
(Refer Slide Time: 37:14)

This is what we have. So, we can try to run the program and we get ϕ as 3.51 ok that isϕ3 ok
3.5. So, we should sayϕ3 was found to be 3.51 and the corresponding r 1 r 2 sorry c 1 and t 1
r 2 c 2 and t 2 r 3 c 3 and t 3 can be found out from the chart, I am not writing it here. So, we
can do that.

So, if you have this value as 3.51. So, I can write P cr is equal toϕ3 of pi square EI by 8l
square. So, this is known, if I know the cross section, if I know the length and if I know the E
value I can find the Euler’s critical, ok friends.

(Refer Slide Time: 38:33)

598
Let us do one more problem, example ϕ you can now see a single way single storey frame on
the screen. So, I can quickly write some inferences, the unrestrained degrees of this problem
are 2, which are theta 1 and theta 2 and the restrained degrees of this problem are 3, 4, 5, 6, 7,
8 and 9.

So, 7 the total kinematic degrees of freedom is 9. So, k uu will be of size 2 by 2 because
unrestrained degree is only 2 and each k, that is k AB, let us say this is my A, this is B, C and
D. So, k AB will be of size 4 by 4, k BC will be of size 4 by 4 and k CD will be of size 4 by
4. Let us write down quickly the labels for this ok, let us quickly write down the labels. So,
let us make a table for the member AB, where is my j th end, where is my k th end ok and
what are the labels.

We will keep for the member A B we will keep this as B and this as A. So, the labels are
going to be quickly rotation at j, rotation at k, displacement at j and displacement at k. For the
member BC we will keep the j th end at B and this at C. So, the labels are going to be 1, 2, 4
and 7 and for the member CD, we will keep the j th end at C, and this at D. Therefore, the
labels are going to be 2, 6, 9 and 8. So, friends I think we can write easily the stiffness
matrices of all the members, let us do that quickly.

(Refer Slide Time: 41:10)

So, k AB will be E times of; so r 1 by l, 1 + c 1 sorry c 1 r 1 by l, then 1 + c 1 of r 1 by l - of

this. Similarly, c l square so, c 1 r 1 by l, r 1 by l, 1 + c 1 of r 1 by l square and - of this. So, 1
+ c 1 r 1 by l square, 1 + c 1 r 1 by l square, 2 times of t 1 r 1 1 + c 1 by l cube - of this, the

599
4th column is simply negative of the 3rd column and the labels are 1, 3, 9, 5; is it not 1, 3, 9,
5.

We can also write quickly for the member k BC, which is going to be E times of ok; r 2 by l,
c 2 r 2 by l, 1 + c 2 of r 2 by l, - of this l square ok c 2 r 2 by l, r 2 by l, 1 + c 2 of r 2 by l
square negative of this. So, 1 + c 2 r 2 by l square, 1 + c 2 r 2 by l square, 2 times of t 2 r 2 1
+ c 2 by l cube - of this. And the 4th column will be negative of the 3rd column the labels are
going to be 1, 2, 4 and 7 you can see here 1, 2, 4 and 7.

So, let us do it for k CD, which is E times of we should remember the length is 2l in this case
is it not it is 2 no sorry, it is 2L in this case. So, we have to be careful ok even this also 2l,
even this also 2 l. So, we should I think write 2l, 2l, 2l, the whole square ok right; k BC is just
only l that is ok and k CD we can enter this, which will be I will do k CD in the next slide.

(Refer Slide Time: 44:30)

So, k CD EI times of which will be r 3 by 2l, c 3 r 3 by 2l 1 + c 3 of r 3 by 2l the whole

square negative and this value will be 2 times of t 3 r 3 1 + c 3 by 2l the whole cube negative
so, negative + of this column. So, the labels are going to be 2, 6, 9, 8 you can see here 2, 6, 9.

600
(Refer Slide Time: 45:47)

So, we have all the 3 matrices, I can assemble them and get k uu which will be E times of ok
substituting we will get this as kuu. So, let us say for example, I should ok, should 1, 3, 9 and
5. So, I am looking for one more. So, I should get this and this. So, r 1 by 2l + r 2 by l so, r 1
by 2l + r 2 by l. 1, 2 and 1 and 2, let us do for 1 2.

So, 1 2 is here, c 2 r 2 by l, c 2 r 2 by l. So, let us say 2 1; 2 1 is here c 2 r 2 by l. So, c 2 r 2

by l let us go for 2 2, 2 2 are 2 places let us say r 2 by l ok r 2 by l + then we also have r 3
with 2l right. So, now, let us set this determinant to 0.

(Refer Slide Time: 47:23)

601
So, I can write that that is going to be r 1 + 2 r 2 by 2l see here, r 1 + 2 r 2 by 2l into 2 r 2 + r
3 by 2l should be equal to I can say should be equal to c 2 r 2 by l the whole square right, c 2
r 2 by l the whole square. So, this gives me rise to the characteristic equation which will be r
1 times of 2 r 2 + r 3 + r 2 times of 4 r 2 + 2 r 3 should be equal to 4 r 2 square c 2 square,
that is my characteristic equation which I will use in MATLAB now and do it.

But further, let us also writeϕ1, see the figure, ϕ1 will be P, ϕ2 will be P whereas,ϕ3 will be
2P. So, let us do that. So, ϕ1 will be P by pi square EI by l square, which is 2l whole square,
which will now become 4P by pi square EI by l square. Let us sayϕ2 is P by pi square EI by l
square whereas, ϕ3 will be 2P by pi square EI by 2l the whole square which will become 8P
by pi square EI by l square.

(Refer Slide Time: 49:15)

So, now I can write a relationship whereϕ1 is 4ϕ2 andϕ3 is 8ϕ2 ok, let us run the MATLAB
program for the 5th problem.

602
(Refer Slide Time: 49:44)

So, this is the 5th problem, we have the program here, we have the control equation which is
here. Set to close to 0 0 ok, not 0, 00 let us run this program, I get these are the charts of r t
and p, I getϕ1, ϕ2,ϕ3 as you see on the screen. So, let us sayϕ2 is 0.38 andϕ1ϕ3 can be
obtained. I will minimize this, I can now sayϕ2 is 0.38, I can findϕ1 andϕ3 does not matter.
So, for this I get r 2 c 2 and t 2. I got r 2 as 3.4738 sorry 32, c 2 as point triple 60 and this is
point. So, this is t 2 and this is c 2 which is 0.6165. So, now, I can say P cr isϕ2 times of pi
square EI by l square, I haveϕ2. If you know E, I and l I can find P c.

(Refer Slide Time: 51:24)

603
So, friends in this lecture we also learned 3 additional numerical examples to find Euler’s
critical law. This book is a very good reference which you must use for stability functions
derivation understanding more problems and computer programs and MATLAB, you can
download them, and this book is available in open access.

So, please see the book and try to use it for your library, recommend it to a library and use it
thoroughly. The MATLAB programs are available in this book, you can copy paste them and
run them these are the typical charts what you will see when you run the program. So, friends
practice more examples in stability problems and if you have any difficulty, you can always
discuss it in the forum, we will try to help you out.

Thank you very much and have a good day bye.

604
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 35
Stability of shells

(Refer Slide Time: 00:17)

Friends, welcome to the 35th lecture on the course on Advanced Steel Design. In this course
we are going to learn more about checking the Stability of cylindrical shells ok or tubes let us
say ok tubular elements. Now the stiffened cylinders are one of the major structural
components of special kind of structures. They are quite common in offshore compliance
structures. So, we are now talking about stiffened cylinders.

They are common structural members in offshore structures. They can be classified as ring
stiffened stringer stiffened and ring stinger stiffened cylinders. So, the whole batch is also
called as orthogonally stiffened cylinders. They are also known as orthogonally stiffened
cylinders.

605
(Refer Slide Time: 02:37)

Now, let us quickly understand what stringers are. Stringers are longitudinal stiffeners which
are either attached externally or internally to the cylinder at equal spacing. The stiffness can
be of cross section there can be a flat bar they can be angles they can be T sections and so on.

They are generally integral welded to the shell, and they help in resisting the lateral loads.
Now the structure is fabricated from hot or cold form plates essentially the welding will be a
butt welding and therefore, we need to check whether the stiffened cylinders are stable for a
given configuration.

The geometric distortion and residual stresses are some of the common problems which you
heard in steel structures which commonly occur under welding process. So, we can say that
welding process initiates geometric distortion and residual stresses. So, upon the types of
failure orthogonally stiffened cylinders generally fail by buckling. This buckling can be in
different modes. They can be shell buckling, panel stiffener buckling, panel ring buckling and
column buckling.

606
(Refer Slide Time: 05:53)

Now, let us try to draw them and show how do they look like. Let us say this is my cylinder.
Now when there are different layers along the length and along the circumference some of
them will have solid concentrations of stresses some of them will have negative. I am
showing the positive ones as firm lines stress concentration factors and the negative ones as
dotted lines.

So, this form of failure is called shell buckling. They generally form a pair; they generally
occur alternatively they generally occur alternatively. The second one is called panel stiffener
buckling. Let us take a shell again cylinder. Now what happens in this case is the panels will
start buckling.

So, I should say and again there will be on different layers the compression tension. So, this
is panel stiffener buckling. Panel ring buckling is a different phenomenon, which occurs in
this form. Again, there will be alternate tension compression inundation and solid bulging.
So, this will be panel ring buckling.

607
(Refer Slide Time: 09:06)

Of course, general buckling looks like this. So, there will be tension and compression stress
concentrations or along the length of the member. So, this is called general buckling whereas,
a column buckling phenomena will be like what we have seen in a column. So, these are
different buckling modes of failure of orthogonally stiffened cylinders. We will do a problem
in relationship with an international code and try to understand how code classifies stability
ok of this circular cylinders.

(Refer Slide Time: 10:29)

608
So, we will do a simple exercise. We will say check the stability of a ring stiffened circular
cylinder under axial compressive stress of 30 newton per mm square against shell buckling.
We will just look only for one mode of failure. So, the data given is like this diameter of the
cylinder is 5 meters, thickness of the cylinder 30 millimeters, length of the cylinder 8 meters,
spacing of the ring stiffness is 800 mm center and take 𝛔 y as 433 newton per mm square.

(Refer Slide Time: 12:17)

So, let us do this. So, let us say the solution. First let us try to find out the stability
requirements, then we will use DNV recommended practice C 202 code. So, according to
section 3.1 of the code when the moment I say code we are referring to this of the code. The
stability requirements of a circular cylinder under axial stress is given by the following
equation 𝛔 d should be less than or equal to f of ksd we call this equation number 1.

Where, 𝛔 d is known as design shell buckling strength ok where f ksd is called characteristic
buckling strength. This is given by f ks byγm where f ks is the characteristic buckling
strength ok not ksd. So, now, the question is how you find f ks.

609
(Refer Slide Time: 14:31)

Let us go for characteristic buckling step f ks. So, we know f ks is given by f y by root of 1 +
lambda 4 we call equation number 2 where, lambda square is given by f y by 𝛔 d of 𝛔 ao by f
ea + 𝛔mo by f em + 𝛔 ho by f eh + tau naught by f e tau equation number 3. And 𝛔 d the
design strength is given by the square root of squares of 𝛔 a + 𝛔 m square - 𝛔 a + 𝛔 m of 𝛔 h
+ 𝛔 h square + 3 tau square we call equation number 4.

They are available in D N V R P C 202 code. Let us substitute the values we know the design
axial stress which is applied on the problem 𝛔 a is 30 which is given in the problem ok and
the design bending stress 𝛔 m we should take it as 0 because we are saying it is under pure
axial stress.

610
(Refer Slide Time: 16:38)

So, then the design circumferential hoop stress 𝛔 h is also 0 in our problem. The design shear
stress tau is also set to 0. Now, I substitute this in equation 4 ok by substituting in equation 4
we get 𝛔 d as 30 mega pascals. The characteristic buckling strength of circular cylinder under
shell buckling is given by this following equation of circular cylinder for check under shell
buckling is given by section 3.4.2 of page 15 of the code.

I am producing the equation again here f E is given by C times of 𝜋square E by 12 1 - mu

square of square t by l the whole square we will call it equation number 5. Whereas, in the
above equation C is called reduced buckling coefficient is given by psi times of root of 1 +
rho ε by psi the whole square. So, equation number 6.

611
(Refer Slide Time: 19:07)

Now, what we do? We use table 3.2 of the code and we say psi is unity. And zeta is 0.702 z l
where z l is l square by r t root of 1 - mu square we call this equation number 7 a this as
equation number 7 b and this further ok this as 7 b and rho is given by 0.5 times of 1 + r by
150 t this is r ok whereas, this is nu - 0.5. We call it as 7 C.

Now we know the spacing the stiffener the spacing of the stiffener given is l which is 800
mm. We also know the cylinder radius which r is 2500 mm we also know the cylinder
thickness which is t is 30 mm we also know the Poisson’s ratio which is nu which is 0.3.

(Refer Slide Time: 21:11)

612
Now substituting this in equation 7 we get Z l as 7.765 zeta as 5.451 and rho as 0.401 now
from equation 6. Let us say equation 6 we get now the reduced strain is it not equation 6
gives me the reduced buckling coefficient is given by 2.404. You can see here in this equation
we know all the values you can substitute and get C.

Now the characteristic buckling strength is given by f E which is equal to C times of 𝜋 square
E 12 1 - nu square of square t by l the whole square you know that is the equation we have.
Let us substitute 2.404 𝜋 square 2.5 10 power 5 we are talking about t ok 12 1 - we have
taken this as 0.3 the Poisson’s ratio. So, 1 - 0.3 square of square t by l the whole square am I
right. So, this value comes to about 641.64 newton per mm square.

(Refer Slide Time: 23:08)

So, now let us check for the stability from equation 3. What is equation 3? Let us see this
equation here from equation 3 we know this value can be now obtained for substitution for
the equation 3 lambda bar square is given by f y by 𝛔 d times of 𝛔 ao by f Ea right remaining
all are set to be 0s is it not which is coming to be 0.675; substituting this in equation 2. What
is equation 2?

See here equation 2 will give you the characteristic buckling strength ok k s f ks is given by f
y by root of 1 + λ4 right. So, let us do that which is going to be 297.514 newton per mm
square. So, now, using section 3.1 page 13 of the code ok, we get γm as 0.85 + 0.60 lambda
bar. If lambda bar is between 0.5 and 10 have a lambda bar 5 lambda bar is 0.675 therefore,
we get γm as 1.343.

613
(Refer Slide Time: 25:44)

Once we get γm I can now find f ksd which is nothing but f ks by γm which is going to be
297.514 by 1.343 is it not which is going to be 221.54 newton per mm square. Now, since the
design strength is 30 which is much lesser than f ksd we can conclude that the circular
cylinder is safe against shell buckling.

(Refer Slide Time: 26:54)

Let us do one more exercise we will say calculate the characteristic buckling strength of the
stiffened cylindrical shell with the following data. Let us see what the data are given. So,
diameter of the shell is 10 meters thickness of the shell is 40 millimeters. Let us say we are

614
using stiffness as follows. Using a flat bar of size 300 by 40 mm as stiffener ring frame
spacing is 600 mm center center. Let us say the axial compressive stress is 40 mega pascal
and circumferential compressive stress is 70.4 sorry is 70 mega pascal and yield strength of
the material is 433.

(Refer Slide Time: 29:02)

Let us solve this problem. So, 1 let us find the characteristic buckling step I am using the
same code rather the same class. f ks is given by f y by root of 1 + lambda 4 we call this
equation number 10 and we know lambda bar square is given by f y by 𝛔 d of 𝛔 ao by f E 𝛔 +
𝛔 m o by f Em + 𝛔 ho by f Eh + tau naught by f E tau equation 11.

We already wrote this equation from the code and we also know 𝛔 d is given by square root
of 𝛔 a + 𝛔 m the whole square - a + m into 𝛔 h + 𝛔 h square + 3 tau square equation number
12. We know the design axial stress 𝛔 a is - 40 because it is compressive design bending
stress 𝛔 m is 0.

615
(Refer Slide Time: 30:57)

Further, design circumferential stress is 𝛔 h - 70. And design shear stress tau is 0. Substituting
in the above equation we will get 𝛔 d as 60.828 newton per mm square. Let us now compute
the elastic buckling stress. So, we use section 3.3.2 on page 14 of the code. Let us talk about
the aspect ratio l by s. Let us see what the value is where, l is the distance of the say spacing
of the ring stiffness which is in this case 1000 mm.

Now, distance between the longitudinal stiffness. We call this as s which is 600 mm. So, the
aspect ratio in our case is 1000 by 600 which is 1.667 which is greater than 1.

(Refer Slide Time: 33:02)

616
When it is greater than 1, then the characteristic buckling strength of the cylindrical shell
checked for shell buckling is given by f E is C times of 𝜋 square E by 12 1 - nu square t by l
the whole square.

We call this equation number 13. Now the reduced coefficient C reduced buckling coefficient
C is given by psi times of square root of 1 + rho epsilon by psi the whole square equation 14.
So, from table 3.2 of the code psi is 4. So, zeta will be0.702 z s is s square by rt of root of 1
-γ2 or nu square rho is 0.5 square root sorry 1 + r by 150 t - 0.5. We call this equation number
15.

(Refer Slide Time: 35:02)

So, now we already know the spacing between the longitudinal stiffness s is 600 mm,
cylinder radius r is 5000 mm, cylinder thickness t as 40 mm and Poisson’s ratio nu is 0.3. So,
substituting these values we get z l as1.638 zeta as 1.15 and rho as 0.369. So, substituting in
equation 14 we get, what is equation 14? This is equation 14 ok we get C we get C as 4.022.

617
(Refer Slide Time: 36:30)

Therefore, the characteristic buckling strength is given by f Ea under axial a C times of 𝜋

square E by 12 1 - mu square t by l the whole square. So, which will become 4.022 times of 𝜋
square 2.1 10 power 5 12 1 - 0.3 square of square 40 by 600 the whole square which now
becomes 3392.78. Now, the elastic buckling strength under circumferential stress is given by
values from table 3.2. So, we say psi is 1 + s by l the whole square ok and ok we call this
equation number we call this equation number 16 a.

(Refer Slide Time: 38:16)

618
Zeta is 1.04 s by l root z s ok and z l is s square by rt root of 1 - nu square. Now, we know rho
is 0.6 s which is the distance between the longitudinal stiffness is 600 mm cylinder radius is
5000 millimeters, Poisson’s cylinder thickness is 40 mm, Poisson’s ratio is 0.3. So,
substituting in equation 14 we get psi as 1.85 equation 14 is what we have here. Equation 16
a psi as 1.85 and zeta as 0.799 and these are all equation 16 b, 16 c and so on.

(Refer Slide Time: 39:58)

So, now substituting this in equation 12 let us go back to this and try to find out capital C as
1.91 one in equation 12 is what C for C here oh not equation 12 equation 14 sorry equation
14. So, therefore, the characteristic buckling strength is given by f Eh C times of 𝜋 square E
by 12 1 - nu square square t by l the whole square which will be 1.911 𝜋 square 2.1 10 power
5 12 1 - 0.3 square square 40 by 600 the whole square which becomes 1612.04 newton per
mm.

Now, lambda bar square is given by f y by 𝛔 d see this equation, equation 11. So, substituting
in equation 11 f y by d 𝛔 ao by f Ea + 𝛔 ho by f Eh which is 0.393 and therefore, lambda bar
will be 0.627.

619
(Refer Slide Time: 42:01)

Hence the characteristic buckling strength is given by f ks, which is f y by root of 1 + lambda
4 which is 402.996 mega pascal.

(Refer Slide Time: 42:34)

So, friends in this lecture we learnt an example. In fact, of finding the buckling strength of
stiffened cylinder in fact, we have used an international code and checked the stability as
recommended by the code. So, that is the procedure how buckling strength or buckling stress
estimates or practiced and applied in design as far as international codes are concerned.

620
So, friends in these like set of lectures on stability we learnt different types of stability. We
learnt stability is a mode of failure and we have learnt the difference between stability and
buckling and we have derived the stability functions for axial compression, 0 axial load and
axial tension.

The rotational and translational functions of stability equations from the first principles and
we really understood that the stiffness matrix formulated for a stability problem is more or
less similar to that of a matrix generated for a simple analysis of statically indeterminate
structure.

Provided you remember the order of the label and remember the stiffness matrix. Once you
have this matrix we have formulated the characteristic equation where we set k uu
determinant to 0 for a non trivial solution and we obtained the characteristic equation. We
used a stability charts or stability tables available in the literature to choose the psi value
sorry the phi value which is the ratio between the axial force and the Euler’s critical load.

And we found out that to satisfy this equation and we determine this value. We also exposed
to you the MATLAB programs useful to plot the stability functions and also subroutines to
solve the problems to estimate the critical buckling loads. In the next lectures we will talk
about unsymmetric bending, curved beams and crane hook design and analysis using
standard equations and MATLAB applications.

Thank you very much and have a good day.

621
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 36
Unsymmetric bending - 1

(Refer Slide Time: 00:18)

Friends, welcome to the 36th lecture on Advanced Steel Design, where we are going to learn
something about Unsymmetric Bending. So, we calls unsymmetric bending, lecture 1. In the
previous lectures, we have learnt the importance of plastic design form resistant design,
material properties at elevated temperature for steel, newly proposed materials as functionally
graded materials and their advantages.

We also did some examples on plastic design and analysis, then for form dominant design we
understood that stability which is one of the modes of failure is important to assess and we
developed stability functions for various application conditions of axial compressive load,
axial tensile load and etcetera and we did some problems and used MATLAB intensively to
find out the critical buckling load so that I can assess stability of this systems.

Now, unsymmetric bending is another area of challenge in advanced design procedures.

Generally, friends, in symmetric bending, bending takes place in the plane parallel to the
plane of applied movement. Let us try to explain this with some graphical illustration. So, let

622
us take a beam. So, let us say this is my x-axis of the beam which is along the span of the
beam and let us call this as y and you know the z-axis is normal to this laptop screen.

So, with that axis I apply a moment about that axis I call that as 𝑀𝑧. So, 𝑀𝑧 is the moment

about z-axis whose trace is seen on the XY plane, right. So, now, if I cut a section, if I cut a
section and if the section cross-section is known to me where this becomes my x-axis now
and I call this as y and this as z-axis, correct? Imagine this figure 1 and figure 2. Figure 1
positive z is normal to the laptop screen, look it from the side.

So, you will see y in the vertical direction and z towards left side, am I right. So, I think I will
remove this write it better here and remove this maybe write it here. So, now, I can say a
about in the cross-section of the beam y-y axis is the trace of applied moment; z-z axis is the
axis about which moment is applied. We call that is y as 𝑀𝑧.

(Refer Slide Time: 05:21)

Having said this, we can now write a condition saying bending takes place in a plane parallel
to the plane of applied moment.

So, therefore, friends we know that moment about y-y axis is kept as zero. Mathematically
algebraic sum of moments about y is 0, we calls equation number 1. Therefore, there are no
external moments acting about y-y axis, right. Since there are no external moments acting
about y-axis the internal resisting moment about y-y axis should be 0.

623
So, can I say mathematically 𝑀𝑦 should be integral of over area A sigma x dA, 0 equation 2.

So, now, we also know

σ 𝐸
𝑦
= 𝑅

𝐸
, hence 𝑅
into sigma x d A of the area a should be equal to 0. So, now, sigma x is the product

of moment of inertia.

(Refer Slide Time: 08:18)

So, let us do that way saying it should be same as over the area A I should say z y dA should
be termed as 0, we call this equation 3. Do not mistake it as stress it is not stress to ensure this
condition, we should imply that they should be satisfied. So, we can say z y dA should be
termed as 0. We call this 3(a) the above equation is only true when Z – Z and Y – Y axis are
principal axis.

The above equation 3(a) is true only when Z – Z and Y – Y axes or principal axes. Hence for
the symmetric bending there are certain conditions. 1 - it is essential that the plane containing
a – principal axes of inertia; b – plane of applied moment; c – plane of deflection should
coincide, then only a symmetric bending can happen.

624
(Refer Slide Time: 10:51)

The second condition for the symmetric bending to happen is, it is also obvious that the
neutral axis will coincide with other principal axes of inertia. So, therefore, friends, when the
trace of plane of applied moment does not coincide with any of the principal axes of inertia,
then it is called unsymmetric bending.

This is also called by another name which is non uniplanar bending. This is another name of
unsymmetric bending.

(Refer Slide Time: 12:55)

625
Now, the question comes if a member or a cross-section is subject to unsymmetric bending
what will be the consequence?

So, we will ask a question what are the consequences if a member is subjected to
unsymmetric bending. 1st, the neutral axis will not be perpendicular to the trace of the
applied moment plane it makes an inclination. 2 – The deflection curve is also not a plane
curve or we should say will not be a simple plain curve in the plane of applied moment.

3 – members which are not symmetric because you may get members which are not
symmetric in cross-section. So, members which are not symmetric about the vertical axis and
have a special cross-section, then these members will undergo twisting under transverse
loads, that is a very serious consequence we have when the members are subjected to
unsymmetric bending.

The 4th could be to avoid twisting of unsymmetric section load should be applied at the shear
center.

(Refer Slide Time: 16:33)

Now, the question comes what is a specialty in analysis if a member is subjected to

unsymmetric bending. The answer is the applied moment, in such cases will cause bending
about both the principal axes of inertia. Therefore, stress at any section or at any point in the
cross section cannot be determined using the classical theory of bending that
𝐸 σ 𝑀
𝑅
= 𝑦
= 𝐼

626
This equation cannot be used to determine the stress. So, it is very complicated problem now.

(Refer Slide Time: 18:41)

Now, what do I do to determine the stress when a member is subjected to unsymmetric

bending?

So, the question is then how to determine the stress at any point in the cross-section of a
member under unsymmetric bending. To do that, it is necessary to transform the problem of
unsymmetric bending into uni-planar bending. So, we need to do a transformation of the
problem.

Now, the question comes how will you transform the problem into uni-planar bending? that is
a interesting question because we need to find the stresses at the cross sections.

627
(Refer Slide Time: 20:16)

Let us look into that. So, the objective is now to transform unsymmetric bending problem to
uni-planar bending that is objective now. Let us take a cross-section. Let us mark these two
axes as z and y; x is normal to the screen. Let us also mark two more additional axis which is
called the v and u axis.

𝑢 = 𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α

𝑣 =− 𝑧 sin 𝑠𝑖𝑛 α + 𝑦 cos 𝑐𝑜𝑠 α

Let us measure the angle of u-axis from z-axis or v-axis from y-axis as α. Let us take a point
whose area is dA, it has got specific coordinates u and v or x, y depending upon our own
choice of coordinates. Let us refer this as figure one a typical cross section of a member.

One can very well see here that uu axis is inclined α to zz axis α is measured from z-z axis
clockwise to mark u-u axis, that is a condition here. So, to be very specific α is measured in
the positive V coordinate. So, we have positive V, we measure from that. So, from this figure
I can write a basic relationship which is very easy write down that. So, from figure 1, we can
write the following equation.

What are they? 𝑢 = 𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α . We call this equation number 1; it is a basic
equation written with reference to figure number. having said this, I want to find the moment
of inertia.

628
(Refer Slide Time: 24:27)

Let us say to find the moment of inertia about uu and vv axes remember they are inclined to
zz and yy. So, it is not that easy to use a principle parallel axes or perpendicular axes theorem
to find out that. They are inclined by an angle α. So, it is not that easy to find out 𝐼𝑢𝑢 and 𝐼𝑣𝑣.
2
So, let us use first principles. So, from first principles, we know 𝐼𝑢 is integral 𝑣 dA for the
2
whole area A. 𝐼𝑣 is integral 𝑢 dA for the whole area a. We call this equation as 2(a).

Now, we have expression for u and v? Do we have expressions for u and v, let us substitute it
here and see what happens. So, now, I v becomes integral over area A u is z of cos α, v is
minus z z sin α. So, minus z sin α plus y cos α the whole square dA. So, I am getting I u, I
am getting I u. So, v is here I am using this relationship. Let us expand this.

629
(Refer Slide Time: 26:36)

So, let us say I u is equal to integral z square sin square α plus y square cos square α minus
2zy sin α cos α of dA. So, which will be sin square α of z square dA whole area A plus cos
square α of y square dA whole area A minus sin 2 α integral zy A dA.

How do we write this? We know sin 2α is 2 sin α cos α. We also know from the first
principles that z square dA for the whole area A can be said as I y integral y square dA for the
whole area A can be said as I z when we know this then I can write I u from the above
equation as,

2 2
𝐼𝑢 = 𝐼𝑦𝑠𝑖𝑛 α + 𝐼𝑧𝑐𝑜𝑠 α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

We will call this equation number 2.

Let us simplify this. Let us copy this equation here. I will use a tool. I will copy this equation.

630
(Refer Slide Time: 29:27)

Let me copy it here. So, I am going to simplify this equation. Let me remove this number.

2 2
𝐼𝑢 = 𝐼𝑦𝑠𝑖𝑛 α + 𝐼𝑧𝑐𝑜𝑠 α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

So, I am going to replace these two values with these new expressions.

2
cos 𝑐𝑜𝑠 2 α = 1 − 2𝑠𝑖𝑛 α
2
cos 𝑐𝑜𝑠 2 α = 2𝑐𝑜𝑠 α − 1
2 1−cos𝑐𝑜𝑠 2α
𝑠𝑖𝑛 α = 2

2 1+cos𝑐𝑜𝑠 2α
𝑐𝑜𝑠 α = 2

𝐼𝑦 𝐼𝑧
𝐼𝑢 = 2
(1 − cos 𝑐𝑜𝑠 2α) + 2
(1 + cos 𝑐𝑜𝑠 2α) − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

𝐼𝑦+𝐼𝑧 𝐼𝑧−𝐼𝑦
𝐼𝑢 = 2
+ 2
cos 𝑐𝑜𝑠 2α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

call this equation number 3. So, let us take equation number 3 as one of the important
reference derivatives.

631
(Refer Slide Time: 32:19)

2
Now, let us move on to 𝐼𝑣. Now, let us find 𝐼𝑣. We know 𝐼𝑣 is integral 𝑢 dA for the entire A;

we already know u we already know u as this value, 𝑢 = 𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α . Let us
substitute it here which is going to be equal to integral for the whole area
𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α the whole square dA. Let us expand this which will be z square
cos square α y square sin square α plus 2z y sin α cos α of dA integrated whole area A.

Now, I can write 𝐼𝑣 as cos square α integral z square dA for the whole area A plus sin square

α integral y square dA for the whole area plus sin 2 α integral whole area z y dA. I think there
is no confusion in writing this expression. Now, I can write this as z square dA is 𝐼𝑦 cos

square α y square dA is I z sin square α plus this is going to be 𝐼𝑦𝑧 sin 2 α. This is 𝐼𝑣, Now, let

us take this 𝐼𝑣 and expand it.

2 2
𝐼𝑣 = 𝐼𝑦𝑐𝑜𝑠 α + 𝐼𝑧𝑠𝑖𝑛 α + 𝐼𝑧𝑦 sin 𝑠𝑖𝑛 2α

We have a substitution.

632
(Refer Slide Time: 34:41)

𝐼𝑦 𝐼𝑧
𝐼𝑣 = (1 + cos 𝑐𝑜𝑠 2α) 2
+ (1 − cos 𝑐𝑜𝑠 2α) 2
+ 𝐼𝑧𝑦 sin 𝑠𝑖𝑛 2α

𝐼𝑦+𝐼𝑧 𝐼𝑧−𝐼𝑦
𝐼𝑣 = 2
− 2
cos 𝑐𝑜𝑠 2α + 𝐼𝑍𝑌 sin 𝑠𝑖𝑛 2α

𝐼𝑦+𝐼𝑧 𝐼𝑧−𝐼𝑦
𝐼𝑢 = 2
+ 2
cos 𝑐𝑜𝑠 2α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

𝐼𝑢 + 𝐼𝑣 = 𝐼𝑦 + 𝐼𝑧

I get this classical relationship, I call this equation number 5. It is a classical relationship.

633
(Refer Slide Time: 37:31)

Let us also find out the cross product which is 𝐼𝑢𝑣. We know 𝐼𝑢𝑣 is given by integral u v dA

over the area A. So, we already have the values for u and v; let us see here in equation. We
have equation 1, we have u and v. So, let me write it here. So, substitute for u and v from
equation 1.

So, let us do that which will be equal to integral z cos α plus y sin α into minus z sin α plus y
cos α of dA for the whole area which now become integral minus z square sin α cos α minus
y z sin square α plus y z cos square α plus y square sin α cos α. This is for the whole area.

634
(Refer Slide Time: 39:18)

Let me copy this equation take it to the next screen for our reference which is now this is 𝐼𝑢𝑣.

So, 𝐼𝑢𝑣 is now. So, we have two terms sin α cos α twice. So, what I will do that is I can write

sin α cos α integral z square dA, let us have a minus sign plus sin α cos α integral y square dA
plus y z dA. we will do it. This way plus y z cos square α minus y z sin square α dA for the
whole area A.

So, now I can expand this further. So, we know this is minus of 𝐼𝑦 sin α cos α and this is 𝐼𝑧.

So, plus 𝐼𝑧 of sin α cos α plus you know y is A dA is 𝐼𝑦𝑧, cos square α minus sin square α, am

I right? Which now amounts to 𝐼𝑧 minus 𝐼𝑦 of sin 2 α by 2 plus 𝐼𝑦𝑧 cos 2 α, equation 6. This is

my 𝐼𝑢𝑣 it is a classical equation.

(𝐼𝑧−𝐼𝑦)
𝐼𝑢𝑣 = 2
sin 𝑠𝑖𝑛 2α + 𝐼𝑦𝑧 cos 𝑐𝑜𝑠 2α

Friends, we must subsequently substitute 𝐼𝑢𝑣 to be 0. So, now subsequently we should

substitute 𝐼𝑢𝑣 as 0. Why do we have to do that? In simple terms, 𝐼𝑢𝑣 need to be indicated as an

understanding of explaining non-symmetric bending.

635
(Refer Slide Time: 42:50)

𝐼𝑢𝑣 = 0;

(𝐼𝑧−𝐼𝑦)
𝐼𝑢𝑣 = 2
sin 𝑠𝑖𝑛 2α + 𝐼𝑦𝑧 cos 𝑐𝑜𝑠 2α

tan 𝑡𝑎𝑛 (2α) =− ( )

2𝐼𝑦𝑧
𝐼𝑧−𝐼𝑦

So, now friends, u and v being principal axes of inertia 𝐼𝑢𝑣 should be 0. So, what is 𝐼𝑢𝑣? I am

just copying this equation. So, if you equate this to 0, I will get tan 𝑡𝑎𝑛 (2α) =− ( )
2𝐼𝑦𝑧
𝐼𝑧−𝐼𝑦
.

We call this equation number 7. So, friends we are now interested to find out stress at any
point in a given cross section.

636
(Refer Slide Time: 44:38)

So, our objective is to find stress at any point in the cross-section. That is objective. So, we
will look into this in detail in the next lecture.

Let us now write the summary what we learnt in this lecture. We learnt what is meant by
unsymmetric bending or non-uniplanar bending. We have also learnt what are the
consequences if a member is subjected to unsymmetric bending. If I want to find the stress at
any point P in a cross section or a member under unsymmetric bending, what do I do what do
we or how do we calculate. So, we said that we need to convert the unsymmetric bending
problem into a uniplanar bending problem.

To do that we need to find the principal moments of inertia and we equated the product to 0
being principal axis and to find the inclination of α. Remember α is an angle is an angle of
u-u axis with z-z axis measuring from the positive z to positive u or positive v to positive y, in
a specific manner. So, if I know this using the equation, we will be able to find the stresses.
That is what we discussed briefly and we will explain the equation how to find stresses in
case of an unsymmetric bending in the next lecture.

Thank you very much and have a good day.

637
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 37
Unsymmetric bending - example problems

(Refer Slide Time: 00:20)

Friends, welcome to the 37th lecture on Advanced Steel Design course, in this lecture we are
going to learn Unsymmetric bending with example problems. So, we learnt in the last lecture
how to identify the u-u and v-v axis which are principal axis of inertia, their inclination with
respect to y z axis of the cross section and the meaning of the angle α and how to compute
this α angle if I know the principal moments of inertia, we have learnt these equations in the
last lecture.

So, the question asked was; if a section or a member is subject to unsymmetric bending, how
do we find the stresses at any point in the cross section?

638
(Refer Slide Time: 01:37)

If a member is subjected to unsymmetric bending, how do we compute stresses at any point

in the cross section because we said that the classical bending equation cannot be used to find
out the stresses. So, how do we get this? So, let us explain this with a figure let us say I have
a cross section which is arbitrary whose area is no, let us say this is my y axis and this
remains as z axis and x is normal.

Let us mark it here let us mark it here as to make it easy, I will mark z here I will mark a
three-dimensional view I will put x here, if this is x axis; obviously, this becomes my
member. length of the member let us say. Now, I draw a plane or I draw the u-u, v-v axis. So,
I draw the u-u axis at an angle α from z axis and also draw the v-v axis here this is u axis. Let
me draw the shaded portion representing a specific plane. So, this hatched portion represents
a specific plane which is x-u plane.

So, I will say this is write it here this is x-u plane this is x-u plane, this is the plane of
symmetry is the plane of symmetry, hence 𝐼𝑢𝑣 should be 0 about that. So, once we establish

this relationship, we can use conventional bending equation to get the stresses let us see how.

639
(Refer Slide Time: 05:09)

Let us say give me y axis and we mark this as my z axis, an axis here I mark this as my
u-axis, this as my v-axis and we know that this angle is α measured from the positive z to
positive u this is. So, this is my first quadrant remember that positive y positive z. So, it is the
first quadrant.

So, I take any point p here which is marked in terms of u and v. Remember; for u and v this
the first quadrant. Now let us put a moment which is 𝑀𝑧 acting about this axis, we call this

moment as 𝑀𝑢 and this moment as 𝑀𝑣. Now I can write a relationship using the sign

convention right screw rule. So, we can say 𝑀𝑢 is 𝑀𝑧cos α and 𝑀𝑣 is 𝑀𝑧sin α, but minus we

call this equation number 8 for continuity what we had from the previous lecture.

𝑀𝑢 = 𝑀𝑧 cos 𝑐𝑜𝑠 α

𝑀𝑣 =− 𝑀𝑧 sin 𝑠𝑖𝑛 α

𝑀𝑢
σ(𝑢,𝑣) = 𝐼𝑢
(v) (Compression)

𝑀𝑣
σ(𝑢,𝑣) = 𝐼𝑣
(u) (Tension)

640
So, interestingly friends if you look at 𝑀𝑢 if you look at 𝑀𝑢 it produces bending this form.

So, tension at the bottom compression at the top this is 𝑀𝑢, can you visualize that? But on the

other hand, if you see 𝑀𝑧, 𝑀𝑧 is reverse the tension at the top compression at the bottom. Let

us take any point u-v we write the equation. So, the stress at the point u v is given by two
combinations 2 components one component will come from 𝑀𝑢 by 𝐼𝑢 into v, the other

component will come from 𝑀𝑣 by 𝐼𝑣 into u.

Looking at these figures the component which is caused by u will be compressive, at the
same location the component caused by 𝑀𝑣 will be tensor, Let us take an extreme top fibre

that is what the situation is.

(Refer Slide Time: 10:11)

Now, we can say

𝑀𝑢 𝑀𝑣
σ𝑃(𝑢,𝑣) =− 𝐼𝑢
(𝑣) + 𝐼𝑣
(𝑢)

but; however, this is minus because this is a compressive stress given by equation 9. So,
friends, equation 9 is similar to a classical bending equation provided I have replaced the
moments properly resolved to u and v axis.

641
And the components of measurement and the moment of inertias or with respect to the
principal axis. So, I am transforming the problem from an unsymmetric bending to uni-planar
bending where the moments of inertia are now transformed to the principal axis of moments
of inertia and I am trying to find out the moments about this axis and using a classical
bending theory. Let us see, what are the steps involved in finding the stress? what are the
steps involved to find the stress at any point of a member under unsymmetric bending.

First step find 𝐼𝑥, 𝐼𝑦, 𝐼𝑧, 𝐼𝑧𝑦 of the cross section if you know this find 𝐼𝑢, 𝐼𝑣, and 𝐼𝑢𝑣. So, then

determine α that is this will help you to fix the u-u axis and v-v axis with respect to z y axis.
Remember, α is measured from the positive psi of positive z axis, then use equation 9 to
compute this stress, these are the steps involved. To demonstrate this, we will take up a
simple problem and gradually transform the problem into a very interesting theory, we will
do that.

(Refer Slide Time: 14:27)

So, let us say example 1 we will take a T section this dimension is 60. And this dimension is
10,70 and this thickness is 10. We will divide this into 2 components we will call see
component number 1 and component number 2. So, now, this is the beam which is simply
supported for a span of 3 meters subjected to an inclined load of 50 kilo newtons, the
inclination is 60 degrees to the vertical. So, let us draw the coordinate axis for this, this is x,
let me do it here this is x axis, this is y axis and this is z axis.

642
Let us mark two points one point is A other point is B. So, the question is, find the stresses at
A and B?. So, we need to locate the neutral axis. So, now, interestingly this has one axe of
symmetry let us locate the z axis here and this is going to be my y axis is a cross section,
because x is along the length of the member see. Let me call this distance as y bar and from
first principles I can easily find y bar let us do that.

So, we know,
Σ𝑎𝑦 (60×10×5)+(70×10×45)
𝑦= Σ𝑎
= (60×10)+(70×10)

𝑦 = 26. 54𝑚𝑚

(Refer Slide Time: 18:36)

Let us say

𝐴(𝑧, 𝑦) = (− 30, 26. 54)

𝐵(𝑧, 𝑦) = (5, − 53. 46)

Having said this let us compute 𝐼𝑦 it has got axis of symmetry. So, we can use simple theorem

to compute this let us do this going to be

3 3
10×60
𝐼𝑦 = ⎡⎢ 12 +
70×10 ⎤ = 1. 86×105 𝑚𝑚4
12 ⎥
⎣ ⎦
𝐼𝑦 = 𝐼𝑣

643
Let us find out I z I will use parallel axis theorem. So, I will say

3 3
60×10 2 10×70 2
𝐼𝑧 = ⎡⎢ 12 + (60×10)(26. 54 − 5) ⎤⎥ + ⎡⎢ 12 + (70×10)(45 − 26. 54) ⎤⎥
⎣ ⎦ ⎣ ⎦
5 4
𝐼𝑧 = 8. 07×10 𝑚𝑚

𝐼𝑧 = 𝐼𝑢

Now I can straight away say since the axis symmetry is matching, I can straight away say u-u
axis will be same as z-z axis, v-v axis will be same as y-y axis because of symmetry
therefore, I can say 𝐼𝑢 is same as 𝐼𝑧 and 𝐼𝑣 is same as 𝐼𝑦. So, I can now say this is 𝐼𝑢 and this is

𝐼𝑣, 𝐼𝑦 is 𝐼𝑣 and 𝐼𝑢 is 𝐼𝑧. Since, y-y axis is symmetric axis I can straight away say 𝐼𝑧𝑦 will be 0

which will give me the angle of α. 𝐼𝑧 will be 0, I do not have to calculate that.

(Refer Slide Time: 22:30)

Now, let us find out 𝑀𝑧. So, 𝑀𝑧 will be in the y- z plane see the original equation. So, it is

going to be

𝑀𝑧(𝑦 − 𝑧 𝑝𝑙𝑎𝑛𝑒) = 50 cos 𝑐𝑜𝑠 60 ( ) =+37.5 kNm (Hogging BM)

3
2

we can write it here itself 𝑀𝑧. 𝑀𝑧 is going to be +37.5 kNm which is a hogging moment I will

say it is positive.

644
The load is downward hogging moment tension at the bottom. Now I can straight away say
𝑀𝑧 is plus 37.5 kNm hogging bending moment. So, I can now say

𝑀𝑢 = 𝑀𝑧 cos 𝑐𝑜𝑠 α = 𝑀𝑧 = + 37. 5𝑘𝑁𝑚

𝑀𝑣 =− 𝑀𝑧 sin 𝑠𝑖𝑛 α = 0

α = 0;

𝑀𝑣 is minus 𝑀𝑧 sin 𝑠𝑖𝑛 α we have this equation with us. So, α 0 so, this is going to be 0. So,

we can also write tan 2 α which will also become; obviously, α is 0.

So, let us compute σ𝐴. So, if you want to find σ𝐴 at

𝑀𝑢 𝑀𝑣
σ𝐴(−30,26.54) =− 𝐼𝑢
(𝑣) + 𝐼𝑣
(𝑢)

6
−37.5×10
σ𝐴(−30,26.54) = 5 ×26. 54+0
8.07×10

𝑁
σ𝐴(−30,26.54) =− 1233. 27 2 (Compression)
𝑚𝑚

(Refer Slide Time: 25:49)

Let us find out σ𝐵 at the point (5,-53.46),

645
 𝑀𝑢 𝑀𝑣
σ𝐵(5,−53.46) =− 𝐼𝑢
(𝑣) + 𝐼𝑣
(𝑢)

6
−37.5×10
σ𝐵(5,−53.46) = 5 ×(− 53. 46)+0
8.07×10

𝑁
σ𝐵(5,−53.46) =+ 2484. 2 2 (Tensile)
𝑚𝑚

Friends, also see this figure the point A is lying on the extreme top fibre. So, point A will be
subjected to compression because the hogging bending mode and the bottom fibre will be in
tension. So, we have got this is tensile and this is compressed.

(Refer Slide Time: 27:40)

Let us do one more problem let us do one more problem which is example 2 here we take a
cantilever beam subjected to a load of 20 kN for a span of 4 meters. So, the beam is going to
bend this way. So, let us mark the axis, let us say this is my y-axis, this is my z-axis and this
becomes my x-axis let me draw the section.

So, this is my section I am looking for a z section. So, the z section is marked here say this is
100 mm and the whole depth is 200 mm and the thickness is 15 mm. So, we all know we got
two axis of symmetry taking this also as 100, now we will mark this axis as my Y-axis and
this axis as my Z-axis an axis somewhere here and I will apply 𝑀𝑧 here where this 𝑀𝑧 will be

equal to 80 kNm.

646
So, here actually the load is applied and the section is going to bend this way and so on. So,
we want to find two points this is A and this is B. So, to find stress as A and B that is the
question.

(Refer Slide Time: 31:27)

Let us find 𝐼𝑦 for this problem which is a straightforward exercise because it has got two axis

of symmetry.

3
15×100 2
𝐼𝑦 = ⎡⎢ 12 + (15×100)(50 − 7. 5) ⎤⎥+
⎣ ⎦
3 3
⎡ 15×100 + (15×100)(50 − 7. 5)2⎤ + ⎡ 170×15 ⎤
⎢ 12 ⎥ ⎢ 12 ⎥
⎣ ⎦ ⎣ ⎦

7 4
𝐼𝑦 = 0. 79×10 𝑚𝑚

647
(Refer Slide Time: 32:18)

One can find I z z is an horizontal axis. So, I can say

3
100×15 2
𝐼𝑧 = ⎡⎢ 12 + (100×15)(100 − 7. 5) ⎤⎥+
⎣ ⎦
3 3
⎡ 100×15 + (100×15)(100 − 7. 5)2⎤ + ⎡ 15×170 ⎤
⎢ 12 ⎥ ⎢ 12 ⎥
⎣ ⎦ ⎣ ⎦

7 4
𝐼𝑧 = 3. 18×10 𝑚𝑚

Since, z-y are axis of symmetry we know 𝐼𝑧𝑦 will be integral z by dA of the whole area.

648
(Refer Slide Time: 35:47)

(𝑧𝑦)𝑑𝐴1 = (100×15)(50 − 7. 5)(100 − 7. 5)

(𝑧𝑦)𝑑𝐴2 = 0

(𝑧𝑦)𝑑𝐴3 = (100×15)(− (50 − 7. 5))(− (100 − 7. 5))

7 4
𝐼𝑧𝑦 = 1. 18×10 𝑚𝑚

7 4
So, add them together 𝐼𝑧𝑦 will become 1. 18×10 𝑚𝑚 . Friends, very important about the z

sections, though it is not act of symmetry 𝐼𝑧 is not 0. Also, there is a note 𝐼𝑧𝑦 can also be a

negative value, but 𝐼𝑧 𝐼𝑦 will be always a positive value. Having said this, we are looking for

let us we will change the point B somewhere here we will mark this as point B and A we
want extreme fibres.

So, therefore, since I z y is not 0 we want to find the angle α. Let us do that.

649
(Refer Slide Time: 38:21)

𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦
𝐼𝑢 = 2
+ 2
cos 𝑐𝑜𝑠 2α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦
𝐼𝑣 = 2
− 2
cos 𝑐𝑜𝑠 2α + 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α

2𝐼𝑦𝑧
tan 𝑡𝑎𝑛 (2α) =− 𝐼𝑧−𝐼𝑦

7
2×1.18×10
tan 𝑡𝑎𝑛 (2α) =− 7 =− 0. 9903
(3.18−0.797)×10

𝑜
α =− 22. 36

So, what does it mean, when I say α is negative what does it mean? Measured clockwise from
positive z axis towards positive view that is what it is.

So, I should say. So, I will mark this figure I will take this figure again.

650
(Refer Slide Time: 41:05)

Let me mark the u-u and v-v axis and we know this angle is 22.36 degrees. So, we have
marked this angle which we obtained it is always measured from positive z towards positive
u that is what you have got. So, now, we know α is minus 22.36 degrees let us substitute that
and get 𝐼𝑢 and 𝐼𝑣. So, substitute and obtain 𝐼𝑢, 𝐼𝑣 also. Check, 𝐼𝑢 + 𝐼𝑣 = 𝐼𝑧 + 𝐼𝑦. you should

check this.

(Refer Slide Time: 42:30)

651
So, now we have 𝑀𝑧 as minus 80 kilo newton meter see here. So, this is a force and 𝑀𝑧 is

creating tension at the top.

𝑀𝑧 =− 80𝑘𝑁𝑚

𝑀𝑢 = 𝑀𝑧 cos 𝑐𝑜𝑠 α =− 80× cos 𝑐𝑜𝑠 (− 22. 36) = − 73. 98𝑘𝑁𝑚

𝑀𝑣 =− 𝑀𝑧 sin 𝑠𝑖𝑛 α =− (− 80)× sin 𝑠𝑖𝑛 (− 22. 36) = − 30. 44𝑘𝑁𝑚

𝑀𝑢 𝑀𝑣
σ𝑥 =− 𝐼𝑢
(𝑣) + 𝐼𝑣
(𝑢)

σ𝑁𝐴 = 0;

𝑀𝑣 𝑀𝑢
𝐼𝑣
(𝑢) = 𝐼𝑢
(𝑣)

So, straight away I can say tan 𝑡𝑎𝑛 β is an angle which is

𝑣 𝑀𝑣 𝐼𝑢
tan 𝑡𝑎𝑛 β = 𝑢
= 𝑀𝑢 𝐼𝑣

Now, what is beta? Beta is the angle of neutral axis measured from the u-u axis positive.

𝑜
β =+ 78. 51

(Refer Slide Time: 45:25)

652
Let us mark this I will go to this figure, pick up this figure again let me mark the beta angle
which will be. So, this be neutral axis and this is my angle beta which is 78.51 degrees. So, I
must know this distance and I must know this distance, to find the stresses. So, that is very
complicated. So, what I am going to do I am not going to compute these distances I want to
find out these distances in terms of u and v that is what we are trying to do.

So, now let us talk about the point A, point A has got z and y as plus 92.5 and plus 100.

A(z,y)=(+92.5,+100)

𝑢𝐴 = 𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α

𝑣𝐴 =− 𝑧 sin 𝑠𝑖𝑛 α + 𝑦 cos 𝑐𝑜𝑠 α

𝑢𝐴 = 92. 5 cos 𝑐𝑜𝑠 (− 22. 36) + 100 sin 𝑠𝑖𝑛 (− 22. 36)

𝑣𝐴 =− 92. 5 sin 𝑠𝑖𝑛 (− 22. 36) + 100 cos 𝑐𝑜𝑠 (− 22. 36)

We already said that look at this axis if this is my z and y if these are my u and v and this is
my α this equation is clear.

(Refer Slide Time: 48:20)

𝑀𝑢 𝑀𝑣
σ𝐴 =− 𝐼𝑢 (𝑣𝐴) + 𝐼 (𝑢𝐴) =+ 29. 36 𝑚𝑚𝑁
𝑣
2 (𝑇𝑒𝑛𝑠𝑖𝑙𝑒)

σ𝐵(𝑧,𝑦) = (− 92. 5, − 100)

653
 𝑁
σ𝐵(−92.5,−100)=−29.36 2 (𝐶𝑜𝑚𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛)
𝑚𝑚

And this indicates its compressive, look at the figure the point A is above the neutral axis and
I have a cantilever. So, I have tension at the top and I get this as tensile and compression at
the bottom get compression.

(Refer Slide Time: 49:31)

So, friends in this lecture we learnt how to solve 2 problems, how to solve example problems
on unsymmetric bending. We know how to use the equations; in the next lecture we will use
MATLAB program to solve this and see how we are able to use MATLAB effectively for
solving such problems.

So, in a z section interestingly when the neutral axis here, this point is tensile, this point is
compressive for a cantilever and we got them exactly verified from the stat. So, friends, this
lecture helps you to compute the stresses of a section subjected to unsymmetric bending by
converting it into uniplanar bending and then using the classical equation for finding out the
stresses with modifications in both M and I values. Use this in more examples are available in
reference textbooks.

Please, see them use the MATLAB programs effectively and try to solve as many problems as
possible. So, that you get good hold on understanding of the problem.

Thank you very much and have a good day bye.

654
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 38
Shear center - 1

(Refer Slide Time: 00:23)

Friends, welcome to the 38 lecture on Advanced Steel Design. In this we are going to learn
more example on unsymmetric bending. We are also going to use the MATLAB program for
solving this problem. We will also start learning more on Shear centre. So, friends we already
said in the last lecture, that unsymmetric bending is caused under special circumstances. And
we know how we have modified the classical equation of bending appropriately to find out
the stresses at any point in a cross section of a member subjected to unsymmetric bending.

655
(Refer Slide Time: 01:23)

We will now take up one more example. I will call this as example 3, because we have
already done couple of examples in the last lecture, I am continuing that; obviously, the
example says determine the stresses at the designated points in the cross section shown
below.

The member is a cantilever of 2-meter span with an edge load as shown. So, let us say; this is
my cross section, this is 30 mm and this is 50 mm and it has got an uniform thickness of 5
mm. And let us say this is a cantilever subjected to an upward loop of 2 km and the cantilever
span is 2 meters.

Let us say this is my z-y axis. The load is applied this way and normal to the screen is my x
axis. Let us say I have a centroidal axis for this. Let me draw the figure again here separately
to mark the geometric properties of this cross section. So, this is 15, this is of thickness 5
uniform and this dimension is 30. We call this as 1 and this as 2. So, there are two things here
this is the first element then the second element.

So, let us say it has got a centroid here. This is going to be my z axis this is going to be my y
axis and I would like to compute these values. I call this as y bar and I call this as z bar. So,
we need to find 𝑦 and 𝑦 we need to find. First principles very simple. let us do that. So,

Σ𝑎𝑦 (30×5×2.5)+(45×5×27.5)
𝑦= Σ𝑎
= [(30×5)+(45×5)]
= 17. 5𝑚𝑚

656
 Σ𝑎𝑧 (30×5×15)+(45×5×2.5)
𝑧= Σ𝑎
= [(30×5)+(45×5)]
= 7. 5𝑚𝑚

First principle is very simple I think there will be no difficulty in doing this. Now, I want to
compute 𝐼𝑍 and 𝐼𝑦, because we know to compute the principal moments of inertia. I need to

compute these values, because the equation is based on this.

(Refer Slide Time: 07:52)

So, now let us compute 𝐼𝑍. Looking at the reference to this figure I can use parallel axis

theorem. So, let me copy this figure for our convenience. Let me copy this figure and put it
here and this dimension is 50 let us mark it here, because we need it. This is 50 let us mark it
here. So, now, let me compute 𝐼𝑍, which is 30 into 5 cube by 12 plus parallel axis theorem, 30

into 5 into 17.5 minus 2.5 the whole square.

This is for piece number 1 plus 5 into 45 cube by 12 plus 45 into 5 into 27.5 minus 17.5. I
think the cg of this piece if you carefully look at this the cg of this piece will be from here,
this value is 45. So, 22.5 plus 5, 27.5 minus 27.5 the whole square this is for piece number 2.
So, this value is 9.45, 10 power 4 mm 4. Let us also find out 𝐼𝑦 again using parallel axis

theorem.

3 3
30×5 2 5×45 2 4 4
𝐼𝑍 = ⎡⎢ 12 + (30×5)(17. 5 − 2. 5) ⎤⎥ + ⎡⎢ 12 + (45×5)(27. 5 − 17. 5) ⎤⎥ = 9. 45×10 𝑚𝑚
⎣ ⎦1 ⎣ ⎦2

657
 3 3
5×30 2 45×5 2 4 4
𝐼𝑦 = ⎡⎢ 12 + (5×30)(15 − 7. 5) ⎤⎥ + ⎡⎢ 12 + (45×5)(7. 5 − 2. 5) ⎤⎥ = 2. 579×10 𝑚𝑚
⎣ ⎦1 ⎣ ⎦2

𝐼𝑦𝑧 = ∫ 𝑦𝑧 𝑑𝑎 = [(30×5)(17. 5 − 2. 5)(15 − 7. 5)]1 + [(45×5)(− (27. 5 − 17. 5))(− (7. 5 − 2. 5))]2

4 4
𝐼𝑦𝑧 = 2. 813×10 𝑚𝑚

So, which is very simple, let us do this 5, 30 cube by 12 plus 5 into 30 into 15 minus 7.5,
because this value is 7.5. This is for piece number 1 plus piece number 2, 45 into 4 cube by
12 plus 45 into 5 into 7.5 minus 2.5 the whole square for 2 which is also equal to now 2.579
into 10 power 4 mm power 4.

Let us compute also 𝐼𝑦𝑧, which will be yzdA for the entire area. Let us do it here. Which will

be equal to 30 into 5 that is for the first piece. So, which is going to be 17.5 minus 2.5,
because this is positive and z bar is also positive, which will be 15 minus 7.5. This is for
piece number 1. Similarly we can do for piece number 2, 45 into 5. And we are talking about
minus going to be minus of 27.5 minus 17.5, because this is going to be below the cg, into
minus of 7.5 minus 2.5.

I think we know very clearly how are we getting this minus. For example, the cg of this piece
is towards the negative direction of z. Therefore, there is a minus here. Similarly, the cg of
this is towards the negative direction. For example, towards negative direction of y therefore,
there is minus. So, if I compute this I get 𝐼𝑦𝑧 as 2.813 into 10 power 4 mm to the power 4.

658
(Refer Slide Time: 13:11)

2𝐼𝑦𝑧 2×2.813×10
4
𝑜
tan 𝑡𝑎𝑛 (2α) =− 𝐼𝑧−𝐼𝑦
=− 4 =− 0. 818
(9.45−2.579)×10

𝑜
α =− 19. 64

I get this value as minus 0.818 degrees. If you work out I get α 1 as minus 19.64 degrees.

𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦 4 4
𝐼𝑢 = 2
+ 2
cos 𝑐𝑜𝑠 2α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α = 10. 457×10 𝑚𝑚

𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦 4 4
𝐼𝑣 = 2
− 2
cos 𝑐𝑜𝑠 2α + 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α = 1. 575×10 𝑚𝑚

So, measured clockwise from positive z to positive u that is what we do. Let us mark that. So,
let me copy this figure. Let me copy this figure, let me put it here. Let me mark the u u axis
19 degrees so, this is my u this could be my v and this angle is α which is 19.64 degrees.

659
(Refer Slide Time: 15:46)

Now, once I have this with me. I can now find 𝐼𝑢 and 𝐼𝑣. 𝐼𝑢 is given by the equation
𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦
2
+ 2
cos 𝑐𝑜𝑠 2α − 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α . I substitute please check you will get this value

as 10.457 into 10 power 4 mm 4. We can also find 𝐼𝑣 from this equation

𝐼𝑧+𝐼𝑦 𝐼𝑧−𝐼𝑦
2
− 2
cos 𝑐𝑜𝑠 2α + 𝐼𝑦𝑧 sin 𝑠𝑖𝑛 2α , which is actually 1.575 into 10 power 4 mm 4. I

do not think we have difficulty in finding out these values by substitution, because α value is
here. We have 2 α and we have 𝐼𝑧𝐼𝑦 everything in the previous step. So, we can easily find

this.

660
(Refer Slide Time: 17:27)

Now, having said this let us compute the components of the moment. So, now, we want to
compute the components of the moment. We know the 𝑀𝑧 value from the equation you know

𝑀𝑧 value will be 2 into 2, which is 4 kNm. Which will cause bending this way. So, tension at

the bottom compresses the top.

So, now 𝑀𝑧 is 4 kilo newton meter. Now, we can say

𝑀𝑢 = 𝑀𝑧 cos 𝑐𝑜𝑠 α

𝑀𝑣 =− 𝑀𝑧 sin 𝑠𝑖𝑛 α

𝑀𝑧 = 4𝑘𝑁𝑚

𝑀𝑢 = 4 cos 𝑐𝑜𝑠 (− 19. 64) = 3. 77𝑘𝑁𝑚

𝑀𝑣 =− 4 sin 𝑠𝑖𝑛 (− 19. 64) = 1. 34𝑘𝑁𝑚

𝑀𝑢
σ(𝑢,𝑣) = 𝐼𝑢
(v) (Compression)

𝑀𝑣
σ(𝑢,𝑣) = 𝐼𝑣
(u) (Tension)

𝑣 𝑀𝑣 𝐼𝑢
tan 𝑡𝑎𝑛 β = 𝑢
= 𝑀𝑢 𝐼𝑣

661
 𝑜
β =− 67. 04

Fiends please note we are not substituting any value for these whether it is hogging or
sagging.

Because these equations are generated from a generic idea we can only interpret whether
stresses will be tensile or compressive based upon the nature of bending. Which we will do
later, it is an interpretation only. Now, we also know to compute the stresses. Stress at any
point x is given by minus 𝑀𝑢 by 𝐼𝑢 into v. I am using the same classical bending theory

equation, but I have resolved M and I n distance of fibres depending upon the principal axis. I
have modified them, plus 𝑀𝑣 by 𝐼𝑣 into u.

So, we say stress at the neutral axis should be 0. So, therefore, can I say stress as neutral axis,
which is equal to 𝑀𝑢 𝐼𝑢 into v plus 𝑀𝑣 𝐼𝑣 into u, which implies tan beta, where beta is the

angle of inclination of the neutral axis from the u u axis. Which can be simply given by v by
u which is 𝑀𝑣 by 𝑀𝑢 into 𝐼𝑢 by 𝐼𝑣. Now, we have all the values.

We have Mu, we have 𝑀𝑢, we have 𝑀𝑣, we have 𝐼𝑢 and 𝐼𝑣. So, can we find beta? Beta is

coming to be minus 67.04 degrees, which is now measured clockwise, because it is negative
from positive u axis to locate the neutral axis. So, let us locate the neutral axis? Let us take
this figure, let us copy this figure again.

(Refer Slide Time: 22:00)

662
Let us copy this figure again, let us draw it here. Let me locate the neutral axis in Blue colour
so, at 67 degrees. So, I should say very close here. This is my neutral axis. Measure 67
degrees further from u-u axis. So, we have mark the neutral axis. now we also know, if you
want to locate a point if you want to locate a point A and the point B, I need to measure the
distance perpendicular from the neutral axis.

Let us say the neutral axis is not passing through this line. So, let us say I want to measure
this distance to find the stresses is not it, but that is very tedious, because it is very difficult to
find out this component. So, instead we have resolved it very simply we can find the 𝑢𝐴 and

𝑣𝐴 component of the point A, 𝑢𝐴 we said it is z cos α plus y sin α we derived it already.

Similarly, this is minus z sin α plus y cos α. Let us first fix the y z points for the corner A to
do that we can easily look at the figure. The y z point will be plus 17.5 and minus 7.5 is it for
the point A. Similarly for the point B, if you want to find y z it should be minus of 15 minus
17.5.

𝑢𝐴 = 𝑧 cos 𝑐𝑜𝑠 α + 𝑦 sin 𝑠𝑖𝑛 α

𝑣𝐴 =− 𝑧 sin 𝑠𝑖𝑛 α + 𝑦 cos 𝑐𝑜𝑠 α

𝐴(𝑦, 𝑧) = (+ 17. 5, − 7. 5)

𝐵(𝑦, 𝑧) = (− (50 − 17. 5), (− 2. 5))

𝑀𝑢 𝑀𝑣
σ𝐴 =− 𝐼𝑢 (𝑣𝐴) + 𝐼 (𝑢𝐴)
𝑣

𝑀𝑢 𝑀𝑣
σ𝐵 =− 𝐼𝑢 (𝑣𝐵) + 𝐼 (𝑢𝐵)
𝑣

This is for y and z straight away it is minus 2.5. We are looking for actually the value of this,
because we know this distance is 7.5 and this is 5. So, I can get this value as minus 2.5. Since,
I know the coordinates y z for the points A and B. I can easily use the above equation, which
is going to be; if you want to find stress at A, I will say minus Mu by 𝐼𝑢 into 𝑣𝐴 plus 𝑀𝑣 𝐼𝑣

into 𝑢𝐴 whereas, α is known to me, I can find 𝑢𝐴 and 𝑣𝐴.

663
Similarly, I can find stress at B minus 𝑀𝑢 by 𝐼𝑢 into va plus 𝑀𝑣 𝐼𝑣 into 𝑢𝐵. We can easily do

this, Now, I want to run this using MATLAB. Let us run MATLAB to compute this. I will
hold the screen we will run the MATLAB now.

(Refer Slide Time: 25:57)

(Refer Slide Time: 26:26)

So, friends what you see on the screen is the MATLAB program to compute unsymmetrical
building l section that is what we are doing. So, let us enter the values of y y as 2.579 𝐼𝑧 and

𝐼𝑦𝑧, which are all taken from the lecture and 𝑀𝑧 is 4. We already have in kilo newton meter.

664
So, now, it is calculating 𝐼𝑢 and 𝐼𝑣 automatically. Then it also calculates the α angle. It

calculates the α angle.

(Refer Slide Time: 26:37)

(Refer Slide Time: 26:55)

Then it computes 𝑀𝑢 and 𝑀𝑣 then we have to enter the coordinates for the point A and

coordinates for the point B. Please see that. Once we enter the program runs. I will get the
stress at point A and point B. Let us run this program. So, the program is run.

665
(Refer Slide Time: 27:07)

(Refer Slide Time: 27:19)

I have the values; I am copying all these values into my screen here and putting them here.
All these values are copied here. Let me also copy this figure and put it here. Let me move
this figure here friends. So, let us check the values. We said 𝑀𝑧 for me problem is 4, yeah I

got 4. Α angle we got is 19.64 program is giving 19.65. 𝐼𝑢, 𝐼𝑣 we have. 𝑀𝑢, 𝑀𝑣 we have. Beta

angle we computed 64.92 degree, it is 67. So, this is the small difference, because of the
round off error in 𝐼𝑢 and 𝐼𝑣.

666
So, beta we have. So, we have already values of 𝑢𝐴 and 𝑣𝐴, 𝑢𝐵 and 𝑣𝐵. Now, the stress at the

point A is minus 1520.9 and stress as point B is plus 1818.54. Friends, we already said it is a
cantilever with an upward. So, the beam is bending like this. This is compression and tension.
So, the top point is receiving compression and the bottom point is receiving tension.

So, friends we have used the MATLAB program effectively to solve this problem and we
have the answers as we see on the screen. So, and this is what we have interpreted we have
these values with us. So, this is what you can run the program and try to show.

(Refer Slide Time: 30:00)

Now, let us start discussing something on shear center. So, we will does something about
shear center to understand shear center. Let us first see the internal force distribution in a
given section. Let us take for example, an I section. And a channel section whatever may be
the dimension.

So, for both these, we say this is my z axis this becomes my y axis and of course, x is normal
to the screen. If I apply a force V upward, let us see the force distribution of this this is how
the internal forces will get distributed, which I call the net force as F1 and F1 F1 and F1. On
the other hand if you look at the channel section and say this is my centroidal axis and apply
a force m bar V here. The force distribution in this case is what I am marking on the screen
now.

667
So, we call this as F1 and this as F1. So, friends in both these figures we can say F1 is the
resultant force of the shear stress due to the applied force v. There will be no net force along z
axis the section will not rotate in this case. As there is no net force along z axis the section
will not rotate. We also know that the vertical force is taken care of by the web.

The vertical force is taken care of by the web, but the flange forces F1 generates a couple.
The horizontal force F1, which is acting in the web, generates a couple correct. This couple is
generated about a point in the cross section that point is called as shear center.

(Refer Slide Time: 34:17)

Now, let us take a cantilever beam. Let me draw the flange thickness. Let us say the section is
a channel subjected to some load v. The cantilever will start deflecting under this load, where
the cross section will also have this effect. So, what we say here is? Cross section gets rotated
or we can say gets twisted. Now, to avoid this load should be applied through the shear
center. Therefore, friends shear center is the intersection of loading plane and the bending
axis.

668
(Refer Slide Time: 36:46)

Now, interestingly let us enlarge this figure and see what happens to my channel section? So,
I have a channel section drawn here again. Let us say this is my length of the member, it is
got 3 axis now let me mark those axis. This is my vertical axis y my horizontal axis z this is
my another axis, which is x. Let us say it is a thinned section the depth of the section is
measured from center to center of the web call that as h.

We call this force, the internal force, the horizontal internal force Fz and Fz. Let us say this is
subjected to a vertical force b. Now, looking at this figure we can say fz into h by 2 of 2,
which is nothing but Fz into h is the moment responsible to cause rotation of the cross
section. This will happen about which point, let us extend this. This lapped amount of point c,
lapped amount of point c, at a distance e from the centroid so, this will happen or this will
occur about c causing twisting of the section.

Now to balance this if P or v is applied at c then imagine that p or v is applied here then we
can say P into e will balance this moment. So, section will not rotate. So, in the whole context
of discussion friends the point c is called as shear center. So, it is ideal that the load should
pass through the shear center to avoid twisting of the cross section. Having said this let us
also say in the same thing whereas, e is the offset of the shear center from the Cg of the
section e is offset.

669
(Refer Slide Time: 41:20)

Therefore friends, shear center is the intersection of loading plane with the bending plane in
my case this is x z. Further shear center is a specific location of the cross section. Load if pass
through the shear center will not cause any twisting moment to the section. So, therefore, to
avoid twisting of the cross section load should be applied or made to pass through the shear
center.

(Refer Slide Time: 43:02)

friends, in this lecture we learnt one more example of unsymmetric bending problem we have
used effectively the MATLAB program to find the stresses at the designated points. We have

670
also started learning about the importance of identifying a shear center and how if load is
applied to the shear center can avoid twisting of the cross section. We will discuss more in
detail about the shear center and more numerical examples of various shape of cross sections
and find shear center using the conventional equations in the coming lectures.

Thank you very much and have a good day, bye.

671
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 39
Shear center - 2

(Refer Slide Time: 00:23)

Friends, welcome to the lecture-39 on Advanced Steel Design course. We are discussing
about the Shear center. So, we will call this lecture as Shear center - 2 because in the last
lecture we discussed about the conditions under which twisting of cross-section should
happen and we stated clearly that if the load is applied or pass through the shear center you
can avoid twisting of the cross section.

Therefore, friend shear center is a geometric point in the cross-section which is actually the
intersection of the loading plane with the bending plane, this is what we said.

672
(Refer Slide Time: 01:19)

Furthermore, if I have an angle section as shown in the screen now we already know that the
internal force distribution will be as 𝐹2 and 𝐹1. If you apply a force V and 𝐹1 and 𝐹2

intersect at the point A. So, one can say that 𝐹1, 𝐹2 intersects at the point A, if the load is

applied or the line of action of load passes through the point A, then one can avoid twisting of
the cross-section.

(Refer Slide Time: 03:16)

673
We already said that shear center is the intersection of the bending plane and the loading
plane. We can also define shear centre in a more geometric sense. Let us say I have a
cross-section of some shape this is my point of action w and this is the shear center offset e
from the Cg and we have applied a load V. Therefore friends, shear center is a point of
intersection of the longitudinal axis of the member that is this axis and the line of action of
the transverse load which is point C.

So, therefore friends, shear center is a geometric point whereas, mass center is identified as C
g. If the load does not pass through the shear center, then the section will be subjected to
twisting moment. So, we can say straight away here for e equals 0; no twisting moment.

(Refer Slide Time: 05:38)

Now, let us quickly see where are shear center very important in what kind of structures. Let
us take for example, marine structures. We already know marine structures are strategic
structures constructed for specific functions. Usually, they are constructed or I should say
fabricated because we use essentially steel as a material using thin-walled elements and
sections are also chosen to be asymmetric.

Therefore, under such conditions the cross-section will fail in torsion because it will be weak
in torsion before it fails in bending. So, this twisting will induce a premature failure of the
members. So, that is a very important status. So, therefore, shear center or location of shear
center are important in such kind of structures where they are thin-walled elements and the
sections are as symmetric in nature.

674
Furthermore, this will also cause warping which will induce geometric instability and which
is one of the most undesirable characteristics in design.

(Refer Slide Time: 08:34)

Friends, shear center location is made comfortable by a simple logic. Shear center will
coincide or will lie on the axis of symmetry. If a section has two axis of symmetry, then their
intersection will be the shear center. So, sometimes in most of the sections shear center can be
located by simple observation. We can show some examples. Let us say rectangular section
where there are two acts of symmetry.

So, this becomes my shear center, here tubular section which is commonly used in marine
structures where there are at least 2x of symmetry that becomes the shear center. If I have a
T-section which has one axis of symmetry, shear center will lie on this axis.

675
(Refer Slide Time: 10:41)

Now, the question comes where are asymmetric sections popular? You must have seen in
crane rail. Crane rail has a section like this. This becomes my crane rail. This will have a
cushion effect here and this is what we call as a slip gap, now this is the free board. So, if the
rail jumps of the crane rail bends still the wheel should not touch the top. This is a typical
cross-section of a crane rail. This is a classical example of an asymmetric section.

So, in heavy industrial structures friends, you will see that asymmetric sections are very
commonly used as load carrying elements which are also crucial elements in the whole setup.

(Refer Slide Time: 12:20)

676
Having understood this, let us try to derive the basic equation for finding the shear stress. Let
us say we have a flange of a channel section. Now, depth d and breadth b and thickness t we
know that this section has an axis of symmetry. Shear center has to lie on this axis. Let us
imagine this along its length at a distance x, let us cut the strip let us say at a distance x let us
cut a strip.

𝑏
𝑀
𝐻 =∫ 𝐼
𝑦 𝑑𝑎
𝑥

𝑏
𝑀+𝑑𝑀
𝐻 + 𝑑𝐻 = ∫ 𝐼
𝑦 𝑑𝑎
𝑥

So, I am drawing that strip in enlarged view here. Let us say this is my plane of reference
thickness be t. Let it be cut at a distance x and a strip is being cut of thickness dx. An area of
this hatched portion is da, and let us say we have an axis which is measured as b distance
from there and along this axis we have H and we have M acting in the cross-section.

Now, I want to find the shear stress in the flange. Now, H will be given by M by I into y of dx
sorry da x to b equation number 1. H plus dH will be M plus dM by I into y da which will
also run from x to b equation, number 2 equations from first principles.

(Refer Slide Time: 16:51)

𝑏 𝑏
𝑀+𝑑𝑀 𝑀
𝑑𝐻 = ∫ 𝐼
𝑦 𝑑𝐴 − ∫ 𝐼
𝑦 𝑑𝐴
𝑥 𝑥

677
 𝑏
𝑑𝑀
𝑑𝐻 = 𝐼
∫ 𝑦 𝑑𝐴
𝑥

Now, I want to find the unbalanced force along the length is given by equation 2 minus
equation 1 which will be dH which will be x to b M plus d M by I y dA minus integral x to b
M by I into y dA. We calls equation number 3. Now, I can say it is dM by I into x b y dA
equation number 3 b we call this 3 a.

Now, for equilibrium, this unbalance force as given by equation 3 b should be equal to the
elemental shear which is opposing this force.

(Refer Slide Time: 18:51)

𝑏
𝑑𝑀
τ(𝑑𝑧 𝑡) = 𝐼
∫ 𝑦 𝑑𝑎
𝑥

𝑏
𝑑𝑀 1
τ = 𝑑𝑍 𝐼𝑡
∫ 𝑦 𝑑𝑎
𝑥

𝑑𝑀 1
τ = 𝑑𝑍 𝐼𝑡
𝑎𝑦

𝑉 𝑎𝑦
τ = 𝐼𝑡

Let tau be the shear stress. Then, the elemental shear force will be tau times of t dz. So, now
for equilibrium this should be equated dM by I integral x to b y da. So, now, I can say the
shear stress is going to be dm by dz 1 by I t integral x to b y da which can be dm by dz 1 by It

678
a y bar because y da integral is a y bar whereas, tau is now going to be V a y bar by I t, where
V is a total shear at the section.

ay bar is a static moment of area; I is the moment of inertia of the whole section and t is the
width of the section perpendicular to shear. Having said this, let us try to do a problem to
locate the shear center for a section.

(Refer Slide Time: 21:17)

So, we will say example one locate the shear center for the given cross-section. Let us draw
the cross-section. Let us mark the geometric dimensions of this. Let us mark the centerline of
this, call this dimension as h, this thickness 𝑡1 and this thickness 𝑡2 and this breadth as 𝑏2 and

this breadth as 𝑏1.

We also know that this line is automatically the axis of symmetry. Let us assume that the
shear center is located here, where the external load V is acting to avoid twisting of this
cross-section. We call this distance of offset as 𝑒1 and 𝑒2. So, our job is to find either 𝑒1 or 𝑒2

because they are connected by relationship 𝑒1 plus 𝑒2 will be equal to h. Let us take this

thickness of the section as t.

So, this is my 1st element, this is my 2nd element and this is my 3rd element where the
intersection is cut here. Let us say the proportional shear shared by the flanges 1 and 2 are
indicators. By the way we know that they are flanges and this is called as a web or indicated

679
as 𝑉1 and 𝑉2 respectively. So, we assume that the total shear V is counterbalanced by these

two shear.

So, we are neglecting shear accounted by the element 3. So, now, let us cut a section at a
distance y of thickness dy and area shaded is da and this is the part above the section under
consideration. This part has a C g let us say that is y bar from here from the axis of symmetry
or from the line where the shear center is being located.

Now, tau is V A y bar by It. This known equation we just now derived, let us put it a small a.
So, the small a for the component 1, this is for element 1, component a is going to be 𝑏1 by 2

minus y that is the hatched portion in neon color 𝑏1 by 2 minus y into 𝑡1. Let us say y bar is y

plus 𝑏1 by 2 minus y half of that which will be y minus y by 2 plus b 1 by 4. So, I can now

say y bar is y by 2 plus 𝑏1 by 4 because we need y bar here we need a here.

680
(Refer Slide Time: 28:31)

Now, I can substitute tau of number 1 is V ay bar by I t which is V by I 𝑡1; ay bar is just now

we computed which is 𝑏1 by 2 minus y of 𝑡1 and y bar is y by 2 plus 𝑏1 by 4. Let us simplify

this tau of number 1 will be V by I 𝑡1 of 𝑡1 𝑏1 by 2 minus y half 𝑏1 by 2 plus y, can I write

like this?

Because I get a product here a minus b and a plus b so, which can be V by 2I as 𝑡1 goes away

2 I of 𝑏1 by 2 𝑏1 by 2 a square minus b square equation 3. This is my tau of the first number.

So, therefore, I can write 𝑉1 is integration of this of the whole area varying from minus 𝑏1 by

2 to plus 𝑏1 by 2, see this figure minus 𝑏1 by 2 to plus 𝑏1 by 2 that is from this section here

from this section here.

So, that is the validity of the shear force 𝑉1 which is shared by the flange 1. Let us substitute

and find out this value which will be 2 times of 0 to 𝑏1 by 2 tau of 1 d a because of symmetry.

So, which will be 2 times of integration 0 to 𝑏1 by 2 tau 1 we already have.

So, V by 2I 𝑏1 by 2 the whole square minus y square of dy of 𝑡1, That is the strip area which

is integrated from minus to plus limits of this, That is da.

681
(Refer Slide Time: 31:40)

So, that is what I am having here. So, I replace this as 𝐼1. So, I can now say 𝑉1 is V by I of 𝐼1

equation 4.

682
(Refer Slide Time: 34:18)

𝑉1 =
𝑉
𝐼 (𝐼1)
3
𝑡1𝑏1
𝐼1 = 12

𝑉2 =
𝑉
𝐼 (𝐼2)
3
𝑡2𝑏2
𝐼2 = 12

𝑉 = 𝑉1 + 𝑉2

𝑉𝐼1 𝑉𝐼2
𝑉 = 𝐼
+ 𝐼

By the same logic I can apply this to piece number 2 or flange number 2 I will get by the
similar logic 𝑉2 will be V by I of 𝐼2 equation 5. Now, we have two equations 𝑉1 is V by 𝐼1 𝑉2

is V by 𝐼2, where 𝐼1 is 𝑡1 𝑏1 cube by 12 and 𝐼2 is 𝑡2 𝑏2 cube by 12. We also know total shear is

sum of these two, shear taken by member 3 is neglected we already stated that. So, the total
force shear will be V 𝐼1 by I plus V 𝐼2 by I.

So, that is the control equation we have with us. Now, let us take moment about the point C.
See this figure, I will copy this figure.

683
(Refer Slide Time: 36:12)

𝑉1𝑒1 = 𝑉2𝑒2

𝑉 ( )
𝐼1
𝐼
𝑒1 = 𝑉 ( )𝑒
𝐼2
𝐼 2

𝐼1𝑒1 = 𝐼2𝑒2

𝑒1 + 𝑒2 = ℎ

Now, let us say we will take moment about shear center C. So, 𝑉1 into 𝑒1 which is clockwise

this is 𝑉1 this is 𝑉2 should be equal 𝑉2 into 𝑒2 which is anti-clockwise. So, what is 𝑉1? Please

see from this equation which is V 𝐼1 by I of 𝑒1 and 𝑉2 see from this equation 𝐼2 by I, V 𝐼2 by I

of 𝑒2.

So, from this relationship we get a concept 𝐼1 into 𝑒1 is 𝐼2 into 𝑒2 we get one relationship

from here. We call equation number 6 here. Also, we know that from the geometry 𝑒1 plus 𝑒2

is h, see here the solving equation 6 and 7 because we know 𝐼1 and 𝐼2 we already have them

here see 𝐼1 is known, 𝐼2 is known. So, I already have 𝐼1 known, 𝐼2 known.

So, there are two unknowns 𝑒1 and 𝑒2, I have two equations. So, by solving equations 6 and 7

we get 𝑒1 and 𝑒2 and I can locate the shear center. That is a very easy approach from first

684
principles. We found out the shear center very easily. Yes friends, it is a very easy illustration
for applying the fundamental logic to obtain the shear stresses. Let us apply on to a problem
and see this quickly. So, we will say the same dimension, I will just copy this figure again.

(Refer Slide Time: 39:09)

𝑒1 𝐼2
𝑒2
= 𝐼1

3
10×50 5 4
𝐼1 = 12
= 1. 04×10 𝑚𝑚

3
20×100 6 4
𝐼2 = 12
= 1. 67 × 10 𝑚𝑚

𝐼 = 𝐼1 + 𝐼2

𝑒1 1.67×10
6

𝑒2
= 5 = 16. 058
1.04×10

Let me give these dimensions as 𝑏2 is 100. I write in a different color maybe 100 which is 𝑏2;

𝑏1 is 50; 𝑡1 is 10; 𝑡2 is 20 and h is 80 all dimension millimeters. So, we have an equation 𝑒1

by 𝑒2 is 𝐼2 by 𝐼1 can you have this equation here. 𝑒1 by 𝑒2 is 𝐼2 by 𝐼1. So, let us compute I 1

for this problem which is going to be 10 into 50 cube by 12 which is 1.04 10 power 5.

Let us compute 𝐼2, this is the second member, this is the first member 20 into 100 cube by 12

which is 1.67 10 power 6. We say the total moment of inertia I is 𝐼1 plus 𝐼2. Let us neglect the

685
I from member 3 this is member 3. So, let us say 𝑒1 by 𝑒2 is 1.67 10 power 6 by 1.04 10

power 5 which is 16.058, equation let us say 8.

(Refer Slide Time: 41:45)

𝑒1 + 𝑒2 = ℎ = (80 + 5 + 10) = 95

𝑒2 = 5. 569 𝑚𝑚

𝑒1 = 89. 431 𝑚𝑚

𝑒1 plus 𝑒2 is h which is, we will say this is 80 in this figure this is anyway h, but we have this

dimension. This dimension the clear dimension is 80. So, h will be now equal to 80 plus 5
plus 10 which is 95 equation 9. So, by solving these two equations I get e 1 as 5.569
millimeters and sorry, 𝑒2 as and 𝑒1 as 89.431 millimeters. This is a simple illustration what

we had to solve this problem.

686
(Refer Slide Time: 42:33)

Friends, in this lecture we learned more about the concepts of shear center, we also did a
numerical example. We derived the equation for shear stress, then we did a numerical
example of an I section to locate the shear center. We did this using first principles and the
derivations are very clear therefore, there is no question or any confusions.

So, friends we will do couple of more examples in the coming lectures to illustrate about the
calculation of shear center more in detail. We will also take up shear center for
unsymmetrical sections. We will try to elaborate this more on coming lectures.

Thank you very much and have a good, bye.

687
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 40
Shear center - 3

(Refer Slide Time: 00:21)

Friends welcome to lecture-40 of the course Advanced Steel Design. We are working on
estimating more examples on Shear center. I will call Shear center lecture-3. In this lecture
we will learn more example problems of locating shear center.

688
(Refer Slide Time: 00:51)

So, example 2. We will take up a channel section. We know the shear flow this is going to be.
So, this is I calls 𝑉1, I call this as 𝑉2 and we also know that this section has an axis of

symmetry.

So, the shear center will lie on this axis let us say somewhere here. Let us assume that the
load is applied at the shear center to avoid twisting of the cross section. We call the
eccentricity or offset of the shear center from the phase of the flange or e. Let us say the
depth of the web is d and the total breadth of the flange is b. Let us say the section has a
uniform thickness t throughout.

So, a thinned section and we already know the green line represents axis of symmetry. Shear
center will have to lie this is the shear center. We will have to lie along this line of symmetry.
Let us take up element 1 this is my element 1, and draw an enlarge view of this.

Let us cut a section at a distance x measured from the tip and this is my da. Let us call this
𝑑
thickness of the strip as dx and the c g of this da from the shear center is 2
from the figure,
𝑑
2
from the figure. And we know that this is b breadth of the flange. Let us copy this figure

and move to the next page.

689
(Refer Slide Time: 04:38)

So, now let us write 𝑉1 will be

So, I can now write 𝑉1 as

690
Now we can also say by symmetry is equal to 𝑉2, is not it from the figure? Further look at

this figure I will copy this figure again.

(Refer Slide Time: 07:19)

We can now say taking moment about c

𝑑 𝑑
𝑉𝑒 = 𝑉1 2
+ 𝑉2 2

2
𝑉𝑡𝑏 𝑑
𝑉1 = 4𝐼

2
𝑉𝑡𝑏 𝑑
𝑉2 = 4𝐼

( )
2
𝑉𝑡𝑏 𝑑 𝑑
𝑉𝑒 = 4𝐼 2

2 2 2 2
2𝑡𝑏 𝑑 𝑡𝑏 𝑑
𝑒 = 8𝐼
= 4𝐼

So, for a given section if I know these values let us say for example, b is 40 and d is 100 and t
is 10.

So, I can find quickly I for this which will be

3
40×10 2
𝐼 = ⎡⎢ 12 + 40×10×45 ⎤⎥×2 + 𝐼𝑤𝑒𝑏
⎣ ⎦
3
10×80
𝐼𝑤𝑒𝑏 = 12

691
So, I get I now. So, now, in this equation I know t, t is 10 mm. I know b, b is 40 mm. I know
d, d is 90 and I know I which is here. So, I can find e. And e will be the offset of the shear
center from the cg. So, friends we have a very easy example these examples were easy
because they had at least one axis of symmetry. So, the shear center was located in that axis
of symmetry.

So, it is very simple that if you have a problem where the section has got a symmetric axis,
we can easily locate the shear center. Let us take up one more example which is again simple,
but slightly tricky in terms of its equation. Let us try to do that. We will take another example.

(Refer Slide Time: 12:31)

We will call this example 3. We say it is a very common section being used in marine
structures a thin section having uniform thickness. So, let us say the dimensions are marked
as shown in the figure this is b.

Let us say this dimension is 𝑏1, the overall depth is h all centroid center, uniform thickness t

throughout, this section has an axis of symmetry. So, from this point let us mark the shear
center somewhere here and we call this offset as e. So, let us say the 𝑉𝑅 and V e are acting

here to avoid twisting of the cross section.

So, now let us mark the shear flow. We call these forces. So, this is 𝑉1 and this is 𝑉2 and this

is 𝑉3 this is 𝑉4 and this is marked as 𝑉5. Let us now start computing these internal shears. So,

692
I want to compute 𝑉1. So, I am drawing an enlarged view of this. Let me cut a section of

thickness dz at a distance z from here and I am marking the centroidal axis or the axis of
symmetry from here. And we know this distance is h by 2 from here, and this was 𝑏1.

So, I have expression for 𝑉1 which will be

Let us try to find 𝑉5. So, let me write this.

(Refer Slide Time: 19:53)

Which is 𝑉1 which is identically same as 𝑉5 because equation number 1.

Now, let us do it for 𝑉2 we will draw a separate figure. This is my axis of symmetry. Let us

mark these dimensions. So, thin section. So, I am marking all the dimension centroid. This is
𝑏1, uniform thickness t, this is b and we know this distance is h by 2. Let me cut a strip at a

distance x of thickness dx from here. We will also divide this into two parts.

693
So, now I want to find 𝑉2. Let us find out what is a y bar of piece 2.

(Refer Slide Time: 25:35)

𝑏1 2⎤
3
⎡ 𝑡𝑏
( ) ℎ 2
3 3
𝐼𝑒𝑛𝑡𝑖𝑟𝑒 𝑠𝑒𝑐𝑡𝑖𝑜𝑛 =
𝑡ℎ
12
+ ⎢ 121 + 𝑡𝑏1
⎢
⎣
ℎ
2
− 2
⎥×2 + ⎡⎢ 𝑏𝑡
⎥
⎦ ⎣ 12 ( ) ⎤⎥⎦
+ 𝑏𝑡 2

So, we have I here now let us take moment about we call this point as O. So,

( )
𝑉𝑒 = 𝑉1𝑏 2 + 𝑉2 ( )2 ℎ
2

We neglect shear 𝑉3.

694
So, now V is a given data this is a shear acting force acting on the section. b h or geometric
data which is known look at this equation. If I know V if I know V if I know t and if I can
compute I. I can compute 𝑉2 and 𝑉4. Similarly, if I know t if I know b 1 and h I can compute

𝑉1 and 𝑉5. So, in this equation 𝑒1 𝑉1 𝑉2 are computed.

So, I know 𝑉1 𝑉2 also. So, I can easily find it which is the offset of the shear center from the

point O. So, it is very simple. So, we can take a quick example.

(Refer Slide Time: 29:37)

Let us say let b be 50, b 1 be 20 or 30, h be 120 and t be 5 mm. All dimensions are given.

So, I know all dimensions can you compute e? Therefore, locate the shear center. It is a very
straightforward problem there is no issue in this. Let us do one more problem.

695
(Refer Slide Time: 30:46)

Which is example 2, 3 we will do example 4. So, the cross section is drawn like this. Let me
mark. Let us say this dimension is b which is centroid center.

Let us take this dimension as b 1 and this dimension as h. Let us say it is having uniform
thickness t throughout. So, we know that this section has one axis of symmetry, this
intersection is marked as O and we mark this as my shear center, we mark this x and t as e
from here. Let us assume that the shear and the reaction forces are acting at this point to
avoid twisting of the cross section.

Let us mark the shear flow now as shown the figure. Say this is 𝑉1, this is 𝑉2, this is 𝑉3 and

this is 𝑉4 and this is 𝑉5. Now to locate 𝑉1. We know it is going to vary say this is 𝑉1. So,

𝑏1
𝑉𝑎𝑦
𝑉1 = ∫ 𝐼𝑡
𝑑𝑎
0

𝑎𝑟𝑒𝑎 = 𝑡𝑧
𝑑𝑎 = 𝑡𝑧

So, let us say area we will consider a strip marked here I will draw this figure separately. So,
we will consider a strip at a distance z from here for a thickness dz.

And I want to know the cg of this from here.

696
 ℎ 𝑧
𝑦 = 2
+ 𝑏1 − 2

𝑏1

𝑉1 = ∫
0
𝑉
𝐼𝑡
(𝑡𝑧) ( ℎ
2
+ 𝑏1 −
𝑧
2 )𝑡𝑑𝑧
()
2 2 3
𝑉 ⎡ℎ 𝑏1 𝑏1𝑏1 𝑏1 ⎤
𝑉1 = 𝐼𝑡 ⎢2 2
+ 2
− 6 ⎥
⎣ ⎦
2 3
𝑉 ⎡ 𝑏1ℎ 𝑏1 ⎤
𝑉1 = 𝐼𝑡 ⎢ 4 + 3 ⎥ ≡ 𝑉5
⎣ ⎦

We will take equation number 1. Identically friends this will also be equal to 𝑉5. Now let us

take the next part that is 𝑉2, I will draw that separately.

(Refer Slide Time: 37:57)

This may line of symmetry. We know this is 𝑏1. I am marking the center line. We know this is

𝑏1 and the thickness is t uniform. I will cut the strip here at a distance x from here and let the

thickness be dx.

Let us hatch this portion and I want to find the cg of this. That is what we want to find. So, let
us divide this into two parts and for our simplicity let us find 𝑉2. So, we know 𝑉2 is going to

integrate you know this dimension is b.

697
 𝑏
𝑉𝑎𝑦
𝑉2 = ∫ 𝐼𝑡
𝑑𝑎
0

( )( ℎ2 )⎤⎥⎦ + ⎡⎣𝑡𝑥(
𝑏1
(𝑎𝑦) = ⎡⎢ 𝑏1𝑡
⎣
+ 2
1
ℎ
2 )⎤⎦ 2

𝑑𝑎 = 𝑡 𝑑𝑥

( )
𝑏 𝑏1ℎ
2
𝑏1
𝑉 𝑥ℎ
𝑉2 = ∫ 𝐼𝑡
𝑡 2
+ 2
+ 2
𝑡 𝑑𝑥
0
2
𝑉 ⎡ 𝑏1ℎ𝑏 𝑏1𝑏 ℎ𝑏
2
⎤
𝑉2 = 𝐼𝑡 ⎢ 2 + 2
+ 4 ⎥
⎣ ⎦

Now I want to find the moment of inertia of the whole section let us copy this figure.

(Refer Slide Time: 41:59)

So, we want to find the moment of inertia of the whole cross section about axis of symmetry.
So, let us do that.

3
⎡ 𝑡𝑏
( 𝑏1
) ℎ 2
3 3
⎤
𝐼=
𝑡ℎ
12
+ ⎢ 121 + 𝑡𝑏1
⎣
ℎ
2
+ 2 ⎥
⎦
×2 + ⎡ 𝑏𝑡 + 𝑏𝑡
⎢ 12
⎣
( ) ⎤⎥⎦×2
2

So, I get I. Now let us take moment about O.

𝑉𝑒 = 𝑉2 ( )×2 − (𝑉 𝑏)×2
ℎ
2 1

698
So, friends 𝑉2 is a function of I the geometric property. 𝑉1 is a function of I geometric

property. So, in this equation 𝑉2 𝑉1 will have V. So, V goes away and for all the dimensions

known one can compute e.

(Refer Slide Time: 44:25)

So, friends we have learnt more examples of working shear center of symmetric sections. We
have understood the use or the importance of shear center in the design. So, in the next
lecture we will start talking about shear center of curved beams. Suppose if we have a section
which is curved how do we locate the shear center? That is what we will focus on in the next
lecture.

Thank you very much friends have a good day bye.

699
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 41
Curved Section

(Refer Slide Time: 00:20)

Friends, welcome to the lecture 41 of Advanced Marine Structures. In this lecture, we are
going to learn example on Curved Sections to Locate the Shear Center. So, we already
studied locating shear center from the first principles simple exercise provided a section has
got at least one axis of symmetry. If axis of symmetry does not exist, then the section will be
subject to unsymmetric bending. Then, we have got to combine both the problems of
unsymmetric welding, locate the u u v v axis, then find the shear center. I will come to that
problem slightly later. We will do that.

So, in the previous examples, we discussed about rectilinear sections, like I section, channels,
t, etcetera. In this example, it is quite interesting we will take a curved section because we are
going to slowly understand the analysis and design of curved beams and crane hooks. So, we
need to at least examine the curved section for shear center calculations. Let us do that. We
will take I will call this example 1 for this lecture. We will talk about curved sections.

700
(Refer Slide Time: 01:56)

So, now, let us say, I have an axis of symmetry of the section. We will mark a curved section
symmetric about this axis. So, let us say it has got a center line, and that center line eta at a
distance R, from here and I call this point as O, ok, this point as O. Let us say it subtends an
angle 2 α symmetric about the center, so this angle is α. Let us believe the section has got
uniform thickness t and the curvature start of this section from the axis of rotation is taken as
β.

So, now; obviously, being this the axis of symmetry shear center will lie on this axis. Let us
mark the shear center somewhere here. If the force and the reaction pass through the shear
center the section will not be subjected in distinct moment. We call this distance of the shear
center from the axis of curvature as e. So, the question is now I want to find e. The given data
for the section are the following, radius of the section, thickness of the section, the included
angle and the subtended angle. Let me write it clear; are known thing.

So, let us now cut the section which is subtending an angle d𝜃 at𝜃 from here. Now, let us say
this is my hatched portion. Now, the shear stress tau is given /the classical equation V /It
integral y dA. Now, my question of integration is going to be, for the element is going to be
varying from β to𝜃. That is my control, is it not? That is my dA. Look at here, this is my dA,
right. That is what I am looking at.

So, let me do that which will be equal to V /It, integral β to𝜃, y will be R cos 𝜃 and R t𝜃 d𝜃, t
d𝜃, R d𝜃 t will be the area of the hatched portion which is green in color. This is y bar. So, let

701
us we can do this. So, the elemental shear stress tau is equal to VR square /I, sin𝜃 - sin β, am
I right. That is my elemental shear stress. Also, friends, we have an interesting relationship
here. From this figure we can say𝜋/2 - β is α. Can I say that? From this figure, from the
geometry this is true.

(Refer Slide Time: 08:16)

Having said this, so we said that tau is VR square /I sin𝜃 - sin β. Now, I can find the
elemental shear force, which is dv, which will be actually tau da. So, now, I can find dv as
VR square /I, sin𝜃 - sin β into R t d𝜃. So, now, this becomes VR cube t /I, sin𝜃 - sin β of d𝜃.
This is my dv. Now, the moment of this shear force about the center is given by dM. dM will
be straight away into R, is it not. The distance is R radius.

So, I can say VR cube t /I, sin𝜃 - sin β d𝜃 multiplied by R. So, this will become VR 4 t /I,
sin𝜃 - sin β into d𝜃. That is my dM. So, once we obtain the elemental moment, we can find
the total moment, yeah which will be integral of dM varying from β to𝜋/2, 2 times of that.

(Refer Slide Time: 10:42)

702
So, the total moment M is twice of integral β 2𝜋/2 dM which will be twice of integral β 2𝜋/2
VR 4 t /I, sin𝜃 - sin β d𝜃. So, which will give me M as 2 VR 4 t /I of - cos𝜃 -𝜃 sin β, apply
the limits from β to𝜋/2. So, that will give me M as 2 VR 4 t /I.

(Refer Slide Time: 12:00)

Before that we also know𝜋/2 - β is α. Can you see this here? Let me use that relationship
here. So, now, saying that M will now become 2 VR 4 t /I - cos𝜋/2 + cos β -𝜋/2 sin β + β sin
β. So, we can say this is 0. And we have cos β sin β, right. Now, I can simplify this equation
as 2 VR 4 t /I. I am replacing β as𝜋/2 - α.

703
So, can I write this as cos𝜋/2 - α - 𝜋 /2 sin𝜋/2 - α +𝜋/2 - α of sin𝜋/2 - α. I am bringing all in
terms of α. Let me copy this equation to the next screen.

(Refer Slide Time: 14:08)

So, M was equal to this which now becomes 2 VR 4 t /I, sin α -𝜋/2 cos α +𝜋/2 - α cos α.
Which is further simplified as 2 VR 4 t /I, sin α -𝜋/2 cos α +𝜋/2 cos α - α cos α. So, this goes
away; which will now land up as 2 VR 4 t /I sin α - α cos α. We call this equation number 1.
It is an interesting equation which is a classical theory available in the literature of a curved
sections. Now, let us take moment about O. This is the figure we have. Let us copy this
figure.

(Refer Slide Time: 15:56)

704
We will take moment about O. So, V into e is equal to that moment. So, which I say is equal
to 2 VR 4 t /I sin α - α cos α. That is what the moment we have here,. I have equated this. So,
now, I can easily get e as sorry, 2 R 4 t /I sin α - α cos α that is my e. If we know the value of
e, I can easily find the shear center, locate the center. So, look at this equation, in this
equation R is a geometric property, t is a section property, α is the angle of this curved section
subtended the center to α rather.

But the question is how about I. I is a moment of inertia of the whole section. Is it not? That
is what we understand. It is a segmented section; it is not a full curve. So, that we can find I
readily; the segmented section. So, the problem is how to find this. Let us do that.

(Refer Slide Time: 18:03)

705
Now, the issue is to find the moment of inertia I of the segment. So, let us draw this figure
again. I will take this figure again,. We know I is y square dA, and I want to integrate it for
the entire section. So, when I integrate I can do this as integration of β to𝜋/2 and 2 times of
this. y we already have with us, it is R cos𝜃, am I right. Look at this figure, R cos𝜃 will be the
y value. That is this value, this is R.

And this value R cos𝜃, is it not. So, R cos𝜃 the whole square and dA is R d𝜃 into t, am I
right? So, now, let us simplify this, so 2 R cube t integral β 2𝜋/2 cos square𝜃 d𝜃. We know
from trigonometry, that cos 2𝜃 is 2 cos square𝜃 - 1. So, hence cos square𝜃 is 1 + cos 2𝜃 /2.
So, substituting, I is 2 R cube t /twice β𝜋/2, 1 + cos 2𝜃 d𝜃.

(Refer Slide Time: 20:58)

706
So, I is R cube t integral β𝜋/2, 1 + cos 2𝜃 d𝜃, which is R cube t𝜃 + sin𝜃 /2 β 2𝜋/2, which is R
cube t𝜋/2 - β + sin 2𝜋/2 /2, - sin 2 β /2. We know already𝜋/2 - β is α, is it not. We know
already have it here, see here. Hence, I will now become R cube t, say𝜋/2 - β. So, this
becomes 0, half of 0 - half of sin twice of𝜋/2 - α. Can I say that? So, this is closed.

(Refer Slide Time: 22:50)

So, now I can say I is R cube t. See, look at this figure. So, this becomes α - this is half of sin
2𝜃, so 2 sin𝜃s are R𝜃. So, half of 2 sin𝜋/2 - α, cos𝜋/2 - α. Can I write like this? This is
argument is 2𝜃, let us say sin 2𝜃 is 2 sin𝜃 cos𝜃. That is what I am writing here. So, these two
goes away now, which will become R cube t α - cos α sin α. So, this is equation number 2.

707
So, now, e is 2 R 4 t /I, sin α - α cos α. That is what we have written here, 2 R 4 t /I sin α - α
cos α. So, now, therefore, substituting e will become 2 R 4 t /R cube t, sin α - α cos α /α - sin
α cos α,. So, which will now become 2 R times of sin α - α cos α /α - sin α cos α, that is my e.
So, we have all the values. So, geometric parameters, I can find e.

Now, I can locate the shear center for the given geometry. So, as a small example, let us copy
this figure, somewhere see this figure.

(Refer Slide Time: 25:36)

So, let us say, let R be 600 mm. Α or 2 α be 60 degrees. Find e. So, e is straight away 2 R
times of sin α - α cos α /α - sin α cos α. And compute this, mark the value and e will be
maybe x millimeters. So, friends, it is very simple that this particular numerical example
explains how to compute the shear center for a curved section as you see here.

708
(Refer Slide Time: 26:54)

We will do one more example. By a slightly a tricky section, let us see how we handle this
section. Locate the shear center of the section shown. So, section we have, ok; let us marks
on the dimensions of this section. Of course, we can locate there is an axis of symmetry for
the section. The section has uniform thickness as t. Let us mark the center line of the section
for our calculation purposes.

And we mark this dimension as b 1 and this dimension as b and this radius as R. We will say
the shear center is located somewhere here, and this is at an eccentricity e from here. Let us
mark the shear flow. We call this arm as V 5, this as V 4, this as V 3, this as V 2 and this as V
1, this as V 1. So, from this figure, we can easily say by symmetry, V 1 will be same as V 5.
Further, V 2 will be same as V 4. Let us cut the section at z from here, that is the thickness of
the strip b dz. This what we hatch. We want to find the cg of this, ok from here axis of
symmetry.

From the figure, you can always write this value as R + b 1 - z /2. Can I say that? Can I say
that? So, now, I can find V 1. V 1 will be the force in the vertical arm which will be
integrated from 0 to b 1, V /It, a y bar of da. So, a will be t into z and da will be t into d, and y
bar will be R + b 1 - z /2. I can substitute this.

709
(Refer Slide Time: 33:39)

So, V 1 is going to be integral 0 to b 1, V /It, t z, R + b 1 - z /2 of t d z. Let me rub this. So,

we can straight away say it is V t /I now integral 0 to b 1, R + b 1 - z /2 of z, d z, which can
be V t /I, R b 1 square /2 + b 1 b 1 square /2 - b 1 cube /6. After substituting the limits, which
becomes V t /I of R b 1 square /2 + b 1 cube /3. This is my V 1, is equation 1. Ok, friends.

So, we will continue this problem in the next lecture. But I want you to do a small homework
on this. I want you to find V 2. And now we have a curved section, we can find V 3 also. So,
try to do this exercise, then we will recollect it how we are able to do it. I leave a small gap
here intentionally, so that I want you to practice this. And try to do for V 2 then V 3 and so
on.

Thank you very much. And have a good day. Bye.

710
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 42
Shear center for curved sections

(Refer Slide Time: 00:20)

Friends, welcome to the 42nd lecture on Advanced Steel Design course, we are continuing to
do estimate or locating shear center for curved sections. So, we are in the middle of a
problem we will do that now in this lecture, hope you have done this. So, this was the last
expression we had for V 1 am I right.

711
(Refer Slide Time: 00:49)

So, we found out V 1 which is given by this equation which is V t by I Rb 1 square + b 1

cube by 3, so that is V 1 here. Now let us try to find out V 2. So, we will copy this figure
again rub this.

(Refer Slide Time: 01:33)

Now, I want to find V 2 ok? Let me cut a new strip now which is here at a distance x from
here thickness of the strip is dx and I want the hatched area is this, hatched area has got 2
components. I got 2 components now. So, I divide this into 2 components and try divide this

712
into 2 components here, I divide this into 2 components here and then try to find out. Let us
see how we do it.

So now, V 2 is integral 0 to b because that is the dimension we have for V 2 V by I t

thickness is uniform a y bar of d a ok, d a is very simple d a is area of this small strip which is
being considered which I am hatching here. Which is t d x there is no doubt, a y bar has got 2
components. So, first let us do it for b 1, so b1 into t of R + b 1 by 2. So, I have done for this
sample I am just marking it here I am marking it here I have done for the sample + this is for
1+ t x into R this is for 2 what is this one.

So therefore, V 2 now will become integral 0 to b V by I t b 1 into t of R + b 1 by 2 + t x into

R of t d x. So, we can simplify apply the limits please check we will get V 2 as V by I t times
of b1 R b + b 1 square b by 2 + b square R by 2 this.

(Refer Slide Time: 05:29)

We can simplify this I get V 2 in a simplified form as V t by I b 1 b R + b b 1 square by 2 + b

square R by 2 equation 2. So, equation 1 we had V 1 equation 2 we had V 2. Now I want to
find for the semicircular portion V 3? We already know this equation let us draw this figure
we are looking for the curved portion, which is this, I will hatch it here maybe with this color
I am looking for this portion we already have an expression for this ok, if you know the
angle, I can do it.

713
So now, the elemental shear stress V by I t will be integral y d a + 𝜃1 + 𝜃 2. Let us say we
will put this Q 1 + Q 2 where Q 1 and Q 2 are the shear of first and second pieces. So, let us
do that V by I t we already have this equation. So, I am writing it here R cos 𝜃 into R d 𝜃 into
t am I right + b 1 times of t into R + b 1 by 2 this is for piece number 1 + b t into R this is for
piece number 2.

(Refer Slide Time: 08:33)

So, therefore, I can now expand this which will become τ V by I R square cos 𝜃 d 𝜃 + b 1 R +
b 1 by 2 + b into R, which will now become V by I R square sin 𝜃 + b 1 R + b 1 by 2 + b R.

So, please understand this integral y d a is applicable only for the curve section ok, for the
remaining section we only have the forces with us into b R. So, V by I let us mark it this way
this y d a is only for the curved section. So, we have now the value for τ which is V a by I R
square sin 𝜃 + b1 into R + b 1 by 2, let me write it here clearly + b R.

Now the elemental shear dv will be τ d a which will be V by I R square sin 𝜃 + b 1 times of R
+ b 1 by 2 + b R of R d 𝜃 into t am I right look at this figure. We are looking for d 𝜃 as a
small segment I will mark it here the segment we are looking for. We know this is 𝜃 and this
angle is d 𝜃 this is the sigma which is y d. So, this R 𝜃 d 𝜃 is very clear from there.

So, now I can find d V as V into t R by I this t has now come into play here of R square sin 𝜃
+ b 1 times of R + b 1 by 2 + b R of d 𝜃. Now we would like to take moment about of this

714
force about the point o this is the point o take moment of this point, the force V is acting here
we will take the moment of this particular d V.

(Refer Slide Time: 12:28)

So, we will see taking moment of this elemental shear about o. So, we call that as d M which
will be τ d a into R see here the distance will be R is it not into R. So therefore, d M now will
become V t R square by I let me write the equation for τ d a, d M will now become V t R
square by I R square sin 𝜃 + V 1 times of R + b 1 by 2 + b R of d 𝜃. Integral d M will be the
total moment M which will be equal to V t R square by I integrated from 0 to pi R square sin
𝜃 + b 1 times of R + b1 by 2 + b R of d 𝜃 equation 3.

So, M will now become V R cube sorry R square t by I R square - cos 𝜃 + b 1 times of R + b
1 by 2 + b R 𝜃, applying the limits 0 to pi. I will copy this equation.

715
(Refer Slide Time: 15:13)

So, we will do V R cube by t by I R square 1 + 1 + pi times of b 1 R + b 1 by 2 + b R which

now become V this is R square sorry V R square t by I 2 R square + pi times of b 1 into R + b
1 by 2 + b R equation number 4. You have the M value now will take moment about the point
o take moment about point o we get V into e should be equal to M is actually equal to M, let
us look at this figure and write this equation we will copy this figure and write this equation.
Selection tool this is V 3let us mark this this is V 3 curved section.

(Refer Slide Time: 17:09)

716
So, now V into e because this was V in this is the point o V into e which is anticlockwise
should be equal to M-V 1 into b by 2 - V2 into R these are for sorry let us do it this way M -
V 1 b twice, because I have V phi also - V 2 into R twice. So, friends I have equation M see
here I have equation M which is having the term V right. So, M is there we also have
equation for V 1 we had the equation for V 1 which also has V.

So, we are equation for V 1 we also have equation for V 2 which also has V so V 2 also here.
Of course, b and R are geometric properties ok, can I find e? So, e will be so V goes away M
by V - 2 V 1 b by V - 2 V 2 R by V I can find e. Now, friends in all these equations of M V 1
and V 2 see here in V 1 I have I in V 2 I have I in M also I have I, where I is the moment of
inertia of the whole section.

So, for first principles you will be able to find out the moment of inertia of all the 5 parts part
number 1 part number 5 part number 2 1 4 and part number 3 I think moment of inertia is
sum of second moment of area of all the pieces you can do it for component 1 and 5
component 2 and 4 + component 3.

I will give you the equation you please check. The moment of inertia of this whole section is
given by R t cube by 6 3 pi + 12 times of b + b 1 by R + 4 times of b 1 square by R into 3 + b
1 by R is the moment of inertia of the whole section you can check this. So, friends we have
learnt how to do the shear center problem, for curved section which has caught a mixture of
rectilinear and curvilinear sections from the first principles we found out how to locate e?

So, locating shear center for geometrically complicated sections not that easy ok, we have got
to gradually divide them into parts into finite elements and keep on assembling them and do it
slowly and patiently it is not that easy. So, when you have got form dominant designs where
the geometric shapes are dependent on the functional requirements obtaining shear center, to
check whether unsymmetric bending will be caused whether twisting will happen whether
stability will be checked all will become a prerequisite.

So, standard software’s cannot handle special kind of structural shapes, it is always better
friends to use the cad models develop these geometric shapes, identify the shear center and
calculate the necessary values yourself with MATLAB program. So, the whole intention of
this course is to make you independent researcher and independent thinker and not making
you dependent on using software as a black box.

717
So, advancement is not only on design process is not only on differing defining equations, not
only on checking stabilities, not only an understanding plastic design it also enables you
inherently to be an independent thinker. So, the MATLAB program is not that difficult is a
very user-friendly software you have got good support from matlab incorporation.

So, they will try to help you answer your queries on time and you can also contact me
through discussion forum, but the whole objective is try to learn have a grip of the problem
and enjoy your creation as a programmer ok; that is the beauty in the whole exercise right.
Having said this let us do a couple of more problems. So now I am going to take a different
section I will call this as example 3.

(Refer Slide Time: 24:06)

So, I am going to use or check an unsymmetric section friends, it is a very tricky problem. So,
far we have been discussing the shear center location for a symmetric section, the section at
least had one axis of symmetry the shear center was located along the axis. But that is not the
true truth all the time you can have unsymmetric sections, unsymmetric sections have their
own complications I think we all know that now by this time we have understood how to
operate an unsymmetric section.

Let us take an example have a section like this unequal channel section with unequal flanges
let us mark the dimensions of the section, let us say this is 40 this is also 40 and this is 20 the
section has uniform thickness of 2 all dimensions or in millimeters. So, the moment I have an

718
unsymmetric section I will mark the axis the z and y axis x axis normal to the screen this is
my y axis this is my z axis. So, immediately these values are required.

I call this as z bar and I will call this as y bar this means C g not the shear center. So, C gwe
will divide this as 3 parts this part number 1 component 2 and component 3, can you find the
z bar and y bar not to locate C g is the first issue here. That is I want to find z bar and y bar of
this problem, we know z bar is given by the simple expression algebra sum of a z bar by a
and y bar is a y bar by a we know this.

So, let us work out z bar. So, 20 into 2 into 10 + this is for piece number 1 40 into 2 40 into 2
into 20 this is for piece number 3 + 36 into 2 into 1 piece number 2 divided by 20 into 2 + 40
into 2 + 36 into 2 which will give me z bar as 10.8 mm I will mark this value.

(Refer Slide Time: 29:24)

Let me copy this figure. Let us find y bar which can be 20 into 2 into 1 piece number 1 + 40
into 2 into 39 piece number 3 + 36 into 2 into 18 + 2 am I right, this is for piece number 2
divided by 20 into 2 + 40 into 2 + 36 into 2 which will give me y bar as 23.96 millimeters.
So, let me mark this value as 23.96.

I also want to find the moment of inertia of this section, because I want to locate the principal
axis in the unsymmetric section is it not. So, next step will be step number 2 to find I y and I
z and of course I y z for this section, then only I can locate the principal axis. So, I can write I

719
z directly here that is horizontal axis I can use parallax theorem. So, 20 into 2 cube by 12 +
20 into 2 into 22.96 squares, this is for piece number 1 +.

Let me do it + 40 into 2 cube by 12 40 into 2 39 - 23.96 the whole square all right this is for
piece number 3, + 2 36 cube by 12 + 36 into 2 into 23.96 - 20 the whole square this is for
piece number 2. If I do this, I will get this value as 4.8 10 power 4 which is I z.

(Refer Slide Time: 33:29)

Let us copy this figure let us do I y, I y is a vertical axis which will be 2 into 20 cube by 12 +
20 into 2 into 10.8 - 10 the whole square this is for piece number 1, + 2 into 40 cube by 12 +
40 into 2 into 20 - 10.8 the whole square this is for piece number 3, + 36 into 2 cube by 12 +
36 into 2 into 10.8 - 1 the whole square this for piece number 2.

So, if you do this I will get I y as 2.57 10 power 4 mm 4 this I y friends this first principles I
mean there is no confusion in this very clear you use parallel axis theorem we have done it.
Let us do I z y, I z y is 20 into 2 into 0.8 into 22.96 very simple you can see this value this
will be here.

And for the first piece see this value this will be here this is for the first piece + 40 into 2 to -
9.2 into - 14.0 you can very well see this C g is located here. So, it is negative in terms of y it
is negative in terms of z and these values can be seen from here, this value is this and this
value is this+ this for piece number 3 + 36 into 2 into 9.8 into 3.96 is for piece number 2.

720
This is 10.8 friends make a correction here I am sorry. So, this value is same as this, and this
is 3.96 it is coming here. So, I z y after you compute will become 1.46 10 power, since I z y is
not 0 the section will be now subjected to unsymmetric bending and z and y will not be the
principal planes we know that. So, once we understand this, I can straightaway locate the
angle of inclination of this figure.

(Refer Slide Time: 38:27)

Let me draw this figure again and let us mark the u u axis and v v axis say this is my v this is
my u axis and we know this angle measured from positive z is α. And we know tan 2 α is
given by - of I y z by z - y now we have all the 3 values with us we will substitute them -
twice of 1.46 10 power 4 by z value is 4.8 and y value is 2.57.

So, we get α as - 26.31 degrees measured clockwise from positive z to positive u. So, this is
26.31 degrees. So, I located the principal axis.

721
(Refer Slide Time: 40:19)

My job is going to find out I u and I v because I need to find I u the moment of inertia about
the principal axis now. So, we know I u is I y + I z by 2 + z - y by 2 cos 2 α - I y z sin to 2 α
this equation we already derived we know. So, substitute them we will be getting 64144.83
mm 4 I v can be also said as I z + I y - I u. So, I can say 9645.17 mm 4 you can check these
values friends. Once we have try this our job is to locate the shear center let us draw this
figure again.

(Refer Slide Time: 41:43)

722
So what I am now going to do is ok, I will redraw this figure slightly gentleα is 26.31. So,
now if I remove this label and say I am going to apply the load v and the reaction V R here
along the u u axis please understand I am writing it normal. So, it is along v axis. So, let us
say this is v these are principal axis.

So, I call this as V R. Now if I take moment about this point and I call this distance as e, I
have an advantage if I write this as V 1 V 2 and V 3 and I call this point as a by taking
moment about A V 2 and V 3 being concurrent about A we will cause no moment. So, I do
not have to work out V 2 and V 3. So, V 2 and V 3 need not be calculated. So, I want to
calculate only V 1.

So, to do that I will cut the section this section will be at a distance z thickness dz and the C g
of this need to be measured from this axis I want this C g to remember that. We want this we
call this V bar am I write where I am transforming everything into the principal axis now. So
now, to find V 1 V 1 is τd a integral which is integral τ d z of t the τ is V A y bar by I t which
is V A, now V bar by I u into t am I right.

(Refer Slide Time: 46:27)

Please look into this equation I am transforming them into principal planes right, therefore V
1 will be V by I u integral A v bar by t into t d z which will be V by I u z t into v bar by t in
into v bar d z because this t goes away right. So now, I want to find this v bar? This v bar is
now equal to we can make use of this figure, if I have 2 axis y and z if you have another 2
axis v and u and this angle is α in simple terms v can be given by - z sin α + y cos α is it not.

723
So, I want to substitute the z value - of - 20 - 10.8 - z by 2 sin α which is - 26.31 degree+ 22.9
cos α which is - 26.31 degrees. So, friends we have the v value with us which can be
simplified and said as - of 9.2 - z by 2this is z let me make it very clear - z by 2 of - 0.44 +
20.5.

So, v now becomes, or v bar becomes - 4.048 + 0.22 z + 20.93 which becomes 16.482 + 0.22
z right.

(Refer Slide Time: 50:28)

So, once we have this value, I can substitute in V 1 I leave a small homework here I want you
to do this before you check this result for my next lecture. So, in this lecture we learnt how to
work core shear center for asymmetric sections. I also left you a homework exercise and you
will check that before you attempt to the next lecture.

Thank you very much and have a good day, bye.

724
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 43
Shear center for unsymmetric section

(Refer Slide Time: 00:18)

Friends welcome to lecture 43 of Advanced Steel Design course. We are now continuing the
problem of unsymmetric section, and we are trying to locate the shear center.

725
(Refer Slide Time: 00:50)

So, this was the unsymmetric section we had. We located the uu and vv axes and we first
located the cg of the section that is we found out y bar and z bar, then we located. We
computed I y I z and y z and then computed alpha angle and located uu & vv axes.

Then we are in the process of finding out the shear center, then we attempted to take moment
of point A where V 2 and V 3 are concurrent, and no moments will be created. So, we wanted
to only find V 1. To do that we wanted to find out the distance v bar because we are using, or
we are transforming the problem to the principal planes. So, we wanted to find out v bar.

(Refer Slide Time: 02:32)

726
And this was the expression we had to find out v bar. So, v bar is 16. 48222 z. So, then we
substitute that v bar here in this equation and get V1.

(Refer Slide Time: 02:48)

So, now, V 1 will be seen here V t by I u integral. Now we can say this will be for length of
20. So, I can say 0 to 20 right. So, see here we have A which is z d z that is what we are
looking for z d z is what we are looking for so z d z. So, let us say V u V t by I u 16.482 +
0.22 z, z d z ok that is what we have here, is it not? V by I u t z d z v bar, v bar is already
here.

We will evaluate this integral, and we will find out this as V t by I u 16.482 z square by 2 +
0.22 z cube by 3 of 0 to 20. If I do that and t is 2 millimeters V1 will now become 0.121 V.
Now we will take moment about this.

727
(Refer Slide Time: 04:36)

So, we will draw this figure. Now we will take moment about A. So, V we will call this as eu,
V *eu will be equal to this is V 1. So, V 1 and this distance 38.

So, already we have V 1 with us. So, V 1 is 0.121 V. So, V into e u is 0.121 V of 38. So, V
goes away e u is 0.121 into 38 which is about 4.59 millimeters. So, this is about 4.6
millimeters, that is how we can find out.

(Refer Slide Time: 06:09)

728
So, friends when you have a problem of unsymmetric sections one need to locate the
principal axes before proceeding to locate the shear center, that is an important point which
we must remember.

(Refer Slide Time: 07:38)

We will do one more example. Say example 4 draw this figure. So, the problem is locating
the shear center of the given field ok section like this. Of course, this has an axisymmetric
which I am marking here. Let us mark the dimensions I think here in mark the dimensions of
this. First let us mark the center lines and mark the shear flow. Let us say this is V 4 this is V
3 this is V 2 this is V 1. Let us mark this dimension that is b 1 and this dimension as b is also
b, this is 45 degrees.

Of course, this is also b 1. The section has uniform thickness t this is a point which we call as
point A and the shear center will be located somewhere on this line. Let us say this is my
shear center C. The vertical force will pass through this, and we measure the distance from
this center from the point A as offset of shear center which is e. So, from this figure by
symmetry we know that V 1 is equal to V 4 and V 2 is equal to V 3 because of the dimension
similarity.

So, we will cut a section. Let the section be cut from here at a distance z and let the thickness
of the section be d z and of course, this thickness is already given as t. We are looking for the
c g of this from here. This is my y bar for piece number 1 ok, for piece number 1. So, now, V
1 is given by integral 0 to b 1 V by I t a y bar of piece 1 into d.

729
So, a of piece 1 is t into z da of piece 1 is t into dz y bar is b sin 45 - b 1 sin 45 + z by 2 sin 45
say b sin 45 let us say you see here. So, looking for this distance ok b sin 45 will be this, b 1
sin 45 you will be reaching here. Then if you do this is you know z by 2 will give to the
center get this right. So, which will be b - b 1 + z by 2 of sin 45 which is y bar.

(Refer Slide Time: 13:28)

Substituting let us get V 1 as integral 0 to b1 V by I t t z into t dz b - b 1 + z by 2 of 1 by root

2 that is sin 45. Integrating and substituting the limits we get V t by root 2 I b 1 b 1 square. It
is not b 1 is b, b b 1 square by 2 - b 1 cube by 2 + b 1 cube by 6.

Which will now become V t root 2 I b b 1 square by 2 - b 1 cube by 3. Therefore, we get v 1

as V t b 1 square by root 2 I of b by 2 - b 1 by 3 which is identically same as V 4. Now we
taking moment about the point A I do not have to work out V 2 and V 3 because they are
concurrent about the point t.

So, by taking moments of all forces about A V2 and V3 or concurrent and hence no moment.
So, there is no need to calculate these 2. So, directly I can use V 1 and V 4, but there is a
issue here V 1 need to calculate I of the entire section. So, this I is a moment of inertia of the
whole section. So, that is a tricky thing here. So, let us do one by one. So, we need to find I of
the whole section.

730
(Refer Slide Time: 16:41)

So, now, to find I of the whole section whole X dash X dash and we call this small axis as uu
and vv. We know and this dimension is of course, b and this thickness is t, this angle is 45
degrees.

We know I uu is tb cube by 12 and I vv is bt3 by 12 and I x dash x dash is I u cos square𝜃 + I

v sin square𝜃 where 𝜃 is 45 degrees here. So, I can say it is tb3 by 12 cos square 45 + bt3 by
12 sin square 45 which will be tb cube by 12 1 by root 2 the whole square + bt cube by 12 1
by root 2 the whole square which becomes tb cube by 24 + bt cube by 24, this is I X dash X
dash.

Now, I have an axis here which is XX. So, now I can say I XX will be I X dash X dash + ak
square which will be tb cube by 24 + bt cube by 24 + ak square which is b into t of the
distance. You know XX axis is not here. We should mark it here look at this figure XX axis is
the tip ok let us mark it here XX. So, which will be b by 2 sin 45 the whole square.

731
(Refer Slide Time: 19:59)

So, which will give me I XX as tb cube by 24 + b t cube by 24 + b cube t by 8 ok, b cube t by

8. So, which will now become 4 tb cube by 24 + bt cube by 24. So, we should say tb by 24 4
b square + t square is my I XX of that arm. Let us do it for the arm b 1 because we have no
we have done for this arm. Let us do it for this arm.

(Refer Slide Time: 21:11)

Let us draw. So, this is my V 1 I am looking for an axis which is XX. Let us also mark
centroidal axis as u u and v v for this and we call this horizontal axis as X dash X dash and
the dimensions are of course, this is b 1 and this is t. So, now we can say I u will be b 1 t cube

732
by 12 and I v will be tb 1 cube by 12. So, I X dash X dash will be I u sin square 45 + I v cos
square 45 which will be b 1 t cube by 12 1 by root 2 the whole square t b 1 cube by 12 1 by
root 2 the whole square.

So, I X dash X dash will be b 1 t cube by 24 + t b 1 cube by 24. So, now, I XX will be I X
dash X dash + ak square we can apply this. So, which will be equal to b 1 t cube by 24 + t b 1
cube by 24 + b 1 t of you know the angle is connected here. This is how it is connected, and
this dimension is b, is it not? So, I should say b sin 45 - b by 2 sin 45 the whole square that is
my k square here. Let me copy this equation.

(Refer Slide Time: 24:20)

So, I XX you know equal to b 1 t cube by 24 + tb 1 cube by 24 + b 1 t b by root 2 - this is b 1

ok b by root 2 - b 1 by 2 root 2 the whole square is it. Simplify I get this as b 1 t cube by 24 +
t b 1 cube by 24 + b 1 t of b square by 2 + b 1 square by 8 - b b 1 by 2 which will be b 1 t
cube by 24 + t b 1 cube by 24 + b 1 b square t by 2 + b 1 cube t by 8 - b b 1 square t by 2 this
is my I XX of the first piece.

733
(Refer Slide Time: 26:08)

So, now we can say I XX of the whole section will be, let us say the piece number this value I
am copying it here ok which I am writing it here tb by 24 4 b square + t square. Let me rub
this then I will copy this down. Let me just write it here + b 1 t cube by 24 + t b 1 cube by 6 +
I am simplifying ok + b 1 b square t by 2 - b b 1 square t by 2 the whole into 2.

Now, I draw this figure take a moment of A this is the point A. So, we can say V into e will
be equal to V 1 into b right, because that is V1 from here and there is same as V 4 so into 2.
So, now I have V I have V 1 I have b. So, find e. So, for example, if b 1 is 40 b is 70 t is 2
millimeters ok you can find A. So, friends in this lecture we learnt 2 examples.

734
(Refer Slide Time: 29:12)

One we learned how to work out the shear center for unsymmetric section and we have seen
the steps involved to locate the shear center. We have also learnt one more example and in
this example, we learnt how to choose the point to take moment of all internal forces to find
the shear center. So, we can choose this point very advantageously, so that our computational
effort can be reduced.

So, friends with this we will be completing the discussion on shear center. There are many
more problems available in the textbook referred by me for this course. Look at those
examples and try to solve them we have also seen curved sections we have seen asymmetric
sections. We have seen symmetric complicated sections which very common sections are
used in mechanical systems as well as structural systems.

For example, this kind of sections are used for crane rails. This kind of sections are used for
pipe racks because you get more space to put the pipe. So, these are all rectilinear shapes, but
arranged in a such a fashion they create very complicated geometry for computational
purposes. So, you should be able to look into these and practice more problems. So, that you
are clear about solving them from the first principles.

Thank you very much and have a good day.

735
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 44
Curved beam - 1

(Refer Slide Time: 00:20) ϕ

Friends, welcome to the lecture 44 of the course Advanced Steel Design. In this lecture we
are going to learn more about Curved Beams, I will call this lecture as curved beams - I.

736
(Refer Slide Time: 00:45)

The foremost question comes in mind is, how curved beams are classified? Curved beams are
classified in 2 ways, beams with small initial curvature and beams with large initial curvature.
Now, what do you understand by small and large? How do you quantify this? If the initial
radius of curvature divided by depth of the section is greater than 10, we call this designated
as beams or curved beams with small initial curvature.

(Refer Slide Time: 02:24)

Now, let us look into the discussions of behavior of curved beams with small initial curvature
in bending. We will look into this now. Let us say, I have a curved beam, , shown in the figure

737
here. Let us say this is now subjected to in a moment as shown the figure. Let us say the
radius indicated is R. So, R is the initial radius of curvature d ϕ is angle subtended at the
center of curvature. I will mark d ϕ, by the element abcd. Let us first mark the element abcd,
say this is the angle subtended, we call this as d ϕ and this element this is a, this is b and this
is c and this is d.

Then, R prime is a radius after applying the moment. See when you apply a moment as
shown in the figure; the radius of curvature now changes. Let us say, we mark a new center
here. And let us see the new angle subtended by this is d ϕ prime and the radius is R prime.
Now, I will take a fiber. Please, note, that R prime is much lesser than R that is it is very
important applied moment is closing the curvature.

See before M is applied, the radius of curvature was R; after M is applied it reduced to R
prime in such a condition that R prime is smaller than R. So, the applied moment actually
closes the curvature. We have to remember this very carefully, right. This is a very important
convention we need to. And, of course, d prime d ϕ prime is now the angle subtended by the
element after deformation.

Let us mark the element layer P, Q and mark it here at a distance y, this is P, Q at a distance y
away from the center. So, let us mark the cross-section somewhere here. Let us mark the axis.
We call this as n-n axis which is the neutral axis. And we take any point which is above this
measured as y, that is what we have marked. So, it is important M is applied in such a manner
that it closes the curvature. It does not open it does not open the curvature. It closes the
curvature.

738
(Refer Slide Time: 09:56)

We consider a fiber at y from the neutral axis. Now, let us see what the original length of this
fiber is. The original length of this fiber will be R + y, d ϕ is it not, before M is applied will
be equal to R + y, d ϕ sees this figure. R is this value, and this is y, d ϕ is a angle that is the
original thing.

Now, after applying this moment M, which is closing the curvature the length of this fiber is
now reduced to R prime + y d ϕ prime, am I right. After M is applied, which tends to close
the curvature length of the fiber is now given by R dash + y d ϕ dash. Therefore, friends
change in length of the fiber because we are looking for strain which will be R dash + y d ϕ
dash - R + y d ϕ.

So, now I can find the strain, epsilon which is change in length by original length which will
be equal to R prime + y d ϕ prime - R + y d ϕ by R + y d ϕ. We will call this equation number
1. But please note that, length of any fiber located in the neutral axis remains unchanged.
There is no change in the neutral axis.

So, friends if I take any fiber in the neutral axis let us say I am marking that fiber here, if I
take any fiber in the neutral axis that is this fiber, this length will remain unchanged. So,
therefore, ds will be R d ϕ which is same as R dash d ϕ dash see in this figure. If I call this as
ds before and after applying the moment, there is no change; because this is a fiber measure
the neutral axis. So, I call this equation number 2.

739
(Refer Slide Time: 14:34)

Now, I substitute 2 in equation 1. And simplify. Let us see what do we get? We get strain;
because this equation is for the strain equation 1, , which will become y times of d ϕ dash - d
ϕ by R + y d ϕ is it ? , call this equation number 3. Now, we know that the distance of the
fiber, y from the n-n axis is small compared to the initial radius of curvature.

We have given a small displacement. Therefore, y is very very less compared to R. Hence, y
may be neglected in comparison to R wherever we have this formation. Therefore, I write
modified strain, epsilon will be y times of d ϕ prime - d ϕ by simply R d ϕ because R + y is
identically equal to R for y being small. We call this equation number 4.

740
(Refer Slide Time: 17:23)

Now, from equation 2, we know that ds is R d ϕ. Let us see what is equation 2, which is also
equal to R dash d ϕ dash. So, d ϕ is ds by R and d ϕ dash is ds by R dash, we call this
equation number 4 a. So, we have equation 4, where d ϕ dash is there substitute equation 4 a
in equation 4. So, equation 4 is for epsilon which is y times of d ϕ dash - d ϕ by R d ϕ, which
will now become y times of d ϕ dash by ds - d ϕ by ds which is equal to y times of 1 by R
prime - 1 by R equation 5.

(Refer Slide Time: 19:16)

741
Further, epsilon is also ratio of stress to this. So, therefore, y of 1 by R prime - 1 by R should
be equal to stress by E. So, stress by y is E times of 1 by R prime - 1 by R we call this
equation as equation number 6. So, we make certain assumptions, while doing this derivation;
let us see what are they.

(Refer Slide Time: 20:05)

There are certain assumptions made in curved beams, which are similar to straight beams. 1.
Every cross section of a curved beam remains plane and perpendicular to the centroidal axis
before and after application of external moment. Now, to satisfy this condition net force
acting on the cross section of the curved beam should be 0. If it is not equated to 0, then it
will cause warping.

742
(Refer Slide Time: 22:21)

So, mathematically integral stress d A for the area should be 0 equation number 7. Now, let
us substitute equation 6 in equation 7. What is equation 6? We have the stress value in
equation 6; here let us substitute that in equation 7. So, you know this should be now equated
E times of integral of y 1 by R dash - 1 by R of d A and equate this to 0, which means E times
of 1 by R dash - 1 by R integral y d A should be equal to 0. So, if you call this equation
number 8, let us copy this equation.

(Refer Slide Time: 23:46)

743
In equation 8, E times of 1 by R dash - 1 by R cannot be equated to 0. Hence, integral ydA
should be only equated to 0; because, if this is equated to 0 if this is equated to 0; then no
curvature. What does it mean? We call this equation number 9. Equation 9 implies that
geometric axis coincides with neutral axis.

How can we say this? You know y is measured from the geometric axis. See this figure. And
this is my neutral axis. So, this condition is now applied, and it is very interesting this is
forming the basis for the derivation of the curved beams.

(Refer Slide Time: 25:36)

Now, since, the curved beam is in equilibrium under the applied moment M, one can,
following statement holds good d A into y should be equated to the moment. Now, from
equation 6, substitute for stress in equation 10 we get, E times of A y 1 by R dash - 1 by R of
y dA is here, which implies that E 1 by R dash - 1 by R integral y square dA should be M. We
have interesting term arrived here, what is this y square d A called second moment of area.
That is a very standard expression we have in mechanics. We know integral A y square dA is
the moment of inertia.

744
(Refer Slide Time: 27:35)

Hence E times of 1 by R dash - 1 by R into I is M, am I right, see here. So, now, I have a very
interesting and standard relationship. M by I is E times of 1 by R dash - 1. I mean this is a
familiar equation similar to that of a straight beam; we call this equation number 11. Now, let
us copy equation 6 here, now I want to copy this equation and then compare. I want to
compare.

Now, let us compare equation 6 and equation 11. So, right-hand side is same. Can you now
say, stress by y M by I is E times of 1 by R dash - 1 by R standard relationship friends for a
curved beam which is more or less looking similar to that of a straight beam is it not,
whereas, we have a new term here 1 by R dash, , that is the theory of flexure which we
recollect that equation M by I is stress by y is E by R the standard relationship we have for
straight beams, is it not. It is more or less similar to that. This is what this is interesting. So,
the curve beams are conserved.

745
(Refer Slide Time: 30:08)

Having said this, let us take our discussion forward for curved beams of large initial
curvature. Now, let us recollect if initial radius of curvature to depth of the section is lesser
than 10. We can call this beam; this curved beam is termed as beam with large initial
curvature.

Can I say that? Because, more than 10 is small, I can say this. So, there are some special
applications and properties of this condition. What are they? There are some special
conditions. 1 The stress variation in such beams across the depth of the section will be
non-linear. What does it mean? Stress in the concave side is generally greater than stress in
the convex side, these are experimental observations.

Experimental studies in the literature show that this statement is valid, that is a general
information we need.

746
(Refer Slide Time: 32:38)

The 2nd point is neutral axis will not pass through the centroidal axis. Why? Because, the
stress on the concave side is not equal to the stress on the convex side; further, stress along
the depth is non-linear. So, these two conditions make this condition valid.

(Refer Slide Time: 33:51)

Now, we will take up this for our discussion. So, I am talking about beam with large
curvature. Let us draw, let us call this angle as d ϕ. Let us cut a new section as d ϕ prime from
the same center. Say, this angle is d ϕ dash. We mark the four corners A, B, C, D. Let us mark

747
a layer P, Q. Let us mark the neutral plane first on the X-X axis let us mark the X-X axis we
call this as X axis.

The neutral plane is lying below this. This is a neutral plane. Let us take a fiber P, Q marked
at a distance y from X axis. Say, this is P, Q ϕ. This is the convex side also called as extrados.
This is the concave side also called as intrados. We are applying a moment which is closing
the curvature that is the nature of the moment. These are the standards in convention we have.

Now, the offset of the neutral axis from the X axis is marked as E. and, the distance of neutral
axis, sorry, X-X axis from the center is marked as R. We will call this as d ϕ we will not have
marked this. We calls this d ϕ that is a strip being cut, that is a strip. Let us say, I project this
project this and draw a cross section in a cross section whatever may be the shape.

I draw a reference line call this line as O O. So, on this I mark the cg. So, the X-X axis will
pass through this cg,. This is my X-X axis. I pick up any point on the line P, Q let us be here
and we know this distance is measured from the X-X axis at y. I can also mark the green line
as N-N axis which is the neutral plane. Now, from the cg, I mark this X-X here; from the X-X
axis I mark the tip as h extrados, I mark this step as h intrados.

And the X axis is marked at a distance R from this plane. So, I mark this dimension as radius
intrados and this dimension as radius extrados and, this as radius neutral axis,. This is a
reference figure here, is that clear. So, this is figure 1, this figure 2. So, we have neutral plane,
we have X-X axis and the neutral plane is offset from the X-X axis towards the center of
curvature.

Neutral plane is shifted towards the center of curvature by y from the X-X axis, correct, see
here by e, not y by e. Having said this, we will now derive the governing equation for this
section. friends, we will discuss that in the next lecture, we will put a summary here for this
lecture.

748
(Refer Slide Time: 42:13)

Friends in this lecture we learnt curved beams with small initial curvature. We have derived
the equation stress by y M by I, E by 1 by R prime - 1 by R and it is more or less similar to
the straight beam, is it not equation. We have also learnt the conditions of curved beam for
large initial curvature. We have also learned a fact stresses in extrados are not equal to stress
in intrados.

And therefore, the stress distribution across the depth of the section of the curved beam is
non-linear. We have learnt this fact. We will continue to discuss this in the next lecture and
derive the control equation for finding out these stresses in extrados, intrados using what we
call as Winkler Bach equation.

Thank you very much and have a good day, bye.

749
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 45
Curved beam with large curvature - 1

(Refer Slide Time: 00:20)

Friends, welcome to this lecture 45 on Advanced Steel Design. We are now continuing to
discuss the derivation for finding out stresses on Curved beams under large initial curvature.
So, we are looking for curved beams with large curvature. So, we discussed the derivation for
finding out stresses at the intrados and extrados of the curved beam for small initial curvature.

And we also said very clearly that the stress distribution along the depth of the section in
curved beams is non-linear. The stress distribution across the cross section of a curved beam
is non-linear therefore, the classical bending equation cannot be used to find out the stresses.
So, we need to really derive the control expression for finding out the stresses in the extrados
and intrados that is the maximum tension and maximum compression on the concave and
convex side respectively.

750
(Refer Slide Time: 02:09)

So, this was the discussion what we had in the last lecture I urge that you should please
redraw this figure again I will hold the screen for a second. So, we have got the condition for
a large initial curvature, please remember that the moment is applied at the ends in such a
manner that it tends to decrease the cross section. So, the convex side and concave side are
marked on the screen which is also called as extrados and intrados and we have marked the
neutral plane with an offset from the x axis by the amount e which is indicating a shift of
neutral plane towards the centre of curvature.

There is a very important statement friends. This is moving towards the centre of curvature
by an amount e. We pick up a fibre at a distance y measured from the x axis which is away
from the centre of curvature and we mark that fibre as pq. So, the figure 2 shows the cross
section of the member any arbitrary shape where the extrados and intrados are marked as Ri
and RO on the screen R stands for the radius and O stands for extrados or outer end I stands
for intrados or the inner end.

Whereas, the xx plane is measured at a distance R from the radius from the centre of rotation
O. Please understand x axis is different from the neutral axis it cannot be same because x-x
the centroidal axis x-x is the centroidal axis and n-n is a neutral axis they cannot be same for
a simple reason the stress distribution along the depth is non-linear its not going to be same
in that case there is no coincidence of this and this shift is indicated as e in the figure.

751
Now, there is a specific way by which we assume m; m is closing the curvature right please
understand these are all standard conventions which has a positive convention in the analysis.
We pick up any point y on the fibre P Q which is offset measured from the centroidal axis it
is nothing to do with the neutral axis. I am always measuring the point respect to the
geometry neutral axis are layer or a plane respect to the stress distribution there are two
different things here we have to be very clear about that.

From the geometry the section depth is marked from the central axis as h i and h o that is
height or depth of the intrados from the centroidal axis and extrados from the central axis it
can be equal may not be equal depending upon the location of cg because the sections can
have also an asymmetric cross sections.

(Refer Slide Time: 06:08)

Having said this we will say let the curved beam has a large initial curvature. The radius of
curvature is indicated as R in the figure where M the applied moment tends to decrease the
curvature.

So, friends, in curved beams there is no hogging and sagging bending moments you have to
classify the applied moment in terms of does it increase or decrease the initial curvature of
the curved beam.

Therefore, friends, m causes tension in the extrados is it not and of course, compression in the
intrados. Now let us consider an element ABCD we will consider an element ABCD we are

752
considering an element A B C D. So, let us consider an element A B C D subtending an angle
d ϕ at the centre of curvature. We see this figure the member or the segment ABCD is making
an angle d ϕ.

Now, under the action of M this element deforms is it not this element will deform now
because the radius is going to change; obviously, this element will deform.

(Refer Slide Time: 09:39)

Let us now draw the deformed shape of this now let me draw the element we call this as A
this is C that is what we have is not yeah and let us call this as B and D this is my element let
me also mark P Q. So, P Q was in crayon color let us mark P Q this is P Q.

Now, let us mark the centroidal axis which is in orange color. I call this as X axis this is the
centroidal axis and we know that the element PQ is at a distance y from the centroidal axis is
it not let us mark that this is y away from the centre of curvature now let us also mark the
neutral axis which is in green color. So, this is my neutral axis and we know this is at a
distance e shifted towards the centre of curvature.

Now, under the applied moment M which is tending to close the curvature or decrease the
curvature the element is going to deform. Let us draw the deformed condition of this element.
So, what we do is, A B C D is deforming we will not disturb one face we will just disturb
only the other face for our simplicity let us mark that. So, at the neutral axis there is no

753
change now let us say it is going to deform. We call this point as C dash at this point D dash.
So, Q will also get extended to Q dash sorry.

So, we call this angle as Δd ϕ. So, now D is shifted to D dash and let us note neutral axis is
shifted towards the centre of curvature. There is an important observation which we need to
understand the shifting towards the centre of curvature.

Now, with reference to this figure for our convenience face AB is fixed that is this face is
fixed face CD is rotated and to C dash D dash. As you can see in this figure you can see here
this is what you have rotated here we have rotated. Now this new face C dash D dash is
subtending an angle del d ϕ sorry Δd ϕ as marked I mark this angle.

Now, I can say the new deformed shape is A B C dash D dash. I think this is there is no
confusion in it. Now, let us consider the fibre PQ or the distance y from the xx axis at a
distance y from the x x axis marked away from the centre of curvature as shown in the figure.
The initial length of this fibre will be R + y of d ϕ call the equation number 1.

Now, when M is applied the length of the fibre changes to PQ dash you can see here this has
become PQ dash this has become PQ dash Q has moved to Q dash. So, the new length of the
fibre or increase in length of the fibre which is the consequence of the applied moment M is
given by y + e d ϕ equation 2, which I call as QQ dash from the geometry can I say QQ dash
is y + e of Δd ϕ.

(Refer Slide Time: 19:01)

754
Having said this let e be the distance of neutral axis from the centroidal axis. Please note
centroidal axis sorry neutral axis is moving towards the centre of curvature. You may ask me
a question how this has happened why it cannot move away from the curvature very
interesting. Since M is applied such that it tends to decrease the curvature neutral axis will
move towards the curvature.

So, we can also put a note M is applied such that radius is closed we are decrease the
curvature. Now, let us write down the equation first here strain epsilon is y + e Δd ϕ change
in length by original length. See here original length is R + by d ϕ change in length is this.
So, I have written a strain equation call this equation as 3.

(Refer Slide Time: 21:23)

Let us assume that longitudinal fibers do not elongate we are only talking about the cross
section. So, I can write stress as E into del d ϕ by d ϕ of y + e by R + y can I write this
equation 4.

Let us see the nature of the stress, I want to copy this figure. Let me project this extrados
point an intrados point and the neutral axis point at any specific reference frame. The
reference frame may not be at the centre there is a reference frame, it need not be at the
centre it can be anywhere I am just putting the point. So, for your convenience if you really
want we can also do it the other way let us stop this, let us put the reference frame somewhere
here and let us project this for our convenience.

755
So, this is my neutral axis. So, let me draw the figure. Now, we know this is going to be
tensile is going to be compressive and the stress is going to be non-linear. So, this is tensile,
this is compressive and this variation will be hyperbolic its non-linear and let us for our
understanding we already know that this is the convex side which is extrados and this is the
concave side which is intrados M is applied a specific style. This is the stress variation and
the stress intensity at any distance y at any distance y is given by this equation and is
non-linear.

(Refer Slide Time: 25:21)

Now, let us say for equilibrium condition total compressive force should be identically equal
to the total tensile force. Now, since the average stress on the concave side on the concave
side is more than the convex side neutral axis shifts towards the centre of curvature. So, stress
dA for the area A should be 0 equation 5 let us substitute equation 4 this is the equation 4 in
5. So, substituting equation 4 in equation 5 we get what is equation 4 ?

See integral E d ϕ by d ϕ y + e + R + y dA should be set to 0.

756
(Refer Slide Time: 27:18)

Which means E d ϕ by d ϕ integral y + e by R + y of dA should be set to 0 we call this

equation number 6. Now friends since E d ϕ by d ϕ cannot be set to 0 and this is a constant
quantity I must. So, one should accept the integral y + e by R + y dA should be only set to 0
equation 7.

So, let us impose this condition. Now, I split this integral into two parts this integral can be
split into two parts as y by R + y dA + e times of 1 by R + y dA is set to 0 equation 7 a. So, I
call this as integral I 1 an integral I 2.

(Refer Slide Time: 29:00)

757
Let us pick up I 1, I 1 is integral y by R + y of dA this integral has the dimension of area
because this is over the area A. So, I could say now this is identically equal to mA where m is
a constant this depends on the shape of the cross section.

The quantity mA is called modified area of cross section. Why modify? The original area got
modified because of application of m in a specific style that is a m is closing the curvature
friends. So, we have got a very interesting integral that y by R + y dA over the area A is
identified as some factor of A which is a cross section property depends purely on the cross
section. So, m should be arrived now for various possible cross sections that can be used in
curved beams its a geometric property.

We will do that we will do some illustrative examples to work out m for different cross
sections.

(Refer Slide Time: 31:30)

Having said this now let us look into the second integral. Second integral I 2 was given by e
times of 1 by R + y dA for the area A which is now equal to e times of R + y - y by R times
of 1 by R + y dA can I write like this of the area A which can be said as e by R, R + y by R +
y dA integral A - y by R + y dA over A. Friends, if you recollect turn back your notes this
particular term is known what is this term? You know this is actually equal to mA.

So, can I now write this as e by R A - e by R mA can I write like this. This is my I 2 integral
2 am I right? So, I have both the integrals now integral 1 I have integral 1 I have this integral

758
1 which is simply mA and integral 2 also I have. Now can you substitute and write equation 7
again what is 7 equations? Can I write equation 7 again? Let us rewrite I will copy these
equations better selection tool.

(Refer Slide Time: 33:47)

Now, substituting for integral I 1 and integral I 2 7 a now becomes mA + e times of I 2 is

actually this value A by R - e mA by R am I right? e mA by R which is now set to 0 that is
what we want is it not set to 0. So, if you do that I can cancel this I know I can write m + e by
R - e m by R is 0. Friends, please note m in the above equation does not represent the
moment. Please understand the common confusion does not represent the moment it
represents or refers to modified area important.

So, we have derived this equation closed form I hope there is no confusion in this let us go
ahead.

759
(Refer Slide Time: 35:36)

Having said this let us say now m is equal to see here m is equal to let us rewrite this equation
or I can say it was m is equal to e times of m by R - 1 by R can I say that? Simplifying so, m
is equal e by R of m - 1 or e which is the shift of neutral axis from the centroidal axis given
by m by m - 1 of R equation 8 that is a very interesting relationship we have.

If I know the modified sectional area if I know the radius of curvature of the beam I can
always find what is the shift of neutral axis. Friends is very interesting now the shift of
neutral x also becomes part of the geometric property because m is a modified area m is a
modified area or modification in the area because of m capital M applied on the curved beam
and e is also now become a geometric property.

760
(Refer Slide Time: 37:09)

Having said this further, m + e by R - e m by R is 0 we have this equation here just see same
equation I am writing m + e by R same equation rewritten. So, now, we can also say m 1 - e
by R is equal to - e by R. So, m can also be said as - e by R by 1 - e by R can I say this ?
Which can be said as - of e by R by R - e by R. So, therefore, I will get m as - of e by R - e
equation 9 is another form to find m or to find e.

This is another form of equation. Having said this let us move forward to work out the
stresses. So, I will draw the figure again let us say we have this cross section with us I will
draw I will copy this figure.

761
(Refer Slide Time: 39:27)

For this figure let me plot the stress distribution. It is my reference plane, and the neutral axis.
So, I should say this is my stress distribution and let us also mark this is convex side, this is
concave side.

That is what we also marked earlier. The bottom is concave side upper is convex side or
extrados and intrados we have marked the same thing here again. Let us now write the
equation which we just now derived stress is given by E times of Δd ϕ by d ϕ of y + e by y +
R is what we already wrote let me call this number of equation as 4 let me mark it here as 4 I
am marking here as 4 because the same equation here written.

We also have e as m by m - 1 of R. I think this equation was 8 let me mark this equation as 8
m is also now given by - of e by R - e which is equation 9. The stress distribution looks like
this and this subjected to a moment such that the curvature is decreased in that we have to
remember very clearly.

762
(Refer Slide Time: 42:21)

So, now the cross section undergoes a non-linear stress distribution which is hyperbolic.
Now, the applied moment M should be equal to the resisting moment. So, we say over area A
into y should be equal to M we call equation number 10. So, now, let us substitute equation 4
in equation 10 we get equation 4 is here just wrote it for you.

So, now, M becomes E del d ϕ by d ϕ integral y + e R + y dA sorry y dA am I right? Just

check this equation 4 and I have one y here y d we call equation number 11 which can be
now written as E times of del d ϕ by d ϕ integral y square + y e by R + y of dA we will calls
equation 11 a. So, we call this as integral I.

763
(Refer Slide Time: 44:32)

Now the integral I is integral y square + y e by R + y dA over the area which is identically
split into two parts integral y square dA by R + y over a + e times of y by R + y dA over A.
We call this integral as I 1 thus integral as I 2 where I 1 is integral y square dA by R + y for
the area A which is y - R y by R + y of dA over A is check.

This is equal is it not? So, I 1 now becomes integral y dA for A - R times of y by R + y dA I

1 is further split.

(Refer Slide Time: 46:03)

764
So, friends, we will continue this discussion in the next lecture, I want you to evaluate this
integral. So, I leave a small homework. So, in this lecture we extended to learn curved beams
with large initial curvature. We are in the process of finding out the equation to obtain the
stress and its distribution across the section of a curved beam under large initial curvature.
There is a small homework given to you evaluate the integral I. We will discuss in the next
lecture.

Thank you very much and have a good day bye.

765
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 46
Curved beam with large curvature - 2

(Refer Slide Time: 00:21)

So, friends, welcome to the lecture-46 of the course Advanced Steel Design. We are now
going to continue to learn Curved beams with large initial curvature. I will put this II. So, in
this we are going to find out the expression for stress at intrados and stress at extrados.

766
(Refer Slide Time: 00:59)

So, in the last lecture we said that the integral I is

2 2
𝑦 +𝑦𝑒 𝑦 𝑦
𝐼 =∫ 𝑅+𝑦
𝑑𝐴 = ∫ 𝑅+𝑦
𝑑𝐴 + 𝑒 ∫ 𝑅+𝑦
𝑑𝐴

2
𝐼 =∫
𝑦
𝑅+𝑦
𝑑𝐴 = ∫ 𝑦 − ( 𝑅𝑦
𝑅+𝑦 )𝑑𝐴
Now this can now become 𝐼1 can become split as 2 integrals

𝐼1 = ∫ 𝑦𝑑𝐴 − 𝑅 ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
This is where we left and I hope you would have done the integration by this time and
substituted back in the original equation you must have got the I value. Let us continue from
this point.

767
(Refer Slide Time: 02:31)

So, now integral y dA for the entire area A should be 0. Why? We are assuming that the
geometric axis and neutral axis coincide. So, I will be 𝐼1 + 𝐼2 which is. So, this goes away.

∫ 𝑦𝑑𝐴 = 0

𝐼 = 𝐼1 + 𝐼2

𝐼 =− 𝑅 ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴 + 𝑒 ∫ 𝑦
𝑅+𝑦
𝑑𝐴

Furthermore, friends we already have this relationship with us. You recall it we have this

𝑦
∫ 𝑅+𝑦
𝑑𝐴 = 𝑚𝐴

is called as mA which has dimensions of area.

768
(Refer Slide Time: 04:26)

Having said this the integral now I becomes

2
𝑦 +𝑒𝑦
𝐼 =∫ 𝑅+𝑦
𝑑𝐴 =− 𝑅(𝑚𝐴) + 𝑒(𝑚𝐴)

This is the original integral here. So, now, I can say this integral is now going to be
− 𝑚𝐴(𝑅 − 𝑒)equation 12.

𝐼 =− 𝑚𝐴(𝑅 − 𝑒)

Now, substitute equation 12 this equation 11 we have substitute equation 12 in the moment
equation. What is the moment equation we had? Please turn your notes. M was equal to
∆𝑑ϕ
− 𝐸 𝑑ϕ
𝑚𝐴(𝑅 − 𝑒). So, now, I have substituted this integral back and got a new equation 1

here.

∆𝑑ϕ
𝑀 =− 𝐸 𝑑ϕ
𝑚𝐴(𝑅 − 𝑒)

∆𝑑ϕ 𝑀
𝐸 𝑑ϕ
=− 𝑚𝐴(𝑅−𝑒)

So, please understand this equation; this is the applied moment and this is cross sectional
property. So, do not get confused with the notation M.

769
(Refer Slide Time: 07:26)

𝑒
We also know that m is given by − (𝑅−𝑒)
I think this is equation number 9, please check that.

So, I have somewhere mA here let me substitute that.

𝑒
𝑚 =− (𝑅−𝑒)

∆𝑑ϕ 𝑀
𝐸 𝑑ϕ
=− 𝑚𝐴(𝑅−𝑒)

∆𝑑ϕ 𝑀 𝑀
𝐸 =− =
𝑑ϕ −( 𝑒𝐴
(𝑅−𝑒)
.(𝑅−𝑒) ) 𝐴𝑒

∆𝑑ϕ 𝑀
𝐸 𝑑ϕ
= 𝐴𝑒

The small m have gone is substituted back is equation 13.

770
(Refer Slide Time: 09:19)

So, substitute this in equation for the stress. Substitute in equation for final stress. It is the
possibly equation 4 you can check that. So, the stress is now given by

σ=
𝑀
𝐴𝑒 ( 𝑦+𝑒
𝑦+𝑅 )
𝑒 = ( 𝑚
𝑚−1 )𝑅
⎡ 𝑦+( 𝑚−1 )𝑅 ⎤
𝑚
𝑀 (𝑚−1)
σ= 𝐴𝑅 𝑚 ⎢ 𝑦+𝑅 ⎥
⎣ ⎦

(Refer Slide Time: 10:33)

771
 𝑀 (𝑚−1) (𝑚−1)𝑦+𝑚𝑅
σ= ⎡ (𝑚−1)(𝑦+𝑅) ⎤
𝐴𝑅 𝑚 ⎣ ⎦
𝑀 1 𝑚(𝑦+𝑅)−𝑦
σ= ⎡ ⎤
𝐴𝑅 𝑚 ⎣ (𝑦+𝑅) ⎦

σ=
𝑀
𝐴𝑅
1−
1
𝑚 ( 𝑦
𝑅+𝑦 )
The classical equation we have for curved beams which is equation 15.

(Refer Slide Time: 12:51)

σ=
𝑀
𝐴𝑅
1−
1
𝑚 ( 𝑦
𝑅+𝑦 )
Let us copy this equation to the next screen. Where we know where m is the applied moment
in such a manner it reduces the curvature. A is an area of cross section of the membrane.
Sigma is the tensile or compressive stress at a distance y measure from the centroidal axis.
Now remember y is measure from the centroidal axis not from the neutral axis. That is a
common confusion people will have.

Generally in classical equation of bending for finding out stresses the distance of extreme
fiber is measured from the neutral axis, but here it is from central axis. Please understand this.
So, very important deviation we have. Then of course, R is a radius of initial curvature we
should say of the unstressed beam before the M is applied the moment is applied; original
initial curvature.

772
Friends this equation 15 is classically known as Winkler back equation; equation 15 is
classically known as Winkler back equation. Winkler back formula Winkler back equation.
So, I think we have derived this stress to be found out at any point distance y from the
centroidal axis. If I know the geometric properties of the curved beam and the value of the
applied moment acting on the curve.

(Refer Slide Time: 16:00)

σ=
𝑀
𝐴𝑅 (1 − ( 1
𝑚
𝑦
𝑅+𝑦 ))
Let us revise some sign conventions for curved beams. Let us say the equation is very generic

stress is given by
𝑀
𝐴𝑅 (1 − (
1
𝑚
𝑦
𝑅+𝑦 )). Now y is measured from the centroidal axis it is
negative when measured towards the concave side. It is positive when measured towards the
convex side. Now one will have a confusion what is the convex and concave side.

So, in a given curved beam this is convex side called extrados. This is concave side called
intrados from the centroidal axis xx when you move towards this y is positive. When you
move towards the convex side.

773
(Refer Slide Time: 18:04)

Let me that is clear, towards the convex side towards the concave side. So, let us apply this
condition and try to find out the equations for introduction next class.

(Refer Slide Time: 18:32)

We also have an alternate equation for stress. Stress is also given by

σ=
𝑀
𝐴𝑒 ( 𝑦+𝑒
𝑦+𝑅 )

774
If you remember we derived this equation this was equation number 14 we derived this. So,
using this stress is intrados will be

σ𝑖 =
𝑀
𝐴𝑒 ( ) 𝑒−ℎ𝑖
𝑅−ℎ𝑖

Look at the figure hi is the depth of the intrados point measured from the central axis. I can
rewrite this equation as which is equal to

σ𝑖 =−
𝑀
𝐴𝑒 ( ) ℎ𝑖−𝑒
𝑟𝑖

I can I write this as Ri negative indicates this is compression. So, intrados will have
compression compressive stress.

(Refer Slide Time: 20:12)

σ𝑜 =
𝑀
𝐴𝑒 ( 𝑦+𝑒
𝑅+𝑦 )
I am looking for the extrados point. So, I am substituting this as ℎ0 that is what I am doing.

σ𝑜 =
𝑀
𝐴𝑒 ( ) 𝑒+ℎ0
𝑅+ℎ0

σ𝑜 =+
𝑀
𝐴𝑒 ( ) ℎ𝑜+𝑒
𝑟𝑜

775
So, this is tensile because it is positive and extrados has tensile stress that is what we also
plotted is not it. The plot was like this. So, this is all compression and these are all tension. it
was 14 b. So, this was 14 a and this is 14 b.

(Refer Slide Time: 21:32)

Let us look at a limiting case. Curved beams with large initial curvature can also become a
straight beam when R tends to infinity, in such condition the stress expression should reduce

to the form ( )𝑦, there is a classical thing for a straight beam.

𝑀
𝐼

(Refer Slide Time: 22:40)

776
Let us show how this is happening. So, we know stress is

σ=
𝑀
𝐴𝑒 ( 𝑦+𝑒
𝑦+𝑅 )
𝑚𝑅
𝑒 =− 1−𝑚

𝑀 (𝑦+𝑒)(1−𝑚)
σ= −𝑚𝐴𝑅 (𝑅+𝑦)

(Refer Slide Time: 23:25)

Now, for a straight beam e is 0. That is there is no shift of neutral axis from the centroidal
axis shift is 0. Hence small m will become 0 as R is infinity therefore, stress now becomes

−
𝑀
𝑚𝐴𝑅 ( 𝑦
𝑅+𝑦 ).
σ =−
𝑀
𝑚𝐴𝑅 ( 𝑦
𝑅+𝑦 )
2
∫
𝑦 𝑑𝐴
𝑅+𝑦 (
=∫ 𝑦 −
𝑅𝑦
𝑅+𝑦 )𝑑𝐴

777
(Refer Slide Time: 24:49)

We also know integral y dA for the area a is 0. Hence this will now become

∫ 𝑦𝑑𝐴 = 0

=− 𝑅 ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
=− 𝑅(𝑚𝐴)

Hence the stress equation IS

σ =−
𝑀
𝑚𝐴𝑅 ( 𝑦
𝑅+𝑦 )

778
(Refer Slide Time: 25:41)

Which means stress is

σ=
∫
𝑀
2
𝑦
𝑑𝐴
( 𝑦
𝑅+𝑦 )
(𝑅+𝑦)

𝑀𝑦
σ= 2

(1+ )∫ ( 𝑦
𝑅
𝑦
1+ 𝑅
𝑦
)
𝑑𝐴

Now, since R is infinitely large the above expression will now reduce to the following form.

(Refer Slide Time: 26:51)

779
 𝑀𝑦
σ= 𝐼

2
𝐼 = ∫ 𝑦 𝑑𝐴

𝑀𝑦 2
σ is 𝐼
where I is 𝑦 dA second moment of area which is applicable to straight beams. Now

friends there is an important statement here; m is a sectional property which is to be

developed or which is to be estimated for different cross section shapes used as curved
beams. That is the job we have additional. Let us do that now for 1 or 2 and see how they can
be estimated.

(Refer Slide Time: 28:15)

σ=
𝑀
𝐴𝑅 (1 − ( 1
𝑚
𝑦
𝑅+𝑦 ))
So, we know we are looking for estimating m factor which is otherwise called as modified
area. How it is modified? You are applying a moment it changes the curvature area is getting
modified. So, where it is being used in the stress equation see the equation we have for the
curved beam 1 minus 1 by m y by R plus y.

So, this is the factor of the cross sectional area. So, it is a geometric shape dependent factor.
Furthermore, y is measured from the centroidal axis not from the neutral axis you must
remember this. There is a common confusion and a common mistake which people generally
make when estimate stresses.

780
(Refer Slide Time: 29:54)

So, we will take up an example of a rectangular section. Draw a rectangular section will say
this is my reference plane from where I am going to measure the curvature. Remember beam
is a curved beam the cross section is rectangular please do not get confused I should say
rectangular cross section. So, it has got the centroidal axis as usual the central axis will be
measured from o o at distance R and if I have this as my width and this as my depth h of the
cross section. We know this is indicated as 𝑟1 in our nomenclature and this is indicated as 𝑟2

in our nomenclature.

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
𝑑𝐴 = (𝑏 𝑑𝑣)

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
(𝑣−𝑅)
𝑚𝐴 = ∫ 𝑣
𝑑𝐴

And from the cg this is hi which is h by 2 and from the cg this is h o which is h by 2 again.
Now let us consider a strip at a distance v from here. Let the thickness of the strip be dv the
strip is measured at a distance y from the centroidal axis as usual y is positive and is

781
measured towards a specific extrados side we know that. So, the equation for mA is integral y
by R plus y dA for the area this is general equation we have.

Let us find out what is dA in this case the hatched area can I say the dA is b into dv? Can I
say that? Now mA becomes integral for the area a y by R plus y dA which I write this as can
I write y as v minus R from the figure can I write y I mean R plus y as v from the figure dA
for the area this has got plane of reference with which the curvature is being measured. I
think this is mA is given by this.

(Refer Slide Time: 33:23)

(𝑣−𝑅) 1
𝑚𝐴 = ∫ 𝑣
𝑑𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

𝑟2
𝑑𝐴
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑟1

𝑟2
𝑏𝑑𝑉
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑟1

𝑚𝐴 = 𝐴 − 𝑅𝑏 𝑙𝑛 ( ) 𝑟2
𝑟1

( )
ℎ
𝑅+ 2
𝑚𝐴 = 𝐴 − 𝑅𝑏 𝑙𝑛 ℎ
𝑅− 2

782
(Refer Slide Time: 35:15)

( )
ℎ
𝑅 𝑅+ 2
𝑚𝐴 = 1 − 𝐴
𝑏 𝑙𝑛 ℎ
𝑅− 2

This for a rectangular section.

(Refer Slide Time: 35:48)

Let us derive it for a circular section. Let me draw a circular section, let me draw the plane of
reference. Let us mark the critical points we say this is my centroidal axis as usual and that is
always measured distance R from the plane of reference let me mark the neutral axis at a

783
distance e from the centroidal axis towards the center of curvature let us mark a strip at a
distance y measured from the centroidal axis.

Let us say this strip is at a distance v from the plane of reference. Let us say the thickness of
the strip is dv. And let us say this is small R and this angle is θ.

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
𝑑𝐴 = 2(𝑟 cos 𝑐𝑜𝑠 θ)𝑑𝑦

𝑦 = 𝑟 sin 𝑠𝑖𝑛 θ

𝑑𝑦 = 𝑟 cos 𝑐𝑜𝑠 θ 𝑑θ

𝑑𝐴 = (2𝑟 cos 𝑐𝑜𝑠 θ)𝑟 cos 𝑐𝑜𝑠 θ 𝑑θ

2 2
𝑑𝐴 = 2𝑟 𝑐𝑜𝑠 θ𝑑θ

(Refer Slide Time: 39:20)

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
𝑣−𝑅
𝑚𝐴 = ∫ 𝑣
𝑑𝐴

1
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

784
 1
𝑚𝐴 = 𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

This is my mA.

(Refer Slide Time: 40:33)

So, now let us pick up this integral which will be

π
2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ
∫ 𝑣
= ∫ 𝑅+𝑦
𝑑θ
−π
2

but now the variable is changing θ is the variable. So, θ has to now vary from

π
2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ
∫ 𝑣
= ∫ 𝑅+𝑟𝑠𝑖𝑛θ
𝑑θ
−π
2

I am covering the whole section.

friends it is a very interesting integral which is dA by v integral a I will stop here. I want you
to evaluate this integral I will continue this in the next lecture.

785
(Refer Slide Time: 41:50)

So, friends in this lecture we understood the stress equation for beams with large initial
curvature. We understood that mA is a modified area is a cross section property. So, we need
to find this m factor for all cross sections which are commonly used for curved beams. So, we
found out this mA factor for rectangular beam. We also showed how the stress the equation
for curved beams get converges to a straight beam when R or the radius of curvature becomes
infinity.

Thank you friends, have a good day bye.

786
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 47
Modified area factor for curved section

(Refer Slide Time: 00:20)

Welcome to the lecture-47 of Advanced Steel Design course. We are continuing to discuss
about the m factor for different cross sections. So, in this lecture we are going to learn m
factor for circular section and few more geometric shapes.

787
(Refer Slide Time: 00:54)

So, this was the circular section what we considered. Look at the screen and we say that v is a
point where the strip is being measured from the plane of reference. And we have derived the
equation written the equation for dA and so on, continue from this point and work ahead.

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
𝑑𝐴 = 2(𝑟 cos 𝑐𝑜𝑠 θ)𝑑𝑦

𝑦 = 𝑟 sin 𝑠𝑖𝑛 θ

𝑑𝑦 = 𝑟 cos 𝑐𝑜𝑠 θ 𝑑θ

𝑑𝐴 = (2𝑟 cos 𝑐𝑜𝑠 θ)𝑟 cos 𝑐𝑜𝑠 θ 𝑑θ

2 2
𝑑𝐴 = 2𝑟 𝑐𝑜𝑠 θ𝑑θ

𝑦= 𝑣− 𝑅

𝑣= 𝑅 + 𝑦

788
(Refer Slide Time: 01:18)

So, we know that mA is ∫ ( 𝑦

𝑅+𝑦 )𝑑𝐴. Now, I can rewrite this as you know look at this figure y
can be simply said as 𝑣 − 𝑅. So, I replace it.

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴 = ∫ 𝑣−𝑅
𝑣
𝑑𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫
1
𝑣
𝑑𝐴

1
𝑚𝐴 = 𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

π π
2 2 2 2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ 2𝑟 𝑐𝑜𝑠 θ
∫ 𝑣
= ∫ 𝑅+𝑦
𝑑θ = ∫ 𝑅+𝑟𝑠𝑖𝑛θ
𝑑θ
−π −π
2 2

π
2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ
∫ 𝑣
= ∫ 𝑅+𝑟𝑠𝑖𝑛θ
𝑑θ
−π
2

789
(Refer Slide Time: 03:31)

𝑅
We will take this equation forward. So, let us say introduce a variable k as a ratio of this 𝑟
.
𝑅
Let us imagine and understand what is 𝑟
. Look at this figure, R is a radius of curvature of

the initial level of the unstressed curved beam whereas, small r is a radius of the cross section
of the circular section. So, there are two different things here. And you know R by r will be
very large number, because R compared to small r will be very large.

𝑘= ( ) 𝑅
𝑟

π π
2 2 2 2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ 2𝑟 𝑐𝑜𝑠 θ
∫ 𝑣
= ∫ 𝑅+𝑟𝑠𝑖𝑛θ
𝑑θ = ∫ (𝑘𝑟+𝑟𝑠𝑖𝑛θ)
𝑑θ
−π −π
2 2

π π
2 2 2 2 2
𝑑𝐴 2𝑟 𝑐𝑜𝑠 θ (1−𝑠𝑖𝑛 θ)
∫ 𝑣
= ∫ 𝑟(𝑘+𝑠𝑖𝑛θ)
𝑑θ =2r ∫ 𝑘+𝑠𝑖𝑛θ
𝑑θ
−π −π
2 2

Now, let us find the simplified value of this expression let us do that.

790
(Refer Slide Time: 05:44)

2 2
1−𝑠𝑖𝑛 θ 1−𝑘
𝑘+𝑠𝑖𝑛θ
= (𝑘 − 𝑠𝑖𝑛θ) + 𝑘+𝑠𝑖𝑛θ

π
2

( )⎤⎥⎦𝑑θ
2
1−𝑘
I=2r ∫ ⎡⎢(𝑘 − 𝑠𝑖𝑛θ) + 𝑘+𝑠𝑖𝑛θ
−π ⎣
2

π π
2 2 2
𝑘 −1
𝐼 = 2𝑟 ∫ (𝑘 − 𝑠𝑖𝑛θ)𝑑θ − 2𝑟 ∫ ⎡⎢ 𝑘+𝑠𝑖𝑛θ ⎤⎥𝑑θ
−π −π ⎣ ⎦
2 2

(Refer Slide Time: 08:57)

791
 π
2 π
2
𝐼1 = 2𝑟 ∫ (𝑘 − 𝑠𝑖𝑛θ)𝑑θ = 2𝑟(𝑘θ + 𝑐𝑜𝑠θ) −π
−π 2
2

𝐼1 = 2𝑟 𝑘(( π
2
+
π
2 ) + cos 𝑐𝑜𝑠 ( ) − cos 𝑐𝑜𝑠 ( ) ) π
2
−π
2

𝐼1 = 2𝑟𝑘π

This gets cancelled 2r of 𝑘π is my 𝐼1.

(Refer Slide Time: 10:07)

Let us try to find 𝐼2 which is

π
2 2
𝑘 −1
𝐼2 = ∫ ⎡⎢ 𝑘+𝑠𝑖𝑛θ ⎤⎥𝑑θ
−π ⎣ ⎦
2

π π π
2
𝑑θ
2

∫ ⎡ 𝑘+𝑠𝑖𝑛θ ⎤≡ ∫ ⎡ 𝑎+𝑏𝑠𝑖𝑛𝑥 ⎤ =
𝑑𝑥 2 ⎡ 𝑎tan𝑡𝑎𝑛 ( )+𝑏
𝑥
2 ⎤
2

⎢ ⎥
−π ⎣ ⎦ −π ⎣ ⎦ 2 2 2 2
2 2
𝑎 −𝑏 ⎣ 𝑎 −𝑏 ⎦ −π
2

( ( )+1
)
θ 2
2 ⎡ 𝑘tan𝑡𝑎𝑛 2 ⎤
= 2 ⎢ 2 ⎥
𝑘 −1 ⎣ 𝑘 −1 ⎦ −π
2

=
2
2
⎡
⎢
𝑘 −1 ⎣
( ) ( )
𝑘+1
2
𝑘 −1
−
−𝑘+1

𝑘 −1
2
⎤
⎥=
⎦
2
2
𝑘 −1
( ) π
2

792
(Refer Slide Time: 13:26)

So, now I can say integral dA by V is 𝐼1 + 𝐼2 that is what we have here.

𝑑𝐴
∫ 𝑉
= 𝐼1 + 𝐼2

2
𝑘 −1
= 2π𝑟𝑘 − 2𝑟 2
π
𝑘 −1

𝑅
𝑘= 𝑟

= 2π𝑟 𝑘 −( 2
𝑘 − 1 )
= 2π𝑟
( 𝑅
𝑟
− ( ) 𝑅 2
𝑟
− 1
)
= 2π𝑟
( 𝑅
𝑟
−
𝑅 −𝑟
2

𝑟
2
2

793
(Refer Slide Time: 15:12)

∫
𝑑𝐴
𝑉 (
= 2π 𝑅 − 𝑅 −𝑟
2
)
2

𝑑𝐴
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 − ∫ 𝑉

2
𝑚𝐴 = π𝑟 − 𝑅(2π) 𝑅 − ( 2
𝑅 −𝑟
2
)
𝑚 = 1−
𝑅(2π)
π𝑟
2 (𝑅 −
2
𝑅 −𝑟 )
2

𝑅 2
𝑚 = 1− 2 ( )
𝑟
+
2𝑅
𝑟
2 𝑅 −𝑟
2 2

that is m for the circular section. So, if we know the values of R and r one can find m. So, it is
a geometric property.

794
(Refer Slide Time: 16:52)

So, now I can find e as

𝑒 = ( 𝑚
𝑚−1 )𝑅
𝑑𝐴
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑉

𝑑𝐴
𝑚𝐴 = 𝐴 − 𝑅 ∫ 𝑉

𝑅 𝑑𝐴
𝑚 = 1− 𝐴
∫ 𝑉

795
(Refer Slide Time: 17:34)

And therefore, e is 𝑒 = ( 𝑚
𝑚−1 )𝑅
⎡ 1− 𝑅𝐴 ∫ 𝑑𝐴𝑉 ⎤
⎢ ⎥
𝑒 =⎢ ⎥𝑅
⎢ 1− 𝐴 ∫ 𝑉 −1 ⎥
𝑅 𝑑𝐴

⎣ ⎦

⎡ 1− 𝑅𝐴 ∫ 𝑑𝐴𝑉 ⎤
⎢ ⎥
𝑒 = 𝑅⎢ ⎥
⎢ − 𝑅𝐴 ∫ 𝑑𝐴𝑉 ⎥
⎣ ⎦

(Refer Slide Time: 18:35)

796
(Refer Slide Time: 18:44)

−𝐴
𝑒 = + 𝑅
𝑑𝐴
∫ 𝑉

𝐴
𝑒 = 𝑅 −
𝑑𝐴
∫ 𝑉

2
∫
𝑑𝐴
= 2π𝑟𝑘 − (
2𝑟 𝑘 −1 π )
𝑉 2
𝑘 −1

I am substituting for k.

∫
𝑑𝐴
𝑉 (
= 2π 𝑅 − 𝑅 −𝑟
2
)
2

797
(Refer Slide Time: 19:47)

𝐴
𝑒 = 𝑅 −
𝑑𝐴
∫ 𝑉

2
π𝑟
𝑒 = 𝑅 −⎛ ⎞
⎝ (
2π 𝑅− 𝑅 −𝑟
2 2
)⎠
⎡ 2 ⎤
𝑟
𝑒 = 𝑅 −⎢ ⎥
⎣ (
⎢ 2 𝑅− 𝑅2−𝑟2 ) ⎥
⎦

So, if you know R and r, you can find e. So, now, I know the modified cross section m, I
know how to locate the neutral axis from the centroidal axis because that is what e is, e is
offset of that. So, all are geometric properties now.

798
(Refer Slide Time: 20:56)

Let us do one more example. I will say this as example 3, example 1 is rectangular section,
example 2 is circular section. We will go for a T-section. So, let me draw the figure of a T
section. Let us say this is my plane of reference o-o this dimension is 𝑏1 and this dimension is

𝑏2 of the T section. Let us say, it has got a centroid axis passing through this point.

This is my cg. I locate a strips at a distance y measured from the cg away from the center. So,
I will call this value as 𝑟1, this value as 𝑟2. And of course, we know that this is R radius of

curvature to the centroid. Let me call this as 𝑟3. Now, let us cut a section let us say this is my

dv and this is at a distance v from the plane of reference.

So, you know the equation mA is given by integral y by R plus y dA for the whole area.

𝑦
𝑚𝐴 = ∫ 𝑅+𝑦
𝑑𝐴

𝑣−𝑅
𝑚𝐴 = ∫ 𝑣
𝑑𝐴

1
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

799
(Refer Slide Time: 23:55)

𝑟2 𝑟3
𝑑𝑉 𝑑𝑉
𝑚𝐴 = 𝐴 − 𝑅 ∫ 𝑏1 𝑉 − 𝑅 ∫ 𝑏2 𝑉
𝑟1 𝑟2

𝑚𝐴 = 𝐴 − 𝑅𝑏1𝑙𝑛 ( ) 𝑟2
𝑟1
− 𝑅𝑏2𝑙𝑛 ( ) 𝑟3
𝑟2

𝑚 = 1−
𝑅
𝐴
⎡𝑏 𝑙𝑛
⎢ 1
⎣ ( ) 𝑟2
𝑟1
+ 𝑏2𝑙𝑛 ( )
𝑟3
𝑟2
⎤
⎥
⎦
𝑚𝑅
𝑒 = 𝑚−1

800
(Refer Slide Time: 25:54)

𝑒 =
( 𝑅

⎣
𝑟

1
( ) ( ))
𝑟
𝑅 1− 𝐴 ⎡⎢𝑏1𝑙𝑛 𝑟2 +𝑏2𝑙𝑛 𝑟3 ⎤⎥
2 ⎦

𝑅⎡

⎣
𝑟2

1
( ) ( )
𝑟3
1− 𝐴 ⎢𝑏1𝑙𝑛 𝑟 +𝑏2𝑙𝑛 𝑟 ⎤⎥−1
2 ⎦

⎰ 1 ⎱
𝑒 = 𝑅 − + 1
⎱ 𝑅
𝐴 ⎢ 1
⎣ ( ) ( )
⎡𝑏 𝑙𝑛 𝑟2
𝑟1
+𝑏2𝑙𝑛
𝑟3
𝑟2
⎤
⎥
⎦
⎰

(Refer Slide Time: 27:48)

That is my e. So, I can locate the neutral axis. So, e is going to be

801
 𝐴
𝑒 = 𝑅 −
𝑏1𝑙𝑛( ) ( )
𝑟2
𝑟1
+𝑏2𝑙𝑛
𝑟3
𝑟2

So, friends you can see from this equation e m are all functions of cross section. They are
geometric parameters. So, for a given value let us say for example, I will just draw this figure
again.

(Refer Slide Time: 29:05)

For b 1 as 100 mm, b 2 as 20 mm, R as 1000 mm, r 1 as 500, r 2 as 520. So, you need to find
actually R capital R it is a cg. So, you must know this dimension, which will be actually I
should say r 2 plus y bar. You can find out this. For b 1 and b 2 and for this dimension which
is h bar which is taken as 70 mm, one can find R. Now, all dimensions are known one can
find m, one can find e.

So, once you locate e you will be able to locate the neutral axis and then you can find the
stresses and plot the stresses. So, friends in this lecture, we learnt how to find the modified
area property and the offset of the neutral axis from the centroidal axis for different cross
sections. We did for circular; we did for T-section. We already did for rectangular section
earlier for first principles.

I think there is no difficulty in handling these kinds of problems and my reference book also
gives you the MATLAB program to find out the stresses for various cross sections for curved
beams. I strongly promote that you should use this textbook and download the programs and

802
use MATLAB intensively for computing the stresses and the cross sections of these shapes
for curved beams of large initial curvature, which is Winkler Bach equation. Please do that.

Thank you very much. Have a good day, bye.

803
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 48
M factor for curved beams

(Refer Slide Time: 00:20)

Friends, welcome to the 48th lecture on Advanced Steel Design course. Here, we are going to
discuss about more on Curved Beams. We will also talk about finding out the stresses and
different cross sections of curved beams. So, we will talk about M factor, we will also learn
more about the crane hook analysis.

804
(Refer Slide Time: 01:07)

So, we will say example 1 for this lecture. We will take up a section, an unequal high section
is a very common case for curved beams are played with large moments. Write down the
properties of this. This becomes my plane of reference o-o and will mark the centroidal axis
somewhere here. We know this distance actually is always measured as R and we call this
dimension as 𝑏1, this as 𝑟1 and this as 𝑟2. And of course, this dimension is 𝑟3.

This is 𝑟4 and this is 𝑟3, this is 𝑏1, this is 𝑏2 and this is 𝑏3. This piece number 1 for us 2 and 3

for our reference. Then, we will cut a small strip which is measured a distance y from here,
we call this as dv where this distance is v from here. So, general expression for mA is given
as

𝑚𝐴 = ∫ ( 𝑦
𝑅+𝑦 )𝑑𝐴
𝑣−𝑅
𝑚𝐴 = ∫ 𝑣
𝑑𝐴

1
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑣
𝑑𝐴

805
(Refer Slide Time: 05:00)

So, now m A is equal to

𝑟 𝑟3 𝑟4
⎰2 𝑏1𝑑𝑣 𝑏2𝑑𝑣 𝑏3𝑑𝑣 ⎱
𝑚𝐴 = 𝐴 − 𝑅 ∫ +∫ +∫
⎱𝑟 𝑣
𝑟2
𝑣
𝑟3
𝑣 ⎰
1

So, I have got different components. Let me I think copy this figure its better and copy this
figure, put it here then I will write m A here. m A is going to be equation is dA which is A
minus R times of see we have integral over the entire area A, I am dividing the three parts.
So, R times of , I will do it here minus R times of let us do from integral 𝑟2 to 𝑟1. So, that

area will be 𝑏1 dv by v. See here, I am doing it for dA.

So, small strip cut here and that strip is going to vary from 𝑟1 to 𝑟2, that is what I am doing

plus integral 𝑟3 to 𝑟2, for this strip it is 𝑏2. Similarly, integral 𝑟4 to 𝑟3 and this is 𝑏3 which will

now become

⎰
𝑚𝐴 = 𝐴 − 𝑅 𝑏1 𝑙𝑛
⎱ ( ) 𝑟2
𝑟1
+ 𝑏2 𝑙𝑛 ( ) 𝑟3
𝑟2
+ 𝑏3 𝑙𝑛 ( )
𝑟4
𝑟3
⎱
⎰

𝑚 = 1−
𝑅
𝐴
⎰
𝑏 𝑙𝑛
⎱ 1 ( ) 𝑟2
𝑟1
+ 𝑏2 𝑙𝑛 ( ) 𝑟3
𝑟2
+ 𝑏3 𝑙𝑛 ( )
𝑟4
𝑟3
⎱
⎰
𝐴
𝑒 = 𝑅 − ⎰
𝑏 𝑙𝑛
⎱ 1 ( ) 𝑟2
𝑟1
+𝑏2 𝑙𝑛( ) 𝑟3
𝑟2
+𝑏3 𝑙𝑛( ) 𝑟4
𝑟3
⎱
⎰

806
So, I can easily calculate this.

(Refer Slide Time: 07:53)

So, therefore, friends if I have a general expression for rectangular sections e will be

𝐴
𝑒 = 𝑅 − 𝑛

( )
∑ 𝑏𝑛𝑙𝑛
𝑖=1
𝑟𝑛+1
𝑟𝑛

This is what I get the e is general expression. Let us take a triangular section.

(Refer Slide Time: 08:36)

807
So, we will draw a triangle plane of reference, triangle has value 𝑏1 here and this is 𝑟1 and

this value is termed as 𝑟2 and triangle has a cg somewhere here and we mark that distance

from here as R. Let us cut the strip at a distance y from here. Let the thickness of the strip be
dv and the strip be measured v from here. Let me draw the frustum of that separately, I will
mark this distance as x, because I want to find the area that is tricky here.

𝑏1
(𝑟2 − 𝑟1)~ 2
(𝑟2 − 𝑣)~ 2𝑥
𝑏1
𝑥
2 (𝑟2 − 𝑟1) = 2 (𝑟2 − 𝑣)
(𝑟2−𝑣)
𝑥 = 𝑏1
(𝑟2−𝑟1)

Let me mark that figure separately. So, anyway this getting converge to become a triangle.
So, this is x and this is 𝑏1, let us say this is 𝑟2 and this is 𝑟1. So, we now write a relationship

𝑟2 minus 𝑟1. The variation is for 𝑏1 by 2, is it not if you look at this triangle. Let us say I am

looking for this part, for a variation of 𝑟2 minus 𝑟1.

Now, we know that this strip is at v is at v. So, therefore, for 𝑟2 minus v, this value will be x

by 2. So, now, x by 2 into 𝑟2 minus 𝑟1 is 𝑏1 by 2 𝑟2 minus v. So, can I say x as 𝑏1 times of 𝑟2

minus v by 𝑟2 minus 𝑟1, that is my x; width of the strip is x.

808
(Refer Slide Time: 12:28)

Now, m A is

𝑑𝐴
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑉

𝑟2
(𝑟2−𝑣) 𝑑𝑉
𝑚𝐴 = 𝐴 − 𝑅 ∫ 𝑏1
𝑟1
(𝑟2−𝑟1) 𝑉

𝑟2
𝑅𝑏1
(𝑟2−𝑟1) 𝑟∫(𝑟2 − 𝑣) 𝑉
𝑑𝑉
𝑚𝐴 = 𝐴 −
1

𝑟 𝑟2
𝑅𝑏1⎰2 𝑑𝑉 ⎱
𝑚𝐴 = 𝐴 − 𝑟 −𝑟
( 2 1) ⎱𝑟∫ 𝑟2 𝑉
− ∫ 𝑑𝑉
𝑟
⎰
1 1

809
(Refer Slide Time: 13:53)

𝑚𝐴 = 𝐴 −
(
𝑅𝑏1
𝑟2−𝑟1
⎡
) ⎢⎣𝑟2 𝑙𝑛( ) (
𝑟2
𝑟1
− 𝑟2 − 𝑟1 ⎤⎥
⎦
)

) ( )
𝑅 ⎰ 𝑟2𝑏1 𝑟2 ⎱
𝑚 = 1− 𝑙𝑛 − 𝑏1
𝐴 ⎱ (𝑟2−𝑟1 𝑟1 ⎰

So, that is my m value for the triangular section.

(Refer Slide Time: 14:50)

Once I have a m value, I can find the e value using this relationship

810
 𝐴
𝑒 = 𝑅 − 𝑟2𝑏1
( )
(𝑟2−𝑟1) 𝑙𝑛
𝑟2
𝑟1
−𝑏1

that is my e to locate the neutral axis.

(Refer Slide Time: 15:22)

In the same logic, let us do for a trapezoidal section. This will be the plane of reference. Let
us say this is 𝑏1, this is 𝑏2, 𝑟1, this is 𝑟2, let us say this has got a cg somewhere here. So, this

becomes a cg axis which is measured R from the plane of the radius R.

Then, we will also then mark a strip at a distance y from the central axis. We call this as dv
and let the strip be cut at the distance v from the plane of reference. Now, I divide this into a
segment like this and I will extrapolate this separately. Then, I can write the similar triangle
equation for this which will be 𝑟2 minus 𝑟1 on the height gives me the variation as 𝑏1 minus

𝑏2.

(𝑟2 − 𝑟1)~(𝑏1 − 𝑏2)

(𝑟2 − 𝑣)~𝑥
(𝑟2 − 𝑟1)𝑥 = (𝑏1 − 𝑏2)(𝑟2 − 𝑣)
(𝑟2−𝑣)
𝑥 =
(𝑟2−𝑟1) (𝑏1 − 𝑏2) + 𝑏2

811
So, if it is 𝑟2 minus v, let us call this variation as x, 𝑟2 minus 𝑟1 into x is 𝑏1 minus 𝑏2 into 𝑟2

minus v. Therefore, x will be 𝑟2 minus v by 𝑟2 minus 𝑟1 of 𝑏1 minus 𝑏2 plus 𝑏2, because I am

looking for this distance is x. So, that is my x value.

(Refer Slide Time: 18:49)

Having obtained this, I can now use the standard relation m A is

𝑑𝐴
𝑚𝐴 = ∫ 𝑑𝐴 − 𝑅 ∫ 𝑉

𝑟2

𝑚𝐴 = 𝐴 − 𝑅 ∫⎡⎢𝑏2 +
𝑟1⎣
( )(
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2 − 𝑣 ⎤⎥ 𝑉
⎦
𝑑𝑉
)
𝑟 𝑟2 𝑟2
⎡2 ⎤
𝑚𝐴 = 𝐴 − 𝑅⎢∫ 𝑏2
⎢𝑟
⎣1
𝑑𝑉
𝑉
+∫
𝑟1
( )
𝑏1−𝑏2
𝑟2−𝑟1
𝑑𝑉
𝑟2 𝑉 −∫
𝑟1
( )
𝑏1−𝑏2
𝑟2−𝑟1
𝑑𝑉⎥
⎥
⎦

𝑚𝐴 = 𝐴 − 𝑅⎡⎢𝑏2𝑙𝑛
⎣ ( ) ( ) ( ) ( )(
𝑟2
𝑟1
+
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2𝑙𝑛
𝑟2
𝑟1
−
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2 − 𝑟1 ⎤⎥
)⎦

𝑚𝐴 = 𝐴 − 𝑅⎡⎢𝑏2𝑙𝑛
⎣ ( ) ( ) ( ) (
𝑟2
𝑟1
+
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2𝑙𝑛
𝑟2
𝑟1
− 𝑏1 − 𝑏2 ⎤⎥
⎦
)

812
(Refer Slide Time: 21:48)

𝑚𝐴 = 𝐴 − 𝑅⎡⎢𝑏2𝑙𝑛
⎣ ( ) ( ) ( )
𝑟2
𝑟1
+
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2𝑙𝑛
𝑟2
𝑟1
− 𝑏1 − 𝑏2 ⎤⎥
( )⎦

So, I can say which will be equal to

𝑚 = 1−
𝑅
𝐴 ⎢ 2
⎣ ( ) ( ) ( )
⎡𝑏 𝑙𝑛 𝑟2
𝑟1
+
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2𝑙𝑛
𝑟2
𝑟1
− 𝑏1 − 𝑏2 ⎤⎥
( )⎦
𝐴
𝑒 = 𝑅 −
⎡𝑏 𝑙𝑛
⎢ 2
⎣ ( )( ) ( )(
𝑟2
𝑟1
+
𝑏1−𝑏2
𝑟2−𝑟1
𝑟2𝑙𝑛
𝑟2
𝑟1
− 𝑏1−𝑏2 ⎤⎥
)⎦

that is my shift of the neutral axis from the centroidal axis which is equation 4 in our original
derivation.

So, friends if you call this as equation 4, in this equation if 𝑏2is 0; what is 𝑏2? See, the figure

if 𝑏2 is 0, we get a triangle. This equation will reduce to that of triangle. If 𝑏2 and 𝑏1 are b,

see this figure if they are b reduced to a rectangle; this equation will reduce to a rectangular
section. You can check that, I will leave it to you. Please check that and check the derivation
carefully.

813
(Refer Slide Time: 24:10)

So, there are some comments which we need to highlight very clearly before we do some
more numerical. e is always measured from the centroidal axis and e it is the distance of the
neutral axis from the centroidal axis, positive value indicates that e is measured towards the
center of curvature that is very important friends. So, now, for curved beams there are some
more simplified equations given in the literature. Let us also try to appreciate them and see.

(Refer Slide Time: 25:31)

So, Wilson and Querean have given simplified form of equations to find the stresses in
extrados and intrados of different cross sections and say the equation is given by σ is equal to

814
k times of applied moment h by I. So, they say the stress you have to use 𝑘𝑖, say 𝑘𝑖 is a factor

for intrados which is ℎ𝑖 minus e by 𝑟𝑖 divided by M ℎ𝑖 by 2 I, 𝑘𝑜 is a factor used for extrados

which is ℎ𝑜 by e by 𝑟𝑜 divided by M ℎ𝑜 by 2 I. So, now I will give you the values of these

tables as recommended by these researchers for our usefulness.

𝑀ℎ
σ= 𝑘 𝐼

ℎ𝑖−𝑒
𝑟𝑖
𝑘𝑖 = 𝑀ℎ𝑖
2𝐼

ℎ𝑜−𝑒
𝑟𝑜
𝑘𝑜 = 𝑀ℎ𝑜
2𝐼

(Refer Slide Time: 27:04)

Let us say I draw the shapes here, this is my plane of reference o o. I write here plane of
reference. So, let us say circular section and for elliptical sections because, elliptical node
sections are also not coming very common now, because they have lot of advantages on
eddies.

We will talk about this shape in some other future lectures. So, let us say we know if this is
my centroidal axis, we know the measurement of the centroidal axis always R from the plane
of reference. This is also R from the plane of reference and from the cg we call this as h is
also h, because it is symmetric and the major axis is h minus axis h again. So, he is trying to

815
work out only in the stresses in the intrados and extrados, only here, not anywhere in
between.

So, he said the values of R by h and he has given the coefficients 𝑘𝑖 and 𝑘𝑜, then he also

compare e when the beam is under pure bending. So, let us see what are these values given by
researchers. So, at different intervals of 1.2 1.4 6 a 2, then 3, 4 and 6; 𝑘𝑖 3.41 2.40 1.96 1.75

1.62 33 23 14; k o 0.54 60 65 68 71 79 84 89.

e value is also given in terms of R when the beam is under pure bending. Its says it is 0.224 R
0.151 R 0.108 0.084 0.009 0.030 0.016 and 0.0070. For rectangular sections, this table is
again given in a different format, becomes the plane of reference.

(Refer Slide Time: 30:27)

Say o o, we draw a rectangle centroidal axis measurement R from the plane of reference. Let
us say this is h is also h. So, R by h 𝑘𝑖 𝑘𝑜 and e when the beam is under pure bending. So, 1.2,

𝑘𝑜 0.305 R 0.204 R 0.149 R 0.112 0.090 0.041 0.021 0.00 93.

So, this is not very interesting reference, my own book advanced steel design written for
CRC. Please refer this book to get more details about the experimental studies conducted by
the authors and how they obtain these coefficients. Let us do a couple of more examples on
circular sections.

816
(Refer Slide Time: 33:03)

Let us say we have a curved beam, we have an applied example now. So, let us say there is a
curved beam as shown in the figure here. Let me mark the gray. Let us have it like, this is
better. Let us mark the cross section, this is 40, this is 20. Let us say this radius is also 40
point, this is central axis. Now, I call this as my o plane of reference because the curve is
measured from here.

And, for the center, I will mark R as usual. I am doing an applied example. Now, let us draw
this cross section separately. So, the cross section has got 40 parallel to plane, see 40 mm is
parallel to o o. So, let me draw the line o o, let us say from the edge from the edge see from
here, see from here this cross section is 20 and this is 40.

The cg from here is R which is 60. Because this is 40 and furthermore another 20 is from 60
point. So, we now say 𝑟1 is 40 and 𝑟2 will be 80 and b so, we say b is 20, 𝑟1 is 40 and 𝑟2 is 80

and R is 40 plus 20 which is 60 mm. Let us say this is subjected to a low P without 20 kN.

Now, this 20 kN P will try to compress this ring. So, M caused by this load will tend to close
the curvature. It will reduce radius. Under that condition extrados will be under tension, on
intrados will be under compressive stress. So, compression is negative for us.

817
(Refer Slide Time: 38:01)

Let us try to find out the direct stress. What will be the direct stress? If you do not apply a
curved beam concept which will be simply P by A which will be minus 20 into 3 by 40 into
20. Why minus? Compression. So, minus 25 Newton per mm square, we say negative stands
for compression.

3
𝑃 −20×10 𝑁
𝐴
= 40×20
=− 25 2
𝑚𝑚

60
𝑀 = 𝑃𝑅 =+ 20× 3 = 1. 2𝑘𝑁𝑚
10

So, we can find the bending stress. Now, M will be actually P into R which will be plus
because it is tending to close the curvature. So, 20 into 60 by 10 power 3 which is 1.2
kilonewton meter. Why it is positive? The moment is positive because it decreases the
curvature, that is the reason. Now, we want to find for the rectangular cross section, what is
the modified area m A.

818
(Refer Slide Time: 39:27)

You have the equation m is given by 1 minus R by A b ln of 𝑟2 by 𝑟1, we have standard

equation for this. So, 1 minus R by 40 into 20 that is A, b is 20 see we already said that b is
20 of ln of 𝑟2 by 𝑟1 which will be 1 minus 60 by 800 20 ln r 80 by 40 which will become

minus 0.0397.

𝑚 = 1−
𝑅
𝐴
𝑏𝑙𝑛 ( )
𝑟2
𝑟1

𝑚 = 1−
60
(40×20)
(20)𝑙𝑛 ( )𝑟2
𝑟1

𝑚 = 1−
60
(80)
(20)𝑙𝑛 ( ) =− 0. 0397
80
40

𝑚𝑅 −0.0397(60)
𝑒 = 𝑚−1
= −0.0397−1
=+ 2. 291𝑚𝑚

So, e will be mR by m minus 1 which will be minus 0397 into 60 by minus 0397 minus 1
which will become plus 2.291 m. So, there is a shift of neutral axis from the centroidal axis
by about 2.3 mm though it is small, but look at the cross sections only 40 by 20 point. So, for
the transaction is significant.

819
(Refer Slide Time: 40:49)

𝑒 =+ 2. 291𝑚𝑚

(ℎ𝑖−𝑒)
(σ𝑏)𝑖 =− 𝑀
𝐴𝑒 𝑟𝑖

( )𝑖
σ𝑏 =−
1.2×10
(40×20)(2.291)
(20−2.291)
40

(σ𝑏)𝑖 =− 289. 87 𝑚𝑚𝑁 2

So, e is plus 2.291 mm. It means positive indicates that neutral axis is shifting towards the
center of curvature. So, I have a cross section which is 20, 40, there is a central axis, this is
my cg. I mean neutral axis is shifted down by a distance 2.291 m. And, for finding out the
stresses for intrados, we know the equation which is minus M by A e ℎ𝑖 minus e by 𝑟𝑖 which

is compressive. Remember, h is always measured from the centroidal axis, not from the
neutral axis please understand that.

So, let me substitute that. So, σ𝑏 bending stress into dos will be minus 1.2 10 power 6 by 40

into 20 of 2.291 of 20 minus 2.291 by r i which is 40. So, I can get this value which is minus
289.87 complex this intrados. Let us do this for extrados. So, bending stress at extrados that is
this point which will be plus M by A e ℎ𝑜 plus e by 𝑟𝑜 tensile. So, I should say plus 1.2 10

power 6 by 40 into 20 into 2.291 of 20 plus 2.291.

820
Because, you know this is 20, divided by 80 that is 𝑟𝑜, we already have it here, on this stress

is plus 182.43 tensile, extrados tensile. So, I have the hook coming here the ring, the ring is
subjected to compression. So, this is the moment. So, extrados will be in tension and intrados
will be in compression. So, we have these values with this.

(Refer Slide Time: 45:06)

σ = σ𝑑𝑖𝑟𝑒𝑐𝑡 + σ𝑏𝑒𝑛𝑑𝑖𝑛𝑔

So, therefore, the total stress will be stress direct plus stress bending. So, if you want to find
at point A which is minus 25 minus 289.87, stress at B which will be minus 25 plus 182.43.
So, B is any point on extrados and A is any point on intrados. I can find these values in
newton per mm square.

This will be compressive, this will be tensile. So, one can also see here, the stress that
extrados and strength intrados are not equal which means the stress variation is also
non-linear. So, you cannot predict that, you have to find the exact fiber distance and then
keep on working out the stresses.

821
(Refer Slide Time: 46:35)

So, friends in this lecture, we learnt the modified area equation for different sections. We also
learnt the stresses to find in extrados and intrados by researchers. We also did a applied
example of a ring under the compression loop. So, in the next lecture, we will do some more
examples on rings and crane hooks etcetera.

We will use MATLAB program and try to find out the use of winkler-bach equation in large
initial curvature problems and try to solve and find the stresses in the sections across the
curved beam.

Thank you very much and have a good day. Bye.

822
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 49
Crane hook

(Refer Slide Time: 00:22)

Welcome friends to the lecture-49 of Advanced Steel Design course. In this lecture, we are
going to do more examples on curved beams, we will also do an example on Crane hook, we
will also use MATLAB program for solving the problem. Friends, the MATLAB coding for
these numerical are available in my standard textbook Advanced Steel Design published by
CRC is given in the list of references of this course.

I request you to please go through that download the program and use it freely for your
MATLAB and MATLAB extensively supports NPTEL courses thoroughly. You can also get a
free training for using MATLAB programs with the MATLAB incorporation. Please
download the software, use it freely so that academic purpose. We encourage very strongly
that you should use MATLAB program, write your own code.

But, in this course we are able to share the code with you can make some small modifications
in the code so that you can make it more I mean compact and look faster to execute. So, we

823
will continue with the discussion on working on problems for the curved beams we said
Winkler Bach equation is a very good compact set of solution for solving a curved beam.

(Refer Slide Time: 02:02)

Let us take an example and try to solve a problem for a curved beam. So, let us say we have a
circular proving ring which is used for loading purpose. Proving ring is a very common thing
used in experimental studies which you must be aware. So, I have a proving ring which is
circular in shape. So, this is my centroidal axis of this and let us say this is subjected to a
downward force of 100 kN acting on the proving ring which is circular in shape.

So, let us also mark the centroidal axis of this curved beam which has a radius R from the
centre and we call this radius as 100 mm. If you cut a cross section of this curved beam the
cross section looks like a T section, if this becomes my plane of reference which is O, O and
let us say the neutral axis is located here and the centroidal axis is somewhere here as marked
on the figure.

So, our job is to even find out this axis, but let us mark the eccentricity of the neutral axis
from the centroidal axis as e which is measured towards the center of curvature. Let us mark
the dimensions of this particular T section. We call this as 𝑏1 which is 120 mm all dimensions

are mm. This has a thickness of 20. Let us call this as 𝑟1 from the plane of reference and let

this be 100.

824
And, we know this is 𝑟2 which is 120 and we call this as 𝑟3 which is 270 which means this

dimension is 150, and this thickness is said as 𝑏2 which is 20. So, we have all the dimensions

and we will say this is my radius which is R. R is to be computed, let us say this is 100. This
is 100. R need to be computed. The intrados is 100. R need to be computed.

So, these are the data we have and we want to find the stresses that intrados and extrados of
this curved beam.

(Refer Slide Time: 06:40)

So, our first step is to find out the centroidal axis locate the centroidal axis. Let us copy this
figure take it to the next screen and we will do the calculation. So, let me call this distance as
y bar. Now, I want to find y bar. So, y bar is given by sigma a y bar by sigma a we will divide
this into two parts, let us do that. So, this is piece number 1 and this is piece number 2.

So, y bar is now equal to 120 into 20 into 10. I take this as my reference plus 150 into 20 into
95 divided by 120 into 20 plus 150 into 20. This gives y bar as 57.22 mm. Therefore, I can
now write R value as 157.22 mm. Therefore, R is 100 plus 57.22, 157.22 mm.

Now, we have an equation to find the modified area factor m. m is given by 1 minus R by A;
refer to the equation for rectangle series of rectangles we already have a generic form I am

writing the general equation 𝑏1𝑙𝑛 ( )

𝑟2
𝑟1
+ 𝑏2𝑙𝑛 ( )
𝑟3
𝑟2
. So, let us say this is my 𝑟3, this is my 𝑟1,

this is my 𝑟2. Let us substitute for this. 1 minus 157.22 by 5400, possibly if you add up this if

825
you add up this should be equal to 5400, as a total area. multiplied by 𝑏1 is 120 you can see

from the figure.

Σ𝑎𝑦
𝑦= Σ𝑎

(120×20×10)+(150×20×95)
𝑦= (120×20)+(150×20)

𝑦 = 57. 22𝑚𝑚

𝑅 = 100 + 57. 22 = 157. 22𝑚𝑚

𝑚 = 1−
𝑅
𝐴
⎰
𝑏 𝑙𝑛
⎱ 1 ( ) 𝑟2
𝑟1
+ 𝑏2𝑙𝑛 ( ) 𝑟3
𝑟2
⎱
⎰

𝑚 = 1−
157.22
5400 {120𝑙𝑛( ) + 20𝑙𝑛( )}
120
100
270
120

𝑚 =− 0. 109

which gives me m value as minus 0.109. Let me write down this value here for our reference
m is minus 0.109. Having said this, let us try to find out the shift of neutral axis from the
centroidal axis.

(Refer Slide Time: 10:55)

𝑚𝑅 (−0.109)(157.22)
𝑒 = 𝑚−1
= (−0.109−1)
= 19. 453𝑚𝑚

3
𝑃 −100×10 𝑁
𝐴
= 5400
=− 18. 52 2
𝑚𝑚

826
 𝑚𝑅
So, step number 2 will be to find e. e is given as 𝑚−1
. So, let us do minus 0.109 into 157.22

by minus 0.109 minus 1 which gives me as plus 15.453 mm that is neutral axis is shifted
towards the center of curvature. So, that is the positive sign. So, step number 3, I want to find
the stresses at intrados and extrados. First let us find the stress of two parts.

One is the direct stress, other is the bending stress. So, let us find the direct stress first. Sigma
direct will be equal to simply P by A which will be now let us see the problem is subject to a
compressive axial force of 100 kilo newton. So, minus 100 10 power 3 by 5400 which is
minus 18.52 minus indicates compressive.

(Refer Slide Time: 13:00)

Now, let us go for bending stress. We will use Winkler Bach equation. To find bending, stress
we will use Winkler Bach equation. The bending stress in intrados is given by minus m by A
e ℎ𝑖 minus e by 𝑟𝑖. It is very easy to remember this equation. Look at this figure. From the CG

you know from the CG, this is h intrados and this is h outtrados extrados.

(ℎ𝑖−𝑒)
(σ𝑏)𝑖𝑛𝑡𝑟𝑎𝑑𝑜𝑠 =− 𝑀
𝐴𝑒 𝑟𝑖

157.22
𝑀 =+ 𝑃×𝑅 =+ 100× 3 =+ 15. 72𝑘𝑁𝑚
10

( )𝑖𝑛𝑡𝑟𝑎𝑑𝑜𝑠
σ𝑏 =−
15.72×10
5400×15.453
(57.22−15.453)
100
=− 78. 69
𝑁
𝑚𝑚
2

827
We always measure from the CG, but I want the distance from the neutral axis. So, that will
be ℎ𝑖 minus e that is what I am writing here. So, now, let us see what is my m value? m value

is P into R, I put a positive sign here because m is m tends to close the curvature. See here m
is applying this form. So, it will try to close the curvature. If it closes the curvature then we
say it is positive.

So, it is positive. Let us see what is the value which will be equal to 100 into 157.22 by 1000
that is 100 kN. So, it becomes 15.72 positive. So, now I want to find σ𝑏 ( )𝑖𝑛𝑡𝑟𝑎𝑑𝑜𝑠 which will
be minus 15.72 into 10 power 6 newton mm divided by A e 5400 and e is 15.453 that is what
e we have of h i is 57.22 see here ℎ𝑖 is 57.22 because you know this distance was 57.22.

Minus e; e was 15.453 divided by 𝑟𝑖; 𝑟𝑖 is 100. So, this gives me the value as minus 78.69

compressive. Let us try to find out sigma b o that is extrados.

(Refer Slide Time: 16:50)

(ℎ𝑜+𝑒)
(σ𝑏)𝑜𝑢𝑡𝑟𝑎𝑑𝑜𝑠 =+ 𝑀
𝐴𝑒 𝑟𝑜

(σ𝑏)𝑜𝑢𝑡𝑟𝑎𝑑𝑜𝑠 =+ 15.72×10
5400×15.453
(112.78+15.453)
270
=+ 89. 48
𝑁
𝑚𝑚
2

828
 𝑁
σ𝑖 = σ𝑑𝑖𝑟𝑒𝑐𝑡 + σ𝑏𝑒𝑛𝑑𝑖𝑛𝑔 =− 18. 52 − 78. 69 =− 97. 213 2
𝑚𝑚

𝑁
σ0 = σ𝑑𝑖𝑟𝑒𝑐𝑡 + σ𝑏𝑒𝑛𝑑𝑖𝑛𝑔 =− 18. 52 + 89. 48 =+ 70. 96 2
𝑚𝑚

Let us see what is my value in the extrados. So, sigma b o is given by positive M by A e ℎ𝑜

plus e by 𝑟𝑜. So, which will become plus we know the m value which is 15.72 by 5400 into

15.453 of ℎ𝑜. You can find out this value geometrically. You can find out this value. This is h

o which will be 112.78 plus 15.453 divided by 270 which will give me this value as plus
89.48 tensile.

So, the final stresses sigma I will be sigma direct plus sigma bending direct was compressive.
So, minus 18.52 bending was also compressive 78.69. This is minus 97.213 compressive.
Sigma extrados will be sigma direct plus sigma bending which is minus 18.52 plus 89.48
which gives me this value as plus 70.96 tensile. Friends let us run a MATLAB program with
this input and see how do we do it.

(Refer Slide Time: 18:54)

So, now I am going to use a MATLAB program. So, the MATLAB program is shown on the
screen for you. It is a T section. We know we are going to apply a load of 100. Please look at
these values. You can very well see from the figure 𝑟1 is 100, I will just mark it here. 𝑟1 is

100, 𝑟2 is 120 and 𝑟3 is 270; radius is 157.22 we already computed that.

829
(Refer Slide Time: 19:34)

And, h intrados and h extrados has been also plotted and 𝑏1 and 𝑏2 are given, and area is now

input. So, we wrote the equation. Now, we run this program and copy these values and take it
back to our screen. I will plot these values and take it back to our screen these are the values I
have let me mark it here. Let me bring it down here.

You know friends we have m value as minus 0.11, let us see what m value we have minus
0.11. So, we had e values 15.48 we have 15.453 it is and stress at point intrados is 97
compressive 97 and extrados 70, you have 70. So, the program is easy to run the Winkler
Bach equation and get to the stresses in the intrados and extrados of this curved beam.

830
(Refer Slide Time: 20:59)

Let us do one more problem on a curved beam. We will say example 2. Let us say we have
again a curved beam, but the cross section is different in this case. I will copy this curved
beam. Take a selection tool copy this curved beam, but the load is 5 kN and this value is 150.
The cross-section I am drawing here which is going to be this is my plane of reference which
is O, O. This value is 𝑏1 which is 80 mm and this value is 𝑏2 which is 30.

Now, let us say the depth of the section is 100 mm. The section has a centroidal axis which is
at a distance R from the plane of reference which we do not know. We have to compute that
this the section also has the neutral axis. We assume that shifted towards the center and we
know this shift is e, if it is positive.

Let us also mark these values. So, this is 𝑟1 which is 150 as given in the figure see here and

this is 𝑟2 from here with 𝑟1 plus the depth of the section 𝑏1, 𝑏2 all data are available to us.

Now, this is my Cg. So, we want to find the stresses at the intrados and extrados, let us mark
the points. So, I want to get the stresses at two positions intrados is A, extrados is B. I say we
want stresses at. So, find stress at A and stress at B.

So, since it is a curved beam I cannot use a standard theory for bending, I have to use a
Winkler Bach equation. I will use that use a MATLAB program and try to solve this. Let us
do that. So, with this data I will directly take you to the MATLAB program, let us see what

831
are the inputs we want. So, one can find the x bar value I think that I leave it as an exercise.
You should be able to get the R value.

Simply using the first principles you should be able to get or locate the centroidal axis and
you should be able to get the R value.

(Refer Slide Time: 25:10)

Let us take you directly to the MATLAB program. So, as you see from the figure height of
the section is 100 mm. 𝑏1 is 80 and 𝑏2 is 30, 𝑟1 is 150 and p is 5 kN.

(Refer Slide Time: 25:44)

832
So, for the geometric property you can calculate the neutral axis and the centroidal axis, I will
run the program. So, I get m as 0.02 and 3.93. I will copy this value directly put it here.

(Refer Slide Time: 25:59)

𝑚 =− 0. 02

𝑒 =+ 𝑣𝑒

𝑒 =+ 3. 93𝑚𝑚
𝑁
σ𝐴 =+ 12. 35 2
𝑚𝑚

𝑁
σ𝐵 =− 10. 05 2
𝑚𝑚

So, friends m is obtained directly from the program as 0.02 negative, e is taken as positive
which is plus 3.93 mm. It means this is shifted towards the center of curvature. After we got
this, we apply the equation and get stress at intrados is 12.35 and stress at extrados is 10.05.
Let us see.

Now, when I apply, I get these values the net stress compressive. this will be sum of this will
be sum of direct stress plus bending stress. So, I get here as plus and minus as a final result.
So, easily one can use the MATLAB program directly with the input available and try to find
the stresses in these sections.

833
(Refer Slide Time: 27:18)

We will take one more problem which is now on the crane hook. Crane hook is also
considered as a standard curved beam which can be used for various purposes crane hooks
are actually used to lift heavy objects, in the construction site. They are also useful to hold the
object during fabrication. Friends, in many circumstances choosing an appropriate crane
capacity itself is a big challenge.

If you use a wrong crane capacity and try to lift a job or an object what they call as a job there
can be fracture initiated in the crane hook which can cause a permanent damage to the crane
hook. So, it is very important for us to know how to compute the stresses in a crane hook for
a given lift of a job load. So, let us say a crane hook is shown in the figure its capacity is 10
tons. Find the stresses in the intrados and extrados when the crane is lifting at its maximum
capacity.

834
(Refer Slide Time: 29:33)

Let us see what is the cross-section of the hook. The crane hook is like this I am just drawing
the figure we will do this later. from here this radius is 70 let us also mark the center line of
this that is where R will be marked. It is now subjected to a load of 10 tons that is what is
going to lift. The points are these two A and B. We want to find the stresses at A and B which
happens to be the intrados and extrados of this crane hook.

Let us see the cross-section of the crane hook. Let us draw the plane of reference. So, this
dimension is 90. This dimension is 30 and depth of the section is 120. Let us say it has got a
centroidal axis and the neutral axis. the centroidal axis is at a distance R from the plane of
reference and the shift of the neutral axis from the centroidal axis is taken as e towards the
center of curvature. Now, we want to locate the C g. We call this as we copy this figure, we
will copy this figure.

835
(Refer Slide Time: 33:12)

Let us say we want to locate this distance. We call this distance as 𝑥 . So, I want to find 𝑥. I
can say x bar is sigma a x bar by sigma a which will be I will divide this into triangle and this
way. So, I will call piece number 1, piece number 2 and piece number 3. We will say which
will be half into base into height into one third of height or piece number 1; half into base
into height into one third of height this is for piece number 3; plus, the rectangle which is 30
into 120 into 60, this is for piece number 2 divided by the whole area half h a plus b I get x
bar as 50.

So, I have this value as 50 mm. So, R is now you know this value from the figure is 70, this is
r 1 is given. So, if I say this is 50, then I can say R as because this is 70; 50 plus 70 which is
120 mm. Now, this becomes my 𝑟1, this becomes my 𝑟2 and 𝑟2 will be now equal to 70 plus

120 which is 190. So, 𝑟1 is 70 and of course, 𝑏2 is 30 and 𝑏1 is 90.

Σ𝑎𝑥
𝑥= Σ𝑎

( 1
2
1
) (
×(30×120)× 3 ×(120) +
1
1
2
1
)
×(30×120)× 3 ×(120) +((30×120)×60)2
3
𝑥= 1
2
(120)(30+90)

𝑥 = 50𝑚𝑚

𝑅 = 50 + 70 = 120𝑚𝑚

𝑟2 = (70 + 120) = 190𝑚𝑚

836
 𝑟1 = 70𝑚𝑚

𝑏1 = 90𝑚𝑚

𝑏2 = 30𝑚𝑚

(Refer Slide Time: 36:09)

Now, let us write the equation to find the modified factor m; m is given by

( )
𝑚 = 1−
𝑅
𝐴
⎡𝑏 +
⎢ 2
⎣ ( 𝑟2 𝑏1−𝑏2
𝑟2−𝑟1 )( )
𝑙𝑛
𝑟2
𝑟1
− 𝑏1 − 𝑏2 ⎤⎥
( )⎦
𝑚 =− 0. 08
𝑚𝑅 −0.080×120
𝑒 = 𝑚−1
= −0.080−1
=+ 8. 89𝑚𝑚

We have this equation derived; you can check that minus 𝑏1 − 𝑏2. Let us substitute for all the

values we already have the values here R we have, 𝑟1, 𝑟2, 𝑏1, 𝑏2 we have. So, we have R, we

have 𝑏2, we have 𝑏1, we have 𝑟2, 𝑟1 and we have area also.

We substitute m will become minus 0.080 and we know e is given by mR by m minus 1. So,
minus 0.080 into 120 divided by 0.080 minus 1 which I get plus 8.89 mm which says that the
neutral axis is shifted by 8.89 mm towards the center of curvature. How can I say towards?
Because it is positive.

837
(Refer Slide Time: 37:51)

Having said this, let us find the direct stress which is P by A which is. You can see very
carefully here friends this will try to create a tensile stress. So, the direct stress will be tensile
plus 100 into 10 power 3 by 7200 that is the area we get plus 13.89 which is tensile. Now, I
want to find the bending stress. So, let me find the moment which will be P into lever arm.

3
𝑃 +100×10 𝑁
𝐴
= 7200
=+ 13. 89 2
𝑚𝑚

120
𝑀 = 𝑃𝑅 = (10×10) 3 =− 12𝑘𝑁𝑚
10

Now, look at this figure P will have a tendency to open. Moment will open the curvature.
When you apply this, I will try to open the curvature node. So, it is negative. So, M will be
now equal to 10 into 10 power 3 is 110 tons. So, 10 into 10; so, I get so many kN into 120
that is my R value divided by 1000 which will be minus 12 kNm. This negative is due to fact
that M will tend to open the curvature.

838
(Refer Slide Time: 40:19)

So, now we can run the MATLAB program and find the stresses sigma i will be minus M by
A. How can we say that let us say I have a hook pulling I have this intrados A. So, intrados
will be subjected to tensile and extrados should be compressive is it not because it is going to
pull.

(ℎ𝑖−𝑒)
(σ𝑏)𝑖𝑛𝑡𝑟𝑎𝑑𝑜𝑠 =− 𝑀
𝐴𝑒 𝑟𝑖

( )𝑖𝑛𝑡𝑟𝑎𝑑𝑜𝑠 =− (7200×8.89)
−12×10 (50−8.89) 𝑁
σ𝑏 70
=+ 110. 10 2
𝑚𝑚

(ℎ0−𝑒)
(σ𝑏)𝑜𝑢𝑡𝑟𝑎𝑑𝑜𝑠 =+ 𝑀
𝐴𝑒 𝑟0

(12×106)
(σ𝑏)𝑜𝑢𝑡𝑟𝑎𝑑𝑜𝑠 =+ 7200×8.89
(70+8.89)
190
=− 77. 84
𝑁
𝑚𝑚
2

So, my M by A e ℎ𝑖 minus e by 𝑟𝑖 which will be minus of minus 12 10 power 6 by 7200 into

8.89 into 50 minus 8.89 by 𝑟𝑖 which is 70 which will give plus 110.10 tensile sigma extrados

will be plus m by A e ℎ𝑜 plus e by 𝑟𝑜 in my case going to be plus times of 12 10 power 6 by

7200 0 8.89 into 70 plus 8.89 divided by 190. I get this value as minus 77.84 compressive.

839
(Refer Slide Time: 42:02)

𝑁
σ𝐴 = σ𝑑𝑖𝑟𝑒𝑐𝑡 + σ𝑏𝑒𝑛𝑑𝑖𝑛𝑔 =+ 13. 89 + 110. 10 = 123. 99 2
𝑚𝑚

𝑁
σ𝑏 = σ𝑑𝑖𝑟𝑒𝑐𝑡 + σ𝑏𝑒𝑛𝑑𝑖𝑛𝑔 =+ 13. 89 − 77. 84 =− 63. 95 2
𝑚𝑚

Now, the final stresses sigma A will be direct plus bending; direct was plus 13.89 and
bending was 110.1 which gets 123.99 which is tensile in the intrados. In the extrados this is
plus 13.89 minus 77.84 which will be 63.95 which is compressive.

840
(Refer Slide Time: 42:51)

We can also run a quickly a MATLAB program for this. I will show it on the screen. Look at
the MATLAB program. The load is in kilo newton. So, it is 10 ton. So, it is 100 kilo newton
and we have given the values h 1, h 2 I mean h 1 is the depth of section 𝑏1, 𝑏2, 𝑟1 and this let

us run the program. So, I get I will just copy these values take it back to my screen you will
see my m value is minus 0.08.

Let us see how do we how much did we get minus 0.08 and this e value is plus 8.92, e got
plus 8.92 then we have the final stresses at A is 123 and at B is minus 63.

841
(Refer Slide Time: 44:10)

So, friends we understood in this lecture more examples on curved beams. We also solved a
problem on a crane hook. We have used MATLAB program to solve these problems and
understood the stress nature generated intrados and extrados of a given crane hook for a given
kind of load.

So, this would be very interesting friends in curved beam examples where we have got
proving rings, we have got crane hooks where a mechanical engineer or a structural engineer
is supposed to check the stress across the cross-section of the crane hook or the proving ring
before is subjected in a specific kind of lift job. So, this lectures will be helpful for you to do
a preliminary study and try to get the stresses on the sections at (Refer Time: 45:10) points
for a given load before it is put to the job work.

Thank you very much and have a good day.

842
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 50
Thin - walled section

(Refer Slide Time: 00:22)

Welcome to the 50th lecture on the course Advanced Steel Design. In this lecture, we are
going to learn more about Thin-walled sections which lead to the discussion for lateral
torsional buckling (LTB).

We all know that structural steel is the most preferred construction material for many
engineering projects. Structural steel offers more advantages compared to other construction
material which is popular which is reinforced concrete. In fact, there are certain applications
where steel is considered to be a non-replaceable only construction material.

843
(Refer Slide Time: 01:20)

So, steel has exclusive application dominance in certain areas like long span roof stresses,
pavilions, cable stayed bridges, stadiums and of course, offshore platforms.

(Refer Slide Time: 02:18)

Steel has got very exclusive advantages; we have seen them in repeated lectures.

844
(Refer Slide Time: 02:37)

Usually, structural steel is available in different shapes. So, we can say they are composite
shaped they have light weight and of course, thin sections are also used. When we talk about
thin cross-sections there is a complexity. In addition to shear and axial resistance, other
parameters should also be checked for design of steel beams and columns when you use thin
sections.

So, there are well laid design procedures in design course design checks are carried out with
advanced application under the structural response. Therefore, in the thin sections special
attention is paid is drawn towards torsion.

845
(Refer Slide Time: 04:33)

Let us look at the figure. I will copy this figure here. This is a circular section subjected to
torsion at one end. In case of pure torsion which is referred as Saint – Venant torsion, so, this
is a classical example of pure torsion which is referred also as Saint – Venant torsion.

So, let us try to understand this slightly in a better manner a circular bar of radius r and length
L is considered. This bar is subjected to pure torsion at one of its ends. This is possible only
when the other end of the bar is fixed. So, this is a case of one end fixed and other end free
which is similar to a cantilever.

(Refer Slide Time: 07:13)

846
Now, we are applying a twisting moment of magnitude T at the free end. This simulates a
cantilever beam as well; we have to keep it in mind. Now, under the influence of this torsion
an initial point is shifted to a new position, see this figure. The initial point B is shifted to a
'
new position B prime. So, under torsion on the circumference is shifted to a new position 𝐵 .

Now, we make certain assumptions we will say that there is no change in diameter of the bar
before and after applying T. So, this imparts a very important condition. This imparts a fact
that any section cut normal to the longitudinal axis of the member is circular despite torsion is
being applied that is a very interesting assumption.

(Refer Slide Time: 09:52)

So, any section cut normal to the longitudinal axis of the member remains circular even
though your twisting moment is applied. So, there is no distortion in the shape of the
cross-section because the cross section initially was circular later on also it remains circular
even though it is subject to torsional collinear. Now, interestingly friends, this condition is
valid only if the material is elastic.

847
(Refer Slide Time: 11:47)

Therefore, in such analysis the material is assumed to be linear elastic. Also, friends please
note twisting moment will generate stresses. They are very small compared to the yield stress
of the material please understand this the magnitude of stresses developed because of this
twisting moment is very small.

Now, please note yield stress is a material property under axial stress strain, please
understand that. It is not under any twisting moment. Yield stress is calculated based upon
axial stress strain, but this is a very important assumption we make, so that is helpful for our
analysis.

The second assumption what we make is the bar undergoes small deformation let us say not
deformation let us call this as a deflection. It is a cantilever bar; it undergoes small deflection.
This is a very important assumption because it conforms to the fact that if large deflection is
expected cantilever, then the behavior is of a different phenomenon.

Therefore, friends, pure torsion isolates the effects that arise from deflection of the member
which will be inherently present in the member, but that is being neglected. In simple terms,
when we focus on effect of twisting moment applied at the free end this end has already
undergone a concept deflection which is oversight, that is a very important statement we want
to make here.

848
(Refer Slide Time: 14:50)

The cantilever end is already undergone a significant deflection. But, under the application of
twisting moment this deflection is over sighted. We are not looking into it, we are considering
it. Now, let us copy this figure. Now, we will equate the external twisting moment to the
internal.

𝐺𝐽θ
𝑇 = 𝑇𝑖𝑛𝑡 = 𝐿

we call equation number 1.

𝐺𝑟θ
τ𝑚𝑎𝑥 = 𝐿

Equation 2. Now, let us substitute equation 2 in 1 and simplify we get T as let us say G theta
by L is tau max. So,

𝐽
𝑇 = τ𝑚𝑎𝑥 𝑟

this calls equation number 3.

849
(Refer Slide Time: 17:54)

Having said this, we also know that

𝑇𝐿
θ= 𝐺𝐽

which can be written from the simplification. Now, τ𝑚𝑎𝑥 in this expression is the maximum

shear stress at the circumference. θ is the angle of twist; G is the shear modulus of the
material; J by r is termed as torsional section modulus. It is very important to note friends,
except shear stress, no other stresses are developed in the cross-section, in the cross-section.

(Refer Slide Time: 19:41)

850
Now, when you look at a member under pure torsion, it enjoys one more benefit. Members
under pure torsion enjoy one additional benefit. What is that? Cross-section rotates as a rigid
body. So, what does it mean? No distortion is developed in the cross section when T is
applied, if it is a pure torsion case.

But, the angle of twist will change along length of the member. Probably, theta will be
maximum at the free end where T is applied and theta will be 0 at the fixed end. So, theta will
keep on changing along the length of the member. It keeps on decreasing or the section
travels towards the fixed end where the twist is 0.

So, we should say theta is maximum at T is maximum; θ=0 where t is 0, there is no twisting
moment at the end, but we have assumed that no distortion occurs. So, then what will happen
to the cross-section? See, you are having a change in theta. So, it has to twist the section, but
we said that section, cross section shape does not change at all under pure torsion, then what
should have happened to the section? The section undergoes warping.

(Refer Slide Time: 22:36)

So, in such condition under pure torsion the section should warp, but pure torsion case
confirms also that it does not cause warping. Therefore, the angle of twist is allowed to
change emphasizing the fact along the length of the member is allowed to warp, but warping
is same at all cross-sections along the member that is a beautiful assumption technically
valid.

851
Section is allowed to warp, but warping is same at all sections along the length of the
member. Therefore, friends warping will not cause any additional stresses in the circular
cross-section.

(Refer Slide Time: 24:46)

So, testing produces only pure shear stress no axial, no bending stress no axial stress and no
bending stress. Such a behavior under pure tension is perfectly valid for circular cross section
in addition to thick-walled beam sections. So, this is also valid for thick-walled beam section
which is circular in cross-section. Having said this, let us see what happens in case of a
closed thin wall section.

852
(Refer Slide Time: 26:12)

Let us say a closed thin-walled section as shown in the figure. In case of a closed thin-walled
section the behavior is different. Torsion of equal magnitude, but opposite in nature will be
applied to the closed thin-walled sections. Let us consider a thin-walled section as shown the
figure.

Let the length of the section be L under the action of applied torsion one can notice that the
point located on the circumference is shift. The amount of shift is not same at both ends of
' '
the member. For example, at one end the shift is 𝑏 𝑎 , at the other end the shift is b-a. Now,
' '
the shift 𝑏 𝑎 is not same as b, they are different.

For a thin-walled section namely tubular or shaft section one can say the Saint – Venant
torsion constant is given by the following equation.

853
(Refer Slide Time: 28:53)

𝑇𝑟
τ = 𝐽

Saint – Venant torsion constant J is given by

4 4
π(𝐷0−𝐷𝑖 )
𝐽= 32

equation number 5 and

4 4
π(𝐷0−𝐷𝑖 )
𝐽= 32

which is 6. Friends, even in closed thin-walled sections only pure torsion is developed. Most
importantly J is not the polar moment of inertia, but it happens to be same as polar moment of
inertia.

Hence friends, Saint – Venant theory is perfectly true for circular sections. So, Saint – Venant
theory of pure torsion is true for circular cross-sections of thin-walled tubes, but not true in
case of thin-walled I section, channel section, L section etcetera. So, only for circular this is
true.

854
(Refer Slide Time: 31:23)

Then you may ask me a question in these following sections. What happens additional in
thin-walled open sections like T section, channel section, angle sections. Additional stresses
will be developed, it will also experience distortion.

(Refer Slide Time: 32:15)

Let us say for example, in I section as shown in the figure. So, look at I section as shown in
the figure. So, this is my plan view, of the top flange, this is plan of the bottom flange, then
this simply view shown. Here it is an open thin-walled I section.

855
Let us consider this figure with I section fixed at one end. A clockwise torsion is applied
externally. Let us apply a clockwise twisting moment external at one end. The section will
twist as you see on the figure, very important observation. Section will twist about it is C g,
first observation.

Second observation, you will see that the top flange is horizontal at the fixed end, but twist at
free end correct same with the bottom flange also. So, now, one can say the twist is seen in
every section along the length of the beam.

(Refer Slide Time: 35:17)

This twist is seen in every section along the length of the beam. In the top flange is twisted to
the maximum we can see top flange is twisted to the maximum at the free end and 0 and the
fixed end, which means that twist is imposed on top flange at every cross section long length
of the member. So, twist is imposed on the top flange at every section along the length of the
member.

Further this twist is not equal at different sections cut along the length of the member,
because how we can say this? It is maximum with the free end and 0 with the fixed end so, it
is not same. So, when I have a twist in the top flange which is happening at every section
along the length of the member which is also not uniform. We can say that top flange is
subjected to distortion.

856
It implies the fact that bottom and top flanges have twisted and bent as well, see the figure.
There is a downward deflection, I am marking the point here from here to here, there is a
downward deflection, but from here to here there is an upward deflection.

(Refer Slide Time: 38:04)

So, what we can say here is in addition to twisting both top and bottom flange planes also
bent. So, therefore, the top and bottom flanges do not remain plain anymore. They are
subjected to warping.

Furthermore, one can also notice the top flange is literally moved to the right side while the
bottom flange move to the left. See the figure here. Top flanges move to the right while
bottom flange move to the left, which means that, both top and bottom flanges move in the
opposite direction laterally so, therefore, both flanges move laterally in the opposite direction.
These lateral movements will now cause additional moments and the ends to contract.

857
(Refer Slide Time: 41:06)

Hence moments occurring at the top and bottom flanges in the opposite direction will cause
warping in the cross-section. In addition, these moments will also induce longitudinal stress
and strain. So, therefore, friends in an open thin-walled section, other than circular
application of torsion resultant shear stress which also caused additional bending stress.

Hence the total external torsion can be expressed as a sum of pure torsion and warping.
Hence the external torsion is now the sum of pure torsion and warping. So, T will be equal to
𝑇 = 𝑇𝑉 + 𝑇𝑊 equation number 7.

858
(Refer Slide Time: 43:08)

Now, having said this please pay attention to this figure where the rotation is indicated as
theta or phi and each displacement laterally of the top and bottom by one delta is summed up
as 2 delta and there is a moment cost or a couple which is effective with a horizontal force H
separated by height of the member.

So, this is a figure which shows twisting in open thin-walled sections other than circular, it is
a nice section. So, the twisting moment applied at one end can be seen as a sum of pure
torsion and additional moment T w. So, I can now say pure torsion plus additional moment.
This additional moment is caused due to the lateral sway of top and bottom flange. So, they
move with the same amount delta. So, the next way is 2 delta, but they are in opposite
direction.

859
(Refer Slide Time: 44:54)

Now, moment developed due to this is called as flexural twist which I call as T w causing
warping. It also idealizes that each of the flange bends is a rectangular beam about its minor
axis. Each of the flange bends about minor axis rectangular b that is an assumption we make.

So, friends, in this lecture we started learning the behavior of thin-walled open sections,
thin-walled closed sections which are circular which represents the fact of pure torsion or
otherwise called as Saint – Venant’s torsion. On the other hand, if you have open thin walled,
I section for example, they undergo warping. So, now, our job is to find out the behavior of
such sections under warping and try to develop a governing equation for the effect of torsion
on such sections. So, we will discuss this in the next lecture.

Thank you very much and have a good day, bye.

860
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 51
Open thin-walled section

(Refer Slide Time: 00:22)

Friends, welcome to the 51st lecture of Advanced Steel Design. In this lecture, we will
continue to learn more about torsion in Open thin-walled sections.

861
(Refer Slide Time: 00:57)

We already said that in case of open thin-walled sections additional stresses will be developed
due to warping. Hence therefore, the applied torsion will be now equal to torsion due to pure
shear + warping. This equation 7, we already have with us. Now, referring to this figure, we
say the torsion is supplied and an axis X, Y, Z and it is about Z axis which can be a sum of
pure torsion + torsion caused because of warping.

And, this is the shift what we have seen which is equally displaced both top and bottom
flanges, in the opposite direction by delta by 2 or x by 2. They create a couple which is H into
d, as we see in the figure. So, T w will be now counteracted by this additional moment. Now,
let us see what is Tv, T v is (Refer Time: 02:53) torsion which is GJ𝜃 by L which can be said
as GJ d β by dz, where we call this equation number 8.

Where, d β by dz is the change in twist at every section along the length of the member. In
the figure, you can easily see the length of the member is measured along the Z axis, am I
right see here. In addition to this, top and bottom flanges undergo later deflection by the same
amount which resulted in additional fractional twist.

862
(Refer Slide Time: 03:57)

Top and bottom flange undergo lateral deflection by the same amount, but opposite in
direction correct. So, these results or this cause an additional flexural twist. So, this additional
flexural twist is expressed as M f which is the flange bending moment which is given by -EIf
d2 x /dz2. We call this equation number 9. I f indicates the moment of inertia of the flange and
- indicates, it is opposite to the applied twisting moment. Also note the second derivative is
along the X axis which is the lateral deflection, as we have seen in the figure.

(Refer Slide Time: 05:50)

863
Therefore, the corresponding shear force is given by H, which is actually equal to dm / dz
because that is the length of span which is - E If d3x /dz3, equation 10. So, If is approximately
half of Iy, that is usually the design practice, where I y is the moment of inertia about the
weaker axis, I should say minor axis.

Also, please understand that the web contribution in moment of inertia is neglected. Now, the
additional moment induced by the lateral deflection of the flange T w is given by this
equation. The additional moment induced by the lateral deflection of the flanges is T w which
is H into d. We call this equation number 12, where d is the depth of the section.

(Refer Slide Time: 07:48)

Now, friends for a very small angle of twist, we can assume the following x is Bd/2. Hence,
Tw will be now -E Iy d2 /4, d3β /dz3, equation 4. Now, we can say the total applied torsion is
given by T which is GJ dB/dz - ECw d3β by dz3, where C w is called warping constant which
is approximately equal to Iy d2 /4. In the above equation, GJ is called torsional rigidity and E
Cw is called warping rigidity.

864
(Refer Slide Time: 09:59)

So, in case of open sections which has an additional flexural twisting, the term E Cw d3β/dz3
arise due to warping. This is a very important concern in design. Having understood how
torsion is initiated and how warping is built in open sections; let us play pay attention to
buckling.

(Refer Slide Time: 11:06)

We already know buckling failure is one of the dominant modes of failure in many steel
structures. Buckling is one of the modes of failure, when a structural member is displaced
laterally out of its plane under compressive stresses. The lateral displacements are associated

865
with the flexural stresses whose magnitude depends on slenderness ratio of the member. We
also know buckling of the beam can be categorized into local and global buckling.

Local buckling is specific to beams whose web or flange is slender. So, buckling will have
two components, local and global. Local buckling will be specific to beams with web or
flange is slender. So, therefore, many small buckles tend to appear along the web. So, along
the web many small buckles appear.

(Refer Slide Time: 12:52)

So, one can see this figure, can see along the web there are lot of small buckling. If the flange
is slender can also happen in the flange. So, in case of flange buckling, the buckling length
varies between 1 to 5 times of the width of the flange. In case of flange buckling, we have a
we have a condition, the buckling length which is we call this as the buckling length, is about
1 to 5 times of the width of the flange whereas, is the width of the flange.

This is an experimental observation made by Hogland in 2006. When the buckling of the
beam is extended to its entire length, it is termed as global buckling. This type of buckling is
further classified based on type of loading and nature of displace. So, this is essentially the
local buckling figure. This is on the flange. This is on the web.

866
(Refer Slide Time: 15:17)

Now, let us pay more attention to the global buckling modes. Due to buckling steel beam can
fail in different modes. The failure modes are governed by type of the load and deformation
shape of the member. Furthermore, they can be subject different types of deformation
simultaneously, bending, torsion and warping. The resistance of the beam depends on the
stiffness and different modes of the deformations.

So, the possible modes of deformations are bending, torsion and warping, each one of them
has different kind of stiffness. So, the resistance of the depends against this kind of bending
or deformation depends on the respective stiffness. Therefore, flexural, torsional and warping
stiffness are essential for estimating buckling resistance.

So, if one need to estimate buckling resistance, one should estimate the flexural, torsional,
and warping stiffness of the beam. Therefore, beams overall buckling resistance is a
combination of flexural, torsional buckling and lateral torsional buckling. Let us see how they
look like in different modes.

867
(Refer Slide Time: 17:57)

So, if I say flexural buckling, if this is my open I section. There will be a shift either in this
format or in this format. If we talk about torsional buckling, the original I section will get
rotated about its cg. When we talk about lateral torsional buckling, then I section is
completely displaced like this.

Flexural buckling can occur only when the compression flange buckles in one plane.
Resistance of the beam to this kind of buckling depends on EI, the flexural rigidity where I is
the moment of inertia of the plane about the plane of flexural. When the beam is caused,
buckling of the beam is caused by compression and induces only twisting then it is resulting
torsional buckling. So, this will occur when the buckling is caused by compression inducing
twisting.

In this case, torsion rigidity will play a role or torsion rigidity and warping rigidity are
important for design. In some cases, when the beam is subjected to axial compression it may
experience both flexural and torsional buckling. This combined mode is called flexural
torsional buckling, also known as lateral torsional buckling.

868
(Refer Slide Time: 22:57)

So, you can also see the picture on the screen. This is about the major axis. This is about the
minor axis. This is flexural buckling. This is torsion buckling. This is flexural torsional
buckling. These are all different modes of buckling of beam under compression. Now, let us
pay attention to lateral torsional buckling.

(Refer Slide Time: 24:05)

Amongst all the modes of buckling, lateral torsional buckling calls for a special attention
because this is very critical in design. Lateral torsional buckling is essentially related to

869
beams. Let us draw a figure to understand the lateral torsional buckling of an I section. Let us
see this figure.

(Refer Slide Time: 25:04)

So, this figure is a classical example of an I section undergoing lateral torsional buckling. So,
now the figure shows you an I section under transverse load. This is a classical case of open
thin-walled section. The top flange of the beam is in compression. The bottom flange of the
beam is in tension.

Under the gradual increase of transverse magnitude of the load, the compression flange acts
as a column and it starts deflecting in the lateral direction. You can see the figure. Just mark,
it starts deflecting in the lateral direction. It buckles out of plane, because there are no
restraints available. On the other plane web offers a resistance, therefore does not deflect.

But, in the it offers an outer plane bending. The bottom flange will anyway have no
deformation, please note that. Practically, the upper half of the section undergoes lateral
deflection, while the lower half of the section does not undergo any deflection. We can write
this.

870
(Refer Slide Time: 28:03)

Therefore, one can also state that the bottom flange is less severely loaded than the top
flange. So, we can infer, but in the case of a column one can agree that both top and bottom
flanges would have undergone the lateral deflection right.

But in this case of an example beam only half of the section is intention does not buckle. So,
the lower half does not buckle because, it is under tension, and this causes overall twist of the
cross section. There is an overall twist of the cross section and under lateral torsional
buckling; we can have the following observations.

(Refer Slide Time: 30:14)

871
The top flange tends to undergo flexural buckling which is similar to columns. This causes
out of plane bending of the top flange. The bottom flange is in tension and does not deflect
and this causes twisting on the cross section in addition to out of plane bending. Please
understand, there is an additional effect. This is identified as lateral torsional buckling.
Therefore, friends in lateral torsional buckling, the lateral deflection of the compression
flange exist along with the twist of cross section.

(Refer Slide Time: 32:38)

Lateral torsional buckling is the lateral deflection of the compression flange along with the
twist of cross section. Now, the question comes we have understood how it is lateral, we have
also understood how it is torsional, because there is a twist there is a later deflection.

Why is it called buckling? Friends, it is very simple, compression flange undergoes lateral
deflection, cross section undergoes twisting. Since, it is compression flange; we term this as
buckling. Since, compression flange is involved, we call this term buckling. In case of open
thin wall section, the compression flange essentially behaves like a column. Due to this
reason, lateral torsional buckling resistance equations have Euler’s buckling load involved in
this.

872
(Refer Slide Time: 34:52)

So, friends in the control equation of lateral torsional buckling, Euler’s buckling load term
will be included. So, in simple terms LTB can be explained like this. When a beam is under
flexural load and deflects laterally, it also results in twisting of cross section. This is called
Lateral Torsional Buckling, shortly known as LTB. When a beam is subjected to lateral
torsional buckling, the cross-sectional resistance of the beam should be derived from the
following quantities.

(Refer Slide Time: 36:32)

873
Hence, resistance of the beam against lateral torsional buckling should be derived from
fractional, torsional, and warping rigidities that is very important. The three components that
contribute to the resistance of the beam subjected to lateral torsional buckling. The
phenomenon of lateral torsional buckling is important in design and it should be controlled,
that is a very serious concern in design. Let us see how we explain this using a simple beam.

(Refer Slide Time: 37:50)

Let us take a simple beam which is undeformed geometry. Let us say it is an I section. Let us
say the beam is a centroid axis as now being marked on the screen. We call this as an X axis
and this as an Y axis and this will be the Z axis. Let us consider a section at a distance z. Let
us consider the section m the distance. So, this is my Y axis, looking for Z axis here, looking
for Z Y and this is an undeformed geometry. Now, let us try to draw the deform geometry.

874
(Refer Slide Time: 40:11)

So, the beam initially was like this. Let us say the beam is now deflected, rotate actually by
an angle phi. Let us say is the new section. Let us say this was my X axis and let us say the
displacement of the new cg from here is - u, move towards the left. The point c has shifted
and rotated to the point c dash and the axis becomes eta. This is the deformed geometry. So,
now, I can also draw the longitudinal view of this beam.

Let me do it in a single line easy. This is my Z and Y axis and my new geometry deforms this
way. Let us cut a section m and let me mark the neutral axis of the new one and let the
intersection be eta and epsilon or zeta. This is m n. This is the deformed geometry. So, simply
supported beam under two conditions deformed and undeformed geometry.

The X axis is the major axis and Y axis is the minor axis or I should say the weaker axis, Z
axis is along the length of the member, that is what the figure shows. Let us take a section
along the span of the member and view this section before and after deformation, that is what
we are drawn. So, as we focus on the lateral deformation, it can be seen that the lateral
deformation of the beam is along the X axis correct.

The compression flange has bent out of plane about its minor axis. The torsion section gets
twisted which can be seen in the side view of the section. Now, the cross section axis are now
zeta eta or zeta eta. Let me mark this zeta eta which corresponds to Z and Y. So, zeta eta
corresponds to Z and Y axis and u and v are the corresponding coordinates.

875
So, now in this figure this point is - v and this point is - u. Now, this will redesignate the new
position after twisting. So, u and v are the corresponding coordinates of the new cg of the
section, after the section is rotated by angle phi. Now, the twisted cross section signifies the
fact that the flange is under flexural bending, while the compression flange is under lateral
torsional buckling.

(Refer Slide Time: 47:21)

So, friends in this lecture, we started learning about lateral torsional buckling. We said the
resistance offered to the beam under LTB has got three components, is it not. It comes from
flexural; it comes from torsional, comes from warping rigidity. We will continue derive this
section and discuss LTB in the next lecture again. Before that give a thorough reading and
understand the figure of deformation and please understand why it is called lateral torsional
buckling.

Thank you very much and have a good day. Bye.

876
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 52
Lateral torsional buckling

Friends, welcome to the 52nd lecture on Advanced Steel Design. We are going to continue to
discuss the Lateral torsional buckling. In the last lecture we explain independently how do we
get lateral, torsional and why it is called as buckling. Let us recapture the last figure, what we
drew for an undeformed and deformed geometry which I have drawn on the screen now.

(Refer Slide Time: 01:00)

So, this is my undeformed geometry. And, the second figure what you see is the deformed
geometry. So, we are considering a simply supported beam, as loaded. The cross-section axis
are now modified, the cross-section axis after deformation are modified as ζ, η axis as you
see in the figure and this corresponds to Z, Y axis which is undeformed,. Let us also
understand that u and v are coordinates of the C g of the section which will now designate the
new position after twisting.

Now, having said this, let us look at the deformed geometry where the C g point is shifted to
C prime and the rotation is ϕ. So, now, let u and v are marked as - u and - v because you
know the positive X and positive Y are in the other direction as you see here. So; obviously,

877
C dash is displaced in the negative coordinates of X and Y respectively as u and v as shown
in the figure.

Now, the twisted cross-section, the phase of the twisted cross-section, signifies the flange
under flexural bending; while the compression flange is under lateral torsional buckling. You
know it’s a simply supported beam we can very well see from the figure that the compression
flange is under lateral torsional buckling. Whereas the other flange is under flexural buckling.
Now friends, this arrangement is like the column buckling.

(Refer Slide Time: 04:36)

Therefore so, elastic buckling equation can be derived for a simply supported beam. Now, let
us consider a double-symmetric I section. This is now subjected to uniform moments at the
ends.

(Refer Slide Time: 05:50)

878
So, let us see this figure, it is subjected to uniform moments at the ends M0. under such
arrangement no shear will happen. No shear will occur. So, the X-Y cross-sections
cross-section axis and the Z-axis or marked as shown in the figure. So, X-Y is the
cross-section axis and Z is along the length of the member. Now, let us look into the bottom
figure where I am marking the components of the end moment.

So, these are the components of the end moment. When you look at the figure showing the
components of the end moment, one can very well see the X-Y plane is rotated to ζ, η plane
by an angle ϕ, correct. See here. Now, when I try to mark the moment components which will
account for lateral torsional buckling so, these are the moment components.

(Refer Slide Time: 08:03)

879
So, now, I can say what are the components of the moments, which will add up to LTB? So,
these are the components of the moment which will add up LTB, where you can see that Mζ
and Mη both are adding together to form the moment about η, ζ axis which is inclined at an
angle ϕ to the original X-Y axis,. So, now, in the transformed plane, what is the transformed
plane? I should say ζ, η. The moments can be resolved as follows. So, M ζ will be equal to M
X cos ϕ which is approximately equal to M x for small rotations as a continuity we mark this
equation number 16.

Similarly, M η will be M x sin ϕ which will be approximately equal to M x of ϕ which can

also be said as M0 of ϕ, ok, that is what the original moment we applied equation number 17.

(Refer Slide Time: 10:07)

880
Now, the torsional component is resolved as below. Will be given by M x sin (- du/dz); as z is
the axis along the length of the member which came you know said as – M0 du/dz equation
number 18. So, one can very well notice here, M x replaces M0 because M0 is the end
moment acting about the X-X plane. Now, in the plane of bending which is the Y-Z plane we
can now write E I ζ d2v by dz2 will be M ζ equation 19.

Equation 19 is a simple bending equation about the major axis. This is written in consensus
with the general equation with respect to axis the beam. Similarly, with respect to minor axis,
we can say E I η d2u by dz2 will be M η, ok, we call equation number 20.

(Refer Slide Time: 12:47)

881
Now, the twisting moment is given by G J d ϕ by dz - E C w d2ϕ by dz2 is Mτ this is equation
21. Now, we can solve all these three equations. So, equations 19 to 21 can be solved
simultaneously. And, we can apply respective boundary conditions for the simply supported
beam the boundary conditions will be applicable to the simply supported beam.

So, what do we get? We get the governing equation for elastic buckling strength. This was
given by Timoshenko and Gere, which is M0, cr is square root of 𝜋 square E I Y by L square
of 𝜋 square by L square E C w + G J. We call this equation as 22 which is the classical elastic
buckling strength equation. This is applicable to very long span beams which is similar to
slender columns.

(Refer Slide Time: 15:32)

Equation 22 is applicable to very long span beams and as they behave similar to slender
columns; short and intermediate span beams will not be applicable or governed by equation
22. So, we can say equation 22 is not applicable to short and intermediate span beams. So,
friends, the elastic buckling strength equation which is governing the lateral torsional
buckling is applicable to long span beams. Now, let us see what the mechanism behind lateral
torsion is buckling as explained by Eurocode.

882
(Refer Slide Time: 16:49)

Let us refer Eurocode, EC3-2005; according to this Eurocode 3, instability is characterized by

large transverse displacements and rotation about the member axis, under bending moment
about the major axis. So, now according to this, instability is characterized by a, large
transverse displacements, and rotation about the member axis. Further bending moment about
the major axis, as per our figure it will be Y-Y axis, is it not.

This very clearly explained by Hoglend in 2006. This instability phenomena involves lateral
bending and torsion of the cross-section. So, this involves 2 aspects; i, lateral bending above
Z-Z axis, Z-axis is the length of the member axis, please understand that. ii, torsion of the
cross-section, ok these 2 are there. We will try to represent this graphically as you see here.

883
(Refer Slide Time: 19:25)

So, graphically the lateral torsional buckling is expressed this is an LTB. Of course, this
represents cantilever beam and the right-hand side figure represents LTB for a simply
supported beam. Now, in both these figures one can see that the beam is subjected to
constantly increased loading in the major axis bending. If the beam is slender, it may buckle
before the section capacity is fully utilized. This buckling involves both lateral deflection this
buckling involves both lateral deflection and twisting this involves both lateral deflection and
twisting.

This is what we call as lateral torsional buckling. The vertically applied load the vertically
applied load in both cases induces compression and tension in the flanges of the beam. It
causes deflection of the compression flange and enables laterally swaying away from the
original position. You can see the compression flange is swaying away compression flange is
swaying away from the original position; we call these phenomena as lateral torsional
buckling. On the contrast, the tension flange tends to keep the beam straight.

884
(Refer Slide Time: 21:32)

Look at figure 16. On the contrary, tension flange tends to keep the beam straight. This lateral
bending of the section creates restoring forces that oppose the moment because the section
tends to remain straight. These actions generate lateral forces which sometimes are not
adequate to prevent the section from lateral deflection and that is what we call as resistance
against lateral torsional buckling which we can see from this figure.

(Refer Slide Time: 22:39)

Let us see, what is the torsional effect under such situation. Now, the forces in the flanges,
cause the section to twist cause the cross-section to twist above its longitudinal axis. It also

885
enables lateral deflection that is what you see in the previous figure. Now, the twisting is
resisted by the torsional stiffness of the section. Now, what is the parameter in the design
which will control this? The parameter in the design, which controls this aspect is the flange
thickness. Let us say for example, we have 2 sections of the same depth.

(Refer Slide Time: 24:35)

For example, there are 2 sections of the same depth, but with different flange thickness. Now,
let us say larger t f, lesser t f thickness of the flange. The larger t f shows high bending
strength. This shows lesser bending strength. If the span is very long then the beam becomes
unstable, even for the small magnitude of the load compared to short span beams and if the
beam is the same length, but different cross-sections are imagined then beams with slender
cross-section buckle.

So, it is a very classical example where, if the thickness of the flange is not sufficient enough
to restrain or offer enough torsional resistance the section or the beam will fail by lateral
torsional buckling.

886
(Refer Slide Time: 26:21)

This is a photograph of a bridge girder failure under LTB. One can very well see the twisting
of the cross-section and the thickness of the flange is phenomenally lesser which cannot
sustain or offer enough torsional rigidity to this. So, now in controlling LTB, let us say what
are the factors in the design. 1, Of course the beam span; if it is too long then it is dangerous.
2, The cross-sectional shape.

So, friends even to control LTB we are anyway aiming at form dominance. We are looking
for the effective cross-sectional shape for controlling LTB. So, now there are various factors
which can contribute to control LTB, Hermann et al. 2014 gave a very interesting list.

887
(Refer Slide Time: 28:05)

Summarized list of factors useful to control lateral torsional buckling, let us see what they
are. Let us see one by one. Let us say the parameter and structural properties. Let us start with
material properties. So, you can material can be useful in controlling or offering resistance
with high shear modulus and with high modulus of elasticity. So, it can offer you a very high
G and E value. Then, the next parameter is the cross-sectional dimensions and span.

So, now, the structural properties governed by this will be torsional constant, warping
constant and moment of inertia about the minor axis. Let us say I Y. The third factor which
contributes to LTB is the boundary conditions or I should say support conditions. When you
look at the support conditions, they can contribute, or they can initiate bending about major
axis. They can also sometimes initiate bending about minor axis. They can also cause
warping of the cross-section.

888
(Refer Slide Time: 30:49)

The fourth factor is a load. So, depending upon the type of load, whether it is uniform
distributed load, whether it is concentrated load etcetera one can also result in lateral torsional
buckling. The next is the point of application of load where the load is applied along the
length of the member and along the breadth of the member. So, these factors are very useful
in designing the beam to control the lateral torsional buckling and that can be avoided.

Now, if you ask me a question, what will be the factors that intuit very high lateral torsional
buckling? Let us ask this question. What are the factors that can cause very high chances for
lateral torsional buckling? Ok, very simple. There are about five factors. We will take an
example of an I section. This means central axis C g; we call this as y let us say this as z.

So, if the section has low flexural stiffness, about the minor axis that is if E I z is very small it
can cause LTB; if the section has low torsional stiffness that is G I t. If the section has low
warping stiffness which is E I w; if the beam has high point of load application, load is
applied on some other plane which is at a different elevation from the beam the last point is
the beam has got very long unrestrained span. So, these are the factors which will cause very
high chances for lateral torsional bucking.

889
(Refer Slide Time: 34:30)

So, further friends it is also interesting to see that in case closed box sections LTB does not
occur, number 1. Number 2, if the section is bending about its major axis, then LTB will not
occur. These are the conditions where LTB does not occur. Now, the question comes, what
precautions, what design controls we can exercise to prevent LTB in the design.

(Refer Slide Time: 35:34)

So, what controls can be exercised against lateral torsional buckling? For preventing lateral
torsional buckling, the control factors what we discussed in the previous slide are very
interesting and very useful. For example, one of the most used techniques is to choose a

890
cross-section with high flexural stiffness about the weaker axis. So, the most used tool is to
choose a section with high flexural stiffness about the minor axis.

Alternatively, admitting the load with the lesser magnitude or providing lateral restraints for
the compression flange can also prevent LTB. They can also control LTB. Lateral torsion
buckling is only possible in major axis bending when the corresponding minor axis about the
corresponding stiffness about the minor axis is weak.

(Refer Slide Time: 37:50)

So, the control point is, LTB can occur only when the stiffness about the minor axis is weak
and the bending is about major axis. If the structural stiffness is same over both the axis, then
LTB can be avoided completely; that is a very interesting statement we have. If stiffness is
same about both major and minor axis, then LTB can be totally avoided. So, that is a very
interesting and simple tool to control LTB in the design itself. Now, offshore structures being
strategic importance, the top side of these platforms are designed within ultimate care and
generally they are designed as a built-up section.

891
(Refer Slide Time: 39:33)

So, that is one of the important reason. So, built-up sections are preferred; because one can
achieve equivalent same stiffness in the major and minor axis it is possible with built up
sections only its possible. Now, what are the possible type of sections we have. Let us draw
them, solid circular sections. If you have an I section, the loading is here then it is, ok; if you
have a box section, the loading is applied here and if this is b and if this is h we need a
condition that h by b should be less than or equal to 2.

So, these are all the sections recommended to avoid lateral torsional buckling. Now, we
discussed also an important factor saying the point of application of load can also cause LTB.

892
(Refer Slide Time: 41:37)

Let us now discuss, what is the effect of point of application of load in LTB. The loads even
with lesser magnitude can minimize the chances of occurrence of LTB. Therefore, when the
load is applied at the bottom flange, it causes a beam considerably more stable than the load
applied on the top flange. So, load applied at the bottom flange makes the beam more stable
than the load applied at the top flange. We are talking about simply supported beam.

One classical example to prove this is the crane gantry. Please, look at the crane gantry
girders; you will observe that the wheel loads are transferred to the bottom flange. You can
notice that whenever you go to any heavy industrial system, please see the crane gantries are
designed in such a manner the wheel loads will always rest on the bottom flange and not on
the top flange.

This is possibly only due to the reason that when the load does not contribute to any
restraining effects. The event of LTB the beam deflects when the load above the centre of
twist generates twisting moment. So, let us explain this graphically like this.

893
(Refer Slide Time: 44:03)

Let us say, if you have a section let us copy the same figure here. When a load is applied
here, there is a line of action of the load. So, C g let us say this is U y and we call this shift as
ϕ. This is the load P. So, the additional moment caused in this case will be the load P into U y.

Imagine that the load same is applied at the bottom flange, its applied here and is the same
shift let us say this is U y and the rotation is same the additional moment caused in this case
is - F into U y, ok; because this is causing an anticlockwise moment whereas, this is causing a
clockwise moment.

So, point of application of load can also cause interestingly the lateral torsional moment. So,
one can see from this figure when the load is applied in the bottom flange, it tends to stabilize
a section. So, it is better; this is not better. Having said this, let us see what the effect of
lateral restraints on lateral torsional buckling will.

894
(Refer Slide Time: 47:15)

These are design parameters, design aspects. So, very interestingly friends if you have an I
section. Let us say, so I can we simply say, there can be similar I sections like this which are
arranged in parallel arrangement like this. We are talking about lateral restraint offered to the
system these are all primary beams. What we can do is we can put lateral restraints in this
direction. At intermittent gaps, let us say I can put this is one beam which we have here. This
is another this is another beam we have here, there is another beam.

So, we can put the lateral restraints between these two at equal gaps. Similarly, between,
these two at equal gaps and so on and between beams b and c only at the ends. So, between b
and c we have lateral restraints only at the ends between a-b, c-d we have at equal intervals.
So, I can provide the lateral restraints along the length of the beam at equal intervals that
becomes a spacing of this lateral restraint they are called intermediate bracings. And, these
are called support bracings or sometimes called end bracings.

Now, one can also find out, what are the possible buckling modes we have with lateral
restraints to avoid lateral torsional buckling.

895
(Refer Slide Time: 51:35)

So, friends we will discuss that in the next lecture. So, in this lecture we learnt about, what
are the design factors that initiate LTB, how LTB can be avoided, what are the control
techniques one can employ in the design to avoid lateral torsional buckling. Apart from
understanding, why do we call this bending as lateral torsional buckling. We will see more
details in the next lecture. We will also work out an example design example of LTB, using
Indian code and one international code to illustrate the design of LTB, against LTB.

Thank you very much and have a good day bye.

896
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 53
Design for LTB - 1

(Refer Slide Time: 00:20)

Friends, welcome to the lecture-53 of Advanced Steel Design Course, we are now continuing
to learn the Lateral Torsional Buckling. So, in the last lecture we discussed about the design
parameters that are useful to control a section against lateral torsional buckling. Now, let us
quickly see what are the various buckling modes with lateral restraints that can be offered to
avoid LTB?

897
(Refer Slide Time: 01:22)

So, now let us see what are the various buckling modes that can happen with different lateral
constraints that are provided to avoid lateral torsional buckling , suppose if you have an I
section which has no restraints then the I section is free to buckle due to lateral torsional
buckling and the buckling will happen like this.

Suppose one restraint the compression flange like for example, I have an I section where
restrain the compression flange. So, in this case no LTB is possible.

(Refer Slide Time: 03:15)

898
Suppose, a tension flange restrained for example, this is an I section, we restrain the tension
flange then there is a possibility that it may deflect like this. So, distortion buckling will
happen then we do not provide any restraints as usual then lateral distortion buckling is also
possible. This will happen only when the beam is slender.

So, it has got flexible webs web thickness is very small and rigid flange if this combination is
there then it can result in distortion buckling as you see on the picture.

(Refer Slide Time: 05:34)

Now, let us try to compare the behavior of a real beam and see how it is different in LT, let us
now see behavior of a real beam. Now, we have seen that the lateral torsional buckling theory
is generally applicable to ideal elastic conditions for beam possessing perfect geometry , but
in reality it is not possible to have an ideal beam like this. So, one cannot get an idealized
design conditions because they will be different in reality. So, now, you will always have a
beam with imperfections.

So, let us try to mark the displacement moment curve and we will mark the real beam
behaves like this whereas and the critical moment beams without imperfections will behave
like this. So, this is the real beam and this is an ideal. So, there is a significant difference in
the response behavior of this and of course, this is what we call as initial deflection del
naught.

899
So, we comparing both the behavior as a real beam and ideal beam, it can be seen as ideal
beam is laterally undeformed until the load reaches the elastic critical moment M cr. So, one
can say ideal beam is undeformed until the load reaches elastic, critical moment marked as M
cr. So, when M cr is reached the beam experiences a different state of instability and results
in significant deflection occurs in the lateral direction since the material is ideally elastic
infinitely large deformations can occur.

So, a new state of equilibrium will be developed in the deflected position every slight
increase in load will cause significant deflection. On the contrary in case of a real beam
which has got lot of imperfections and residual stresses present in the beam during welding
when such beam is subjected to apply load a lateral imperfection initial deflection del naught
will occur.

So, in real beam, lateral deflection del naught will occur due to imperfection, the lateral
deformation increases with the increase in magnitude applied load closer to M cr the
diffraction increases spontaneously without reaching theoretical value of M cr. So, the
moment capacity of the real beam is far lesser than M cr. So, this is governed by material
property in plastic range non-linear geometry and possible local buckling.

(Refer Slide Time: 10:43)

Now, let us quickly see what are the factors that cause reduction of capacity? So, we have
understood that the real beam has got reduced moment capacity is it not that is M is lesser

900
than M cr, let us see what are the factors that cause this. 1 it may be due to non-linear
material response that is plastic behavior of the material that could be one reason.

Second reason could be initial imperfectionsthe beam may not be initially straight etcetera,
thirdly can also be due to presence of residual stresses during fabrication, during
manufacturing etcetera. The fourth reason could be local buckling of the beam in class 4
sections. The last reason could be piercings, unsymmetry and defects in the fabrication.

Now, these effects can be considered in design through design buckling curves which
simulate the real beam behavior. There are design buckling curves which are given the design
codes which can help you to have a modified factor of reduced moment capacity on account
of each one of these factors.

(Refer Slide Time: 13:11)

Now, friends let us see the design procedure for lateral torsional buckling. Let us see this.
Now, the elastic critical moment the elastic critical moment M cr is the primary design
parameter. So, how do you define M cr? M cr is defined as the maximum value of the
bending moment supported by the beam which is free from any imperfections. In practical
design M cr can be estimated using a software or by performing hand calculations.

Euro code helps the estimate M cr using a three factor formula. It is one of the most used
analytical formula to estimate the elastic critical moment as suggested by Lopex et al in 2006.

901
(Refer Slide Time: 17:44)

Now, let us see, what is the three factor formula for M cr here? M cr is referred as elastic
critical moment , which is the design parameter to control LTB in beams. So, the basis of this
formula is to apply correction factors and compute M cr. Now, the reference beam used as the
basis is a double symmetric cross section which is simply supported at both the ends
subjected to constant moment at both ends that is the ideal condition for which this material is
linearly elastic.

The cross sectional dimensions are assumed to be smaller than the curvature radius and the
deformations are considered to be very small. Let us try to see a simply supported beam
which is considered for the analysis.

902
(Refer Slide Time: 21:16)

So, we will have a simply supported beam which is undergoing an initial deflection will mark
the beam with this. So, there are n constraints let us call this end as A this end as B let us say
this beam is subjected to a momentum M y at the ends as shown. Let us say the span of the
beam is L meters and if I draw a neutral axis of this beam at any section let us draw the axis
normal and tangent we call this as x dash and z dash whereas, the original axis is X and Z this
is an elevation.

So, double symmetric section let us say the general cross section is looking like this having a
C g at the center. So, this becomes my y and z axis, if I have a wide flanged bottom when the
C g is shifted this becomes y and z and C g is shifted down, both are double symmetric
sections anyway double symmetric and single symmetric you can say. So, this is double
symmetric, and this is single symmetric, is it not?

903
(Refer Slide Time: 25:41)

Having said this the elastic critical moment of the perfect section that is of the standard
section is given by Mcr is 𝜋/L √(G I τ EIz (1+ 𝜋2 EIw /L2 GI τ) equation 23 we are
continuing from the previous derivation. So, the standard section what we say is a double
symmetric section that is the standard symmetric.

If you want to expand this M cr equation for a single symmetric section then equation 23 will
be modified to apply for single symmetric sections using the correction factors C1, C2 and
C3. These correction factors account for imperfections, these factors can be determined either
from the tables or from the figures they can also be estimated using closed form expressions
available in the literature.

904
(Refer Slide Time: 28:10)

Now, considering the warping degrees-of-freedom also and lateral rotation at the supports,
the equation is further modified with two more factors namely kz and kw. So, now the
modified 3 factor formula for singly symmetric sections is given by Mcr equals
0.5
2 ⎧⎡ 2 2 ⎫
𝐶1
π 𝐸𝐼𝑧

(𝑘𝑧𝐿) ⎨⎢⎣
2

⎩
( )
⎢
𝑘𝑧
𝑘𝑤
𝐼𝑤
𝐼𝑧
+
(𝑘𝑧𝐿) 𝐺𝐼τ
2
π 𝐸𝐼𝑧
(
+ 𝐶2𝑍𝑔 − 𝐶3𝑍𝑗 ⎥
⎥
⎦
2⎤
) ( )⎬
− 𝐶2𝑍𝑔 − 𝐶3𝑍𝑗
⎭
was equation

number 24.

So, one can see in this equation there are different factors C 1, C 2, C 3 which are accounting
for the imperfection conditions in addition to that by allowing warping degrees of freedom
and lateral rotation at the ends because the ends are simply supported k z k w are also
involved. Now, this above equation has got some limitations, now this equation is applicable
only to symmetric and singly symmetric sections. This equation includes the effects of
loading apply above or below the shear center.

905
(Refer Slide Time: 32:28)

Let us see how this included. So, let us talk about the application of law, let us say I have the
original section of this format nicer. Now, this can also have a marginal twist when the load is
applied at P which can be causing a restoring moment initially the load is applied here. So,
one can say here M cr 1 is greater than M cr because we have restoring moment available.

Let us take another example where the standard I section twists to this form where initially
the loading is applied at the shear center and now the loading is here. So, in this case M is
actually equal to M cr. Suppose we have a section where the load is applied here and the
section is twisted this will cause an additional moment in this case M cr 2 will be less than M
cr. Ideally speaking, if you have an I section is this the point where load is applied and this is
the point where the shear center is located the shift of the shear center from here is actually Z
g.

So, friends in this equation we can also write down that E is modulus of elasticity, G is the
shear modulus, I z the second moment of area about the weaker axis, I tau torsion constant, I
w warping constant, L span of the beam or center distension distance between the lateral
constraints.

Let me write here k z effective length factor, k w is the effective length factor related to
lateral bending, k w is again effective length factor, but related to warping. Z g the distance
between the point of application j center as I show here and shear center, Z j is the distance
related to effects of asymmetry about the y axis.

906
(Refer Slide Time: 38:04)

And this can be given by a separate equation Z j is Z s minus 0.5 times of integral over A y
square plus z square z d A by I y equation 25, where Z s is the distance between the shear
center and center of gravity. C 1 is the factor that accounts for shape of the moment diagram,
C 2 is a factor which accounts for point of application of load with respect to the shear center,
C 3 is a factor which accounts for asymmetry about y axis. Now, let us quickly see; what are
these moment correction factors C 1, C 2 and C 3?

(Refer Slide Time: 40:03)

907
The C 1 factor is equivalent uniform moment factor also referred as moment gradient factor,
this is valid when the load is acting at the shear center. In reality beams are often loaded on
the top and bottom flanges and not on the shear center. Therefore, a second correction factor
C 2 is applied to account for the effects of not loaded and the shear center. So, what will
happen when the loading is not occurring at the shear center?

Very good question, when the point of application of load is not at the shear center it causes
an additional moment, we should say additional twisting moment. Therefore, the load applied
under shear center stabilizes the beam the load above shear center destabilizes the beam. We
can also illustrate this with the figure which we already draw but let us once again do it.

(Refer Slide Time: 43:12)

Let us say if we have a beam which is twisted, and this is my point of C g and the load is
applied here the line of application of load is the shift. So, this is what we call as u t. So, in
this case the additional moment will be equal to F into u tP into u t if it is supplied in the
bottom flange that is below a shear center if it is applied below the shear center still also the
shift is u t, but M additional now is minus F u t oh sorry P u t.

So, this causes a destabilizing moment, but this causes a stabilizing counteracting moment.
So, this is better. Now with the factors C 1 and C 2 the three factor formula can calculate the
elastic critical moment M cr for double symmetric beams under various loads and point of
application.

908
(Refer Slide Time: 45:11)

So, the three-factor formula or equation can be used to compute M cr for a double symmetric
section of beams under various loads and points of application of load because C 1 and C 2
are accounted for this, but still the equation is valid only for double symmetric section. Now,
if you want to use this equation for single symmetric section then an additional factor C 3 is
used to apply this equation for single symmetric sections.

(Refer Slide Time: 46:43)

So, there are various factors given as correction factors which can be summarized here. So, I
should say correction factors for uniform moment conditions. Let us say the loading

909
condition bending moment diagram k z C 1 C 2 and C 3. Let us say I have a simply supported
beam with UDF of some value, we know the bending diagram is like this.

k z is 1.0 for 1.0 k z 1.12, 0.45 and 0.525 or for C 1, C 2, C 3, if k z is 0.5 for the same
condition of loading C 1 is 0.97, C 2 is 0.36 and C 3 is 0.478. If the beam has got central
concentrated load P the bending moment diagram is triangular for k z of 1.0 and 0.5 these
values are 1.35, 0.59 and 0.411, 1.05, 0.48 and 0.338.

If the beam is subjected to a load which is l by 4, l by 4, l by 4 and these are the loads which
is l the bending moment diagram goes like this and for a condition of 1.0 and 0.5 these values
are given as you see on the screen. So, now, friends we are going to now perform a design
check for lateral torsional buckling which we will discuss in the next lecture.

(Refer Slide Time: 49:59)

So, in this lecture we have learnt about the three-factor formula or equation to compute Mcr to
design for beams under lateral torsional buckling. The lecture and the tables have got good
reference Advanced Steel Design book written by me for CRC, then there is another book
Design Aids of Topside written by me for CRC. So, these two books are standard references
which can help you to learn more about the design against LTB which is explained more in an
elaborate manner with an example given from euro codes.

Thank you very much and have a good day, bye.

910
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 54
Design check for LTB - 2

Friends, welcome to the 54th lecture of the course Advanced Steel Design. We are now
continuing to discuss about the Lateral Torsional Buckling. In this lecture, we will learn how
to do design check for lateral torsional buckling. Now, if you look into the international code
as well as Indian code and see how these codes handle the design for members against lateral
torsional buckling.

(Refer Slide Time: 01:01)

For example, we will discuss first with Eurocode 3. The general procedure is described in this
code. So, the general procedure to design against LTB is by introducing a buckling factor.
And, if this buckling factor psi is less than or equal to 1 considered to be, so now, this
buckling factor reduces the capacity of the member. To be specific, when I am using this
check for lateral torsional buckling; so, I should say for lateral torsional buckling, this factor
is indicated as psi LT.

So, LT stands for lateral torsional bucking. Remember, this factor also accounts for all the
effects which can decrease the load capacity. We have seen many factors which affects the

911
capacity of the member under buckling. So, this factor psi LT will address all those concerned
parameters which will contribute to decrease in load capacity or the moment carrying
capacity of the member under buckling.

(Refer Slide Time: 03:40)

Now, the moment capacity is given by the following equation, MED /Mb R D is less than or
equal to 1, call this equation number 26. Because I am doing the continuation what we had in
the previous lectures, were M b R D is actually given by the factor of psi LT which is W y f y
by gamma M 1, equation number 27. Now, let us see W y is a bending resistance
corresponding to the cross-sectional classification of the member. I think we all know even
the cross section is classified, according to the section being used.

And of course, f y is the yield strength of the material, gamma M 1 is the partial safety factor
and psi LT is a reduction factor for lateral torsional buckling. Now, this reduction factor can
be calculated by two methods. One you can have a general method and you can have an
specific method to calculate psi LT. We will see both, specific method depends depending on
the cross section of the member. There are two procedures available in the literature.

912
(Refer Slide Time: 06:47)

Now, for different class of members certain conditions also apply varying from class 1, class
2, class 3 and class 4 of the members, following conditions also apply. Let us see what they
are, W y will be W pl, y for class 1 and 2, W y is taken as W el, y for class 3 and W y is W
effective y for class 4. This is as per the advice of the Eurocode. Now, Eurocode also gives us
a table for correction factors of moment gradient conditions.

(Refer Slide Time: 07:56)

So, this is actually the correction factor for moment gradient conditions given by the code.
One can see different types of possible bending moment distribution and you can also look

913
into the factors C 1 and C 3 depending upon psi f less than or more than 0, for different
loading conditions. We will use; we will be using this correction factor when you do a
problem. Now, let us discuss the general method.

(Refer Slide Time: 08:47)

As we just now said the code suggests two alternate methods to check the stability design of
the beams under bending. The designer can choose any method appropriate to the
specification as recommended by the local codes, both follow a similar procedure. The
general method recommended for lateral torsional buckling of beams is given based on the
flexural buckling of columns. So, the general basis for LTB is flexural buckling of columns.
The parameters are derived for the behavior of members of the bending.

So, Eurocode part 3-1-1, part [Link] is what we are specifically referring to. This specific
factor gives the buckling factor. This is given by the equation as written from the code, 1 by
phi LT of square root of squares of phi square LT - lambda square LT for psi LT less than 1
because 3 28 , where phi LT, LT stands for Lateral Torsional bucking is 0.5 times of 1 +
alpha LT of lambda LT - 0.2 + lambda bar square LT, equation 29.

So, one can very easily write a note here, looking at these two equations one can say the
buckling behavior is strongly dependent on the slenderness ratio of the member.

914
(Refer Slide Time: 12:25)

The slenderness parameter psi LT is given by the following

equation, sorry not psi LT lambda bar LT. So, lambda bar LT is given
by √Wy fy/ Mcr. I think we know all these terms earlier; we have
been discussing this. For example, we know M c r is the elastic
critical moment based on classical buckling theory. We can also
recollect that this M c r accounts for the critical moment effect,
imperfections, and residual stresses.

There are correction factors know, three factors which we were discussed in the last lecture.
There is one important comment which we like to make here. The comment is the code does
not provide any information on determining M c r. But, it states that it should be based on the
gross cross section, the loading condition, real moment distribution and lateral restraints used
to control the lateral torsional buckling.

915
(Refer Slide Time: 15:18)

So, we can also make another comment on this procedure, that the general method is a
standard procedure which accounts for the bending moment distribution which is applicable
only to determine the slenderness parameter. How can I say that? The slenderness parameter
depends on W y and we have seen a table which talks about the moment gradient conditions
for different moment distributions. So, we can very well say that the general method is
concerned about the moment distribution pattern and the boundary conditions as applicable
determine the slenderness parameter.

Now, alpha LT which is called the imperfection factor, in fact, we should say not alpha LT,
we should say alpha LT into lambda LT - 0.2, I mean that is the whole factor here. This
accounts for imperfection factor. This aids to reduce the LT capacity , for lambda bar LT
greater than 0.2 , that is why it is - point. So, in simple terms the lateral torsional buckling
curves will have a plateau length of 0.2 are only applicable under this condition.

We have also seen the table which gives me the bending moment distribution and the
corresponding factors for that which can be used in parallel to the lateral torsional buckling
curves which will be using them in the analysis.

916
(Refer Slide Time: 18:23)

Then, we will talk about the buckling curves, because that is a part of the general method.

(Refer Slide Time: 18:40)

I will copy this figure and put it here. The one what you see on the screen is a
non-dimensional lambda bar which accounts for the reduction factor. It is a buckling curve
given for different designation as a, b, c, d and a naught curve. Now, these curves are based
on; these curves are based on experimental studies conducted to investigate the beam under
bending. Lateral torsional buckling curves applicable different cross sections were developed
subsequently right.

917
So, it has got a basis of the basis is followed from Ayrton Perry formula for lateral torsional
buckling of beams. This was developed much early in 1886. So, this states very interesting
condition that buckling will occur in beams before bending failure, if the beam is slender. So,
buckling failure preludes the bending failure. So, therefore, in doing this analysis, the
buckling factor becomes a very important parameter.

(Refer Slide Time: 21:29)

So, the buckling factor which is psi LT is equal to 1 by lambda bar square LT, where lambda
bar LT is given by W y f y by M c r. So, from this we can say M c r is 1 by lambda bar square
LT of W y f y, can you say that? We call equation number 31. Now, we can see in this curve
there are different grades like a naught, a, b, c, d. They are different choice of buckling curves
which depends on the cross section, type of steel quality and manufacturing method. Let us
have how do they vary. Let us see that.

918
(Refer Slide Time: 22:58)

Please look at this table. The choice of buckling curve depends on this is choice of buckling
curve. How can you say this choice of buckling curve? I will just show you, you can see here
there are varieties of curves being recommended, buckling curves. It depends on various
parameters. Number 1, it depends on the cross-sectional layout, 2, it depends on type of steel.

To be very specific, Eurocode puts us quality of steel, S 235, S 275, S 355 and so on , you can
see here. It also of course, depends on manufacturing method. If you carefully look at this
code, the values of the choice of curve depends on whether it is manufactured in the standard
procedure, whether it is welded I section that is built up section, if it is an hollow section or a
box section and so on.

So, depends upon how you manufacture, how do you fabricate the choice of curve is
different, number 1. Number 2, it also limits the thickness of the flange. You may wonder
why flanges thickness is limited. We already know flange governs buckling, is it not;
thickness governs buckling, we already saw that in the three parameter formulae in the last
lecture about ecr Mcr.

So, we got that value and understood how it is. Also remember friends, if it is buckling about
major then the curve choice is different, it buckles about minor then the curve choice is
different. Of course, these curves are close to each other for a larger value of lambda, but still
there is a variation. And, these are all depend on experimental finding.

919
Furthermore friends, there is also a ratio of the geometric section dimension that is depending
upon h by b, h is the overall depth of the section, b is the overall breadth; depending upon h
by d h by b for the section you also have different curves. So, that is very interesting to see
how they are classified well in detail in the code. Furthermore friends, in a short summary we
can say that if it is an let us say type of section, limit value and the recommended buckling
curve.

If you look at this summary very quickly from this table for I or H sections which are rolled,
if h by b is less than or equal to 2, we can look for a, if h by b is more than 2, look for
buckling curve b. If it is welded, then if h by b is less than 2 equal to 2, look for curve c. If it
is more than 2, you can look for curve d. For all other sections, you only use curve d. These
are all the recommended buckling curves as per the code, as a summary extracted from the
table what you see on the left side.

Furthermore, for every buckling curve, one has also recommended the imperfection factor
alpha LT called imperfection factor. So, for a curve this factor is 0.21, for using b buckling
curve it is 0.34, for c curve it is 0.49 and for d curve it is 0.76. These are all type of buckling
curves. Having said this, if the theoretical buckling factor is plotted for different values, the
relationship is actually given as you see in figure 324 ; as you see in this figure it is plotted.
The corresponding limited values are also given in the table as you see here.

There is a curve what we have; this is the curve what we have. This is the curve what we
have and there is a table recommended by the code. And, the buckling curve depends on
various parameters and depending upon the type of manufacturing and the section limitations,
we have recommended certain curves by the code whose imperfection factor is also given.
So, now we can say very clearly some summary, very quick observation on this.

920
(Refer Slide Time: 30:02)

For higher slenderness values, the resistance offered by the beam approaches the value of
perfect beam. Furthermore, it is also seen the influence of imperfection and residual stresses,
though they are considered in the buckling curves they are small. Furthermore, the design
buckling factor are much lower than the ideal buckling factor. So, it gives me a conservative
design of this general approach.

(Refer Slide Time: 31:50)

There is another method by which we can use, this is second method. The second method is
useful to calculate the lateral torsional buckling factor. Let us quickly see how this method

921
works. This is referred as a special case method in the code. This method is applicable for
beams with hot rolled an equivalent welded section only. Now, what is the variation? The
major change is reflected in the lateral torsional buckling curves.

So, according to this method, the lateral torsional buckling reduction factor which is psi LT is
given by 1 by phi LT + squares of phi LT square - beta lambda bar square of LT subject to a
value which is minimum of 1 by lambda bar square LT; equation 32, where lambda bar LT is
the slenderness parameter.

(Refer Slide Time: 34:39)

Further, phi LT can be computed using this expression 1 + alpha LT of lambda bar LT -
lambda bar LT at 0 + beta lambda bar square. If you look back the previous equation, you
would remember and recollect immediately that this was flattening of 0.2 earlier. So, that is
now changed significantly by the alternate method as I am discussing now. We call equation
number 33 is it not, yeah 33.

So, now considering these two equations 32 and 33, the shape of the lateral torsional buckling
curves significantly changed compared to the general method. Now, the changes are
reflected, changes in the LT buckling curves are reflected by two parameters. One parameter
is referred as beta which is seen here. The other parameter is alpha bar sorry lambda bar LT
coma 0 which is indicated here.

922
So, now the code also imposes a maximum limit for lambda bar LT naught. The maximum
limit imposed by the code is 0.4. Therefore, designers do not have to count on reduction
factor for lateral torsional buckling, if lambda bar LT is less than 0 and 4 is the slenderness
range. And, there is a minimum value of beta, recommended by the code which is 0.75.

So, friends as we now understand the lateral torsional buckling curves are significantly
modified by the second method compared to the general method. The second method is
called a special case method which is significantly different from the general method
because, of these two parameters and the limits are given here.

(Refer Slide Time: 38:02)

In addition to this, Eurocode also proposed a modifier called as f factor method, code also
recommended a modifier called as f factor. Now, this f factor is useful to address the load
distribution, because you see in the general method the bending moment distribution was
addressed by a tabular value depending upon thickness of the flange, method of
manufacturing, all those things were there.

So, the second method that is also taken care of by a factor called f factor. So, now using this
f factor, the lateral torsional buckling resistance of a beam by the second method is given by
M b R d, which is psi LT modify W p l, y f y by gamma M 1 which is given by psi LT by f,
W p l, y f y by gamma M 1; call equation 34 right, yeah 34. Now, in this equation you will
notice there is a factor f.

923
This factor is given by 1 - 0.5 times of 1 - k c of 1 - twice of lambda bar for LT - 0.8 the
whole square and this is limited to 1. It is a reduction factor. In this, k c is called correction
factor. What does it correct? It corrects the bending moment distribution value, depending
upon the bending moment distribution pattern this factor is recommended. So, now let us see
how this factor looks like in the code.

(Refer Slide Time: 41:46)

For various types of bending moment distribution, let us say the correction factor k c for BM
distribution. So, we say if it is uniform, where this is + 1 k c is 1.0. If it is triangular varying
from - 1 to + 1, then k c is given by 1 by 1.33 - 0.33 psi. If it is distribution of this order 0.94,
if it is distribution of this order which can be a fixed beam 0.9, if it is eccentric on overhang is
0.91.

So, now psi is a factor, how do you work out this factor? If I have the fixed end moments of
this order and span moment of this order, if we call this as M and this as psi M. So, psi is a
fact is a factor and this is my maximum M naught.

924
(Refer Slide Time: 43:47)

So, if we have a triangular distribution, then k is 0.86, if you have a distribution of this order
which is linear 0.77. These are all for point loads, I think you can recollect the bending
moment diagrams very easily related to the type of load. We can say here psi is a ratio
between the end moments, that is what we have written there. It can vary anywhere from - 1
to + 1, it is a ratio between the end moments.

(Refer Slide Time: 44:39)

Now, code also advises very interestingly a statement; lateral torsional buckling effect can be
neglected provided lambda bar LT is lesser than lambda bar LT 0 and M ED by M c r is lesser

925
than equal to lambda bar square LT 0. If this condition is satisfied, then you can neglect the
lateral torsional effect in the design. Now, the buckling curves are also having different
designations.

(Refer Slide Time: 45:31)

Based on the special method depending upon the type of section and depending upon the
limits, the buckling curves are also recommended, and this is available in EC 1-1. For I or H
sections which is rolled, if h by b is less than or greater than 2, then I can use b or c. For I or
H section, which is welled, that is built up section, if h by b is less than or more than 2, I can
use c or d curve. This is recommended by the code, when you start using the second method,
special case method.

926
(Refer Slide Time: 46:39)

So, friends in this lecture, we learnt what are the factors that affect design for lateral torsional
buckling. 2 methods by which the buckling capacity can be computed which is one is a
general method; one is the specific case or special case method. We have also seen there are
different buckling curves which as a, b, c, d etcetera which can be selected to compute the
buckling capacity depending upon the type of section, the sectional dimensions, bending
moment distribution, type of lateral restraints and quality of steel etcetera.

So, in the next lecture, we will take up a design example and solve this using Eurocode as
well as Indian code and illustrate step by step the decision procedure to design or check the
beam against lateral torsional buckling.

Thank you very much. Have a good day.

927
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 55
LTB example problem

(Refer Slide Time: 00:20)

Friends, welcome to the lecture 55, where we are going to do an example problem for Lateral
Torsional Buckling. We will take a simply supported beam for a span of 5.7 meters subjected
to two kinds of loads. One load is a dead load, which is about 9.58 kilo newton per meter.

And over and above we have an impose load which is uniform distributed, and this is
imposed load otherwise called as live load of intensity 6.25 kilo newton per meter. We will
call this end as A, this end as B. Now, we are going to use a specific kind of trial beam
section. The beam section what we chose is UK beam 356 by 171 by 51, which is S 275
grade.

928
(Refer Slide Time: 02:23)

So, for the benefit of the users, let us show the cross section of the beam. This one the cross
section of the UK beam 356, 171 and 51. Based on this, let us write down certain dimensions.
Now, the first step is to estimate the maximum bending moment. We call this as M ED. Let us
see what the factored loads w are, which is going to be 1.35 into 9.58; this value plus 1.5 into
6.25.

So, I am using a different factor for live load and dead load which is going to be 22.31 kilo
newton per meter. So, I can quickly find out M ED as w l square by 8 because simply
supported section. So, it is going to be 22.31 into 5.7 square by 8, which will give me 90.6
kilo newton meter, this is my demand.

Now, for the section shown on the screen the properties are like this; h is 355, that is why the
section is called 356 355; b which you see here is 171.5. These are all millimeters friends; tw
which you see here thickness of the web is 7.4 millimeter, tf thickness of the flange which
you see here is 11.5 millimeters, r is 10.2 millimetre which is actually the radius of curvature
being used for fabricating the ends of this rolled steel section.

These are available in the standard tables. They are referred as UK steel sections. Similar also
available Indian code. We have Indian code sp 61 which will give you the section properties
and so on. So, the steel grade which is being used is S 275, which means sigma Y is 275.
Yield strength is 275 newton per mm square. We will use fy because that is a symbol used in
Euro code.

929
Let us check quickly, what is the maximum thickness which is permitted for this kind of steel.
The maximum thickness in the section available is 11.5. This is lesser than 40 that is a check,
that is a check. So, this is according to I will write here this according to EN 1993-1-1 2005;
table 3.1, that is what I have used.

(Refer Slide Time: 06:20)

Now, let us come to section classification. Section classification depends on square root of
235 by f y which comes to 0.92. This is as per table 5.2, I am using the same code. Let us talk
about the outstanding flange. The outstanding flange which is under uniform compression.

We will work out the c value is b minus t w minus 2r by 2, which will be b is 171.5, thickness
of the web is 7.4 minus 2 times of 10.2, that is the radius given; divided by 2 which gives me
71.85 millimetres. So, a factor c by t f is 71.85 by 11.5 which is 6.25 which is less than 9
epsilon. So, this gives me that it is class 1 element. This is as per table 5.2 of the code. So, it
is a class 1 element now.

Let us talk about the internal compression member. So, now, we are considering web in
bending. Bending can also have compression, bending compression. Let us compute that for c
is equal to h minus 2tr sorry 2tf minus 2r which is going to be 355 minus 2 times of 11.5
minus 2 times of 10.2 which gives me 311.6 millimetres. Now, c by t w which is 311.6
divided by 7.4 is 42.1 which is less than 72 epsilons, which also guarantees and classifies as
class 1 element. This is as per table 5.2.

930
(Refer Slide Time: 09:28)

Having said this, we are now using; so, now, the section used for trial is class 1 section, we
have we have clarified that. Now, let us do the second step which is to compute the moment
of resistance which is M b R d. We will use class 6.2.4 of Euro code 1993-1-1 2005,
according to this M b R d is given by W pl y into f y by γ M naught. So, we will say 896 is
capital W, 275 is my yield strength. And γ m naught is taken as 1.0 to 10 power minus 6.

So, so many kilo newton meter which gives me 246.4 kilo newton meter. Now, friends please
note, the capacity which is moment of resistance is higher than the demand, which is M, what
we have used the symbol M ED. So, one can say. So, since that is 260 sorry 246.4 is greater
than 90.6 kilo newton meter, we can say the section is safe against bending. We have checked
it; we are now checking a trail section.

931
(Refer Slide Time: 11:57)

Now, let us check against LTB. So, to check against LSTB, LTB, we are going to check the
elastic critical moment. Now, to compute the elastic; so, based on the support conditions and
the loading applied at the top flange level. So, now, support conditions we should say simply
supported, load application top flange. We can calculate M cr. We already derived this
equation.

M cr is given by that is three parameter formula

0.5
2 ⎧⎡ 2 2 ⎫
𝐶1
π 𝐸𝐼𝑧
⎢
(𝑘𝑧𝐿)2 ⎨⎢⎣
⎩
( )
𝑘𝑧
𝑘𝑤
𝐼𝑤
𝐼𝑧
+
(𝑘𝑧𝐿) 𝐺𝐼τ
2
π 𝐸𝐼𝑧
(
+ 𝐶2𝑍𝑔 − 𝐶3𝑍𝑗 ⎥
⎥
⎦
2⎤
) (
− 𝐶2𝑍𝑔 − 𝐶3𝑍𝑗 )⎬
⎭
. This equation

only we wrote the previous lecture, please refer back. So, now we have to use the figures and
tables.

So, the conditions are the beam is simply supported. Therefore, the effective length factor
which is k w which is also equal to k z which is 1.0. Now, since the load is separated in top
flange Z g will reduce the section by half. So, what does it mean? Z g will be actually h by 2,
which is 355 by 2, which is 177.5 millimeters. C 1 from table which you already said is 1.12,
C 2 from table is 0.45.

932
(Refer Slide Time: 15:47)

Then, from the table we will also know I z for the section which is 968 centimeter to power 4,
I w is 0.286 d m to the power 6, this is given in UK steel tables. I am just taking the value
directly from there. So, 1 decimeter is 10 centimetre. Let us have this connection here. I tau is
23.8 centimeter 4 and E of course, is 2.1 10 power 5 newton power square. And G is 77 kilo
newton per m square.

Now, we know all the values. See here we wanted to know I am just checking we wanted to
know E sorry C 1, we have E we have , I z we have k z, we have L of course, we have. I w
we have, I z we have, G we have, I tau we have, Z g C 2 we have, Z g we have. So, you can
compute these values directly and I will get M cr as 121.9 kilo newton meter,.

933
(Refer Slide Time: 18:06)

So, the next step is to estimate the buckling factor, which is psi LT. This depends on
slenderness. So, slenderness parameter which is λ bar LT is given by square root of W y f y
by M cr. We already have this value see here W y we have , f y we have, now M cr also we
have. So, let us substitute that. It is going to be square root of 896 into 10 power 3 into 275
divided by 121.9 10 power 6. I get this value as 1.423, you can please check this.

So, now we consider this section as hot rolled. So, we will use this special case method.
There are two methods by which we can do. So, we know from the table for h by b which is
equal to 355 by 171.5 which is 2.05 greater than 2, we should use buckling curve C, am I
right?

We have seen this in the last lecture. For rolled section for h by b more than 2, what buckling
curve we should use? A special case method. Please refer to your notes. Now, from the
buckling curve for curve C, we can directly find α LT as 0.49.

934
(Refer Slide Time: 20:31)

Then we can compute ϕ LT which is given by 0.5 times of 1 plus αLT times of λ bar LT
minus λbar LT 0 plus beta λ bar square of LT. So, let us work out ϕLT in this case 0.5 times of
1 plus 0.49 1.423 minus 0.4 plus 0.75 1.423 square, which gives me 1.510. So, now, I can say
psi LT is given by 1 by ϕ LT plus square ϕ square LT minus beta λ square LT. And which
should be the minimum of 1.0 and 1 by λ square LT.

So, now I get psi LT as 1 by 1.51 plus square of 1.51 square minus 0.75, that is the beta value
into 1.423 square, which comes to 0.42 which is less than 1 and also 1 by λ bar square LT is 1
by 1.423 square which is 0.493, right. So, the value what we got is 0.42. So, I can now say
psi LT can be taken as 0.4. It satisfies both the conditions no.

935
(Refer Slide Time: 23:01)

Now, we have the beam under uniform distributed load. So, for this load distribution or
specifically, for this bending moment distribution, from the table which we discussed in the
last lecture, I can find k c as 0.94. Please see the table. Now, I want to calculate the f factor
because I am using the second method.

So, f factors required; f is given by 1 minus 0.5 times of 1 minus k c of 1 minus twice of λ bar
LT minus 0.8 square should be less than or equal to 1. Let us check 1 minus 0.5 times of 1
minus 0.94 of 1 minus twice of 1.423 minus 0.8 the whole square, we get this value as 0.993
which is less than 1,. So, f factor is now computed.

936
(Refer Slide Time: 24:28)

Therefore, we can then find the M o R which is actually M b of R d. M b of R d is given by

psi LT by f W pl, y f y by γM 1. We will substitute. 0.42 that is psi LT, is it not? 0.42 f factor
is 0.993 and W p is 896 10 power 3 into 275 by 1. We already said γ m is 1 for this kind of
distribution. So, this value comes to be 104.2 kilo newton meter.

So, now the demand is M ED which is 90.6 kilo newton meter. The capacity M b, R d which
is 104.2 kilo newton meter. Since the demand is lesser than the capacity, the beam is safe
against lateral torsional buckling. So, we are just checking whether the beam is safe. So, what
is the comment here? No additional lateral restraints are required. Otherwise, you have got to
provide lateral restraints as we discussed in the previous lectures.

937
(Refer Slide Time: 26:39)

Let us solve the same problem using IS code. We are using IS 800-2007, that is the steel
design code by checking the section. So, we will take the section as ISMB 400 at 61.6 kg per
meter. So, let us say the beam is simply supported over a span of 8 meters. The total moment
under uniform distributed load is 50 kilo newton meter, f y 250 grade is to be used. So, check
the beam for lateral torsional buckling that is the question asked. Let us do that.

(Refer Slide Time: 28:08)

Let us first see the section properties. Area of the cross section which is 78.46 centimeter
square which is 7846 millimeter square, depth of the section h 400 mm, thickness of the

938
flange 16 mm, breadth of the flange 140 mm, thickness of the web 8.9 mm. This can be seen
from steel tables,. Moment of inertia about the minor axis also given in the table which is
622.1 centimeter 4 which is 622.1 10 power 4 mm. I xx which is about the major axis which
is 20458.4 which is 20458.4 into 10 power 4 mm 4.

(Refer Slide Time: 29:37)

Then we will do the section classification. The first step is section properties, we know f y is
250 newton per meter square. Let us look into table 2 of IS 800 2007. So, epsilon is given by
square root of 250 by f y which is equal to 1 and b by t f in our case is going to be over an.
So, 140 by 2 divided by 16 which is 4.375 less than 9.4 epsilon 4 epsilon equal to 1,. And d
by t w which is 400 minus 2 into 16 divided by 8.9, which is 41.35 which is less than 85
epsilon,.

So, friends in this equation if you note this b actually is b f by 2. Look at figure 2 of IS 800.
And of course, d is d w, only the web. So, now, based on these data we can classify the
section as , classify as plastic section. Furthermore, h by b f is 400 by 140 which is 2.86
greater than 1.2 and t f is 16 mm less than 40 mm.

939
(Refer Slide Time: 32:12)

So, from Table 10 of the code buckling class about XX axis is a and buckling class about YY
axis is b. So, then the next step is to find out M cr that is elastic lateral buckling moment.
Now, the section is symmetric. It is prismatic and it is simply supported , because we should
have the bending moment distribution also these are the conditions.

For these conditions, M cr is given by square root of pi square E I y by L LT square of G I tau

plus pi square E I w by L LT square. This is as per class 8.2.2 2.21, sorry. So, in this let us
substitute the values. E is 2.1 10 power 5. So, we have E mu is 0.3, therefore, G is E by twice
of 1 plus mu which gives me 2.1 10 power 5 by twice of 1 plus 0.3. I get this value as 80.77
10 power 3,. So, I have G, I y already we have. Let us see what is I y 622.1.

940
(Refer Slide Time: 34:55)

So, I y we have, 622.1 10 power 4. Effective length which is L LT, is 1 point L which is 8
meters. So, this is according to table 15 class 8.3 for torsional fully restrained, but warping
not restrained in both the flanges. For this condition you have taken this effectively. So, I tau
torsional constant which is summation of bi ti cube by 3, this is for open sections.

So, let us find out for our problem, I tau twice of 140 into 16 cube by 3 plus 368 into 8.9 cube
by 3 , which comes to be 468.77 10 power 3 mm 4, warping constant. So, let say I tau we
have, L LT we have, we want to find I w. So, warping constant I w is 1 minus beta f beta f I y
h y square, this is for I sections.

So, now what is beta f? Beta f is I fc by I fc plus I ft where I fc and I ft are moment of inertia
of tension and compression flanges about the minor axis of the section. Let us compute I fc
that is very easy. So, I fc about minor axis. So, p t cube by 12, I can now say this is going to
be 3.66 10 power 6 mm 4. So, we have the same as I ft tension and compression flange is
minor axis moment of inertia is same,.

941
(Refer Slide Time: 38:30)

Having said this, let us now find what is h y which is actually the distance between the shear
center of two flanges of the cross section. So, h y in my problem will be h minus t f by 2 into
2 which will be 400 minus 16 which is 384, beta f which is I fc by I fc plus I ft which 3.66 10
power 6 by 3.66 10 power 6 into 2 which is 0.5.

Now, let us work out I w which is given by 1 minus beta f beta f I y h y square. Let us
substitute them. 1 minus 0.5, 0.5 I y 621.1 that is here, 622.1 10 power 4, h y is here 384
square which gives me the value as 2.293 into 10 power 11 mm 6,. Now, I can find M cr.

(Refer Slide Time: 40:40)

942
M cr let us copy this equation of M cr back again is given by this equation. Let us substitute
which will be equal to square root of pi square 2.1 10 power 5 that is my E value, I is 622.1
10 power 4. This is 8000 square. I am doing it in millimetres, in newtons. Then 80.77 10
power that is my G value see here , that is my G value I tau is 468.77 10 power 3 that is my
tau value plus pi square 2.1 10 power 5.

I warping constant we already have it here 2.293 , divided by 8000 square. So, I get M cr as
95.51 into 10 power 6 newton mm which will become 95.51 kilo newton meter,.

(Refer Slide Time: 43:04)

Now, let us compute the design bending moment. The design bending moment is M d which
is beta b Z p f bd. This is as per class 8.2.2 of IS 800. So, what is beta b? Which is 1 for
plastic sections. We already declared this the plastic section. What is Z p? Is the plastic
section modulus, and f bd? Is a design compressive bending stress which is actually equal to
psi LT f y by γ m naught. So, let us say what is psi LT, we already know this ϕ LT plus ϕ LT
square minus λ square LT,.

943
(Refer Slide Time: 44:29)

Φ LT is 0.5 times of 1 plus α LT into psi LT minus 0.2 plus α bar LT square α LT is the
imperfection factor which is 0.21 for rolled sections given in the code. Let us calculate λ LT
which is actually equal to square root of beta b Z p f y by M cr which is less than or equal to
square root of 1.2 Z e f y by M cr. So, let us say what is Z p; 1176163.26. This is for ISMB
400 Z ce is 1022.9 10 power 3 mm cube,.

We can refer these values from SP 6 1 steel tables. So, now, α λ LT will be 1 into 1176163.26
into 250 by 95.51 10 power 6 which will become which should be less than or equal to 1.2
times of 1022.9 10 power 3 of 250 by 95.51 10 power 6. So, now we get this value as 1.79
less than or equal to one point this is 75 79. So, it is ,. Once I have this value we can
substitute back in ϕ LT and cut ϕ LT, let us do that.

944
(Refer Slide Time: 47:20)

So, therefore, ϕ LT is going to be 0.5 times of 1 plus LT LT minus 0.2 plus λ bar square LT
which will be 0.5 1 plus 0.21 that is what my α LT is 0.21 into 1.75 minus 0.2 plus 1.75
square , which comes to 2.194. So, now can calculate this value as 1 by ϕ LT plus ϕ LT
square minus λ square which will be 1 by 2.194 plus 2.194 square minus 1.75 square which
comes to 0.284. Now f bd bending stress in compression is LT fy by γ M naught.

Now, γ M naught is equal to 1.1 which is the partial safety factor of member to resist
buckling. So, let us substitute that which will become 0.284 into 250 by 1.1 which comes
64.616.

945
(Refer Slide Time: 49:53)

So, now the design bending strength is governed by the lateral torsional buckling. So, M d is
beta b Z p f bd which is 1.0 1176163.26 into 64.616 which is 75.99 or 76 kilo newton meter,.
The applied moment is 50 kilo newton meter, the capacity is 76 kilo meter. Therefore, we can
say the beam is safe against lateral torsional buckling no additional restraints are required,.

(Refer Slide Time: 51:31)

So, friends, in this lecture, we learnt two examples of checking the design for lateral torsion
work. 1, we did the Euro code, 2nd one we did with Indian code. In both the cases, since the

946
capacity of the section is larger than the demand of the section. So, we declare they are safe
against LTB.

So, the procedure is very straight forward. It is well illustrated in both the course, equations
and supporting tables are available to you. So, only I got to practice more number of
problems. So, take some sample sections, work it out and see whether the chosen section is
safe against lateral torsional buckling,.

Thank you very much and have a good day.

947
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 56
Ice loads

(Refer Slide Time: 00:18)

Friends, welcome to the-56th lecture on Advanced Steel Design. In this lecture, we are going
to learn more about Ice Loads. Generally, sea surface or sea ice develops in winter this is
essentially due to the cooling of surface seawater if you look at a simple statistics sea ice is
about 15% of worlds ocean.

There is another fact which is bothering as a designer to us is the ice thickness and ice
coverage on the sea surface increases progressively in winter which is basically a fact which
all of us know. Further to that hardness of ice also increases with age. these are studies
interestingly made by Mahoney et al., 2004 which will be helpful for additional reading for
you.

948
(Refer Slide Time: 02:53)

Friends, during ice formation needle-like crystals of ice develop first. Let us look into the
process of ice formation needle-like crystals of ice first develop they get modified in elastic
ice crest from further freezing. Elastic you can say ice crest on further freezing. Ice deforms
more with the brittleness of the ice crest, more the brittleness of the ice crest more is the
deformation of ice. It causes separation of ice into small cylindrical pieces of about 20 meter
in diameter.

More details of this progress of ice can be seen in one of my publications 2019. Small
diameters of ice are called pancakes. They will be of diameter 3 meter. Large diameter of ice
is called ice cakes pancakes can cause impact forces on offshore structures.

949
(Refer Slide Time: 06:07)

The natural period of these impact forces increases with the increase in wave height period of
this will also increase in the presence of current. Now, Ice cakes which of larger diameter,
further freezes to form a continuous ice sheet. They are also called as ice floe.

Ice floes further freeze to form ice fields. They are extended over a large area which will be
about 10 square kilometres. Under the presence of in ice floe structural motion of ice is
governed by the wavelength. Ice floes also impact offshore structures they are governed by
the drift of velocity of ice.

(Refer Slide Time: 09:12)

950
Continuous action of wind wave and current transforms this ice sheet into pressure ice fields.
These pressure ice fields have rough surfaces. These ice sheets can also pile up one over the
other. You have an ice sheet of certain thickness; there can be another ice sheet of another
thickness they can get piled up. This is what we call as piling.

They generally pile one over the other in an irregular manner and that will result in creation
of ice pressure ridges. For example, this intersection of this piece with the bottom one is a
classical example of ice ridge and so on. Further, friends, ice sheet remains attached to the
shore such ice sheets are called shore fast ice. They are usually developed in winter.

(Refer Slide Time: 12:00)

This sheet will get detached from the shore line during spring season from the shore line, this
happen generally during spring season. Once they are detached, they are called pack ice. Pack
ice usually has a width about 10 kilometres. They remain continuous motion. The velocity of
pack ice is similar to the current velocity.

Now, let us talk about water glaziers. Now, glaziers are essentially ice shelves which actually
originate from land by this island by compaction of snow leaves and freezing of freshwater.
They result in glaziers. Flow of glazier’s results in formation of ice bergs.

951
(Refer Slide Time: 15:10)

The chunk of ice breaks due to buoyancy of seawater. Further friends, icebergs are more
governed by the wind speed current velocity and so. Temperature below and above the water
surface causes non-uniformity in the formation of ice where is the non-uniformity occurring?
The non-uniformity comes from non-uniformity in the melting of icebergs.

This therefore, results in breaking of ice unless otherwise it is not broken it will remain as a
continuous sheet. It also results in tilting of the ice bergs. It can also cause capsizing of ice
bergs.

(Refer Slide Time: 17:51)

952
Now, friends breaking of icebergs is in smaller form is called as growlers. They are also
others called as bergy bits. Therefore friends, structural design of platforms under ice loads is
governed by 1 – sheet of ice; 2 – pack ice and 3 – ice bergs. So, these are the three structural
form of ice which can cause worries the design of the platform under ice loads. Ice loads
acting on offshore structure depends on many factors.

(Refer Slide Time: 19:51)

Let us see what are they. The first and foremost factor is the structural geometry of the
platform depends upon the shape of the platform, size of the platform etcetera. So, friends
this is also reminding us that design of structures and rise load is a form dominant design
procedure. 2 – it also depends on location of the platform. It also depends on sea state at that
location. So, both are location factors. I will say this is 2 (a), this is 2 (b).

The 3rd factor that governs ice load calculation is the ice properties itself. For example,
thickness of ice, velocity with which ice is moving crushing strength of ice and lastly, the ice
structure interaction. So, ice load calculation depends on many such factors.

953
(Refer Slide Time: 21:56)

Having said this, let us ask a question. How does ice force develop? What is the analogy
behind this development? Friends, please listen when ice hits the vertical member of an
offshore platform, ice failure occurs. This failure is attributed to continuous action of wind
wave and current.

Now, this results in horizontal force of the member. This is continuous in nature and hence is
a dynamic load. Secondly, it also causes transient vibration to the platform. Why this is
developed? This is due to the fact that formation of pressure gradient during development of
ice.

954
(Refer Slide Time: 24:50)

Let us talk something about ice-structure interaction. To understand the ice-structure

interaction, let us talk about the common failures that occur due to ice loads. 1 – limit stress
failure. 2 – limit force failure. Now, failure of ice occurs and the ice interaction surface.

When the environmental forces act on the ice like wave, wind and current, their magnitude
increases and the magnitude very higher than the ice strength and therefore, ice face like
waves, wind and current have higher magnitudes then the ice-strength which causes ice
failure.

(Refer Slide Time: 27:07)

955
Let us see what are the common modes of ice failure. Ice can fail by buckling of ice sheet. 2 –
it can also fail due to crushing of ice sheet. Once the ice sheet fails ice structure interaction
surface causes movement of ice ridges around the structure.

Ice ridges will move around the structure to cause an additional impact force on the platform.
Interestingly friends, this motion of ice which is relatively important, but to the motion of the
structure reduces the force structure. So, the advantage is this process reduces the force on the
structure; to be very specific, the ice force on the structure.

(Refer Slide Time: 29:01)

So, let us ask a question what are the factors that govern ice sheet failure. Ice sheet failure
depends on ice thickness. It also depends on the dimension of the structure to which ice
impact is considered. Dimension moment I say is size; shape of the platform. The third one is
a strain rate which is very important.

956
(Refer Slide Time: 30:04)

There are other modes of failure; 1 is they failed by shear, 2 radial and circumferential
cracking, 3 – it can also be due to creep, 4 – the failure can also be due to spalling. Now,
interestingly friends, when the ratio of ice thickness to leg diameter of the structure is very
less that is thickness size is very more compared to the leg diameter of the member diameter.
If this ratio ice thickness by member diameter is very less and this will induce larger
cross-section of the member to withstand ice force.

So, as a result the diameter of the member is very large. Encountering ice force is very large.

(Refer Slide Time: 32:14)

957
Having said this, ice failure generally takes place by creep at lowest strain rate or by crushing
at highest strain rate. So, friends, the ratio of ice thickness to the diameter of the member – if
it is increasing then the failure modes will shift from conventional failure to radial or
circumferential failure. So, the failure modes will get shifted from the conventional modes of
failure to formation of radial or circular cracks.

Having said this, one should use carefully the suitable empirical relationship to estimate ice
force on offshore structures.

(Refer Slide Time: 34:36)

In the literature if you look at the ice loads there is a term called level ice. Level ice actually
induces random vibration on the members of the structure. Out of all the scenarios of failure,
the most sensitive parameter is crushing ice failure. Crushing ice failure is considered as a
worst scenario because crushing ice failure induces maximum ice force on the member.

958
(Refer Slide Time: 36:11)

So, friends as a summary, in this lecture we started learning more about ice loads; we learnt
also about the process of ice formation; we also learnt about the failure modes of ice; we also
learnt about the crushing ice failure which is a very important phenomena of failure because
this induces maximum ice force on the member.

Have a thorough reading, I can recommend a textbook and Vinodhini 2019 which is talking
about analysis and design of offshore complaint triceratops platforms, where the theory about
formation of ice in the process of formation and the factors governing the ice load are clearly
illustrated in the textbook.

Thank you very much and have a good day.

959
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 57
Ice spectrum

(Refer Slide Time: 00:19)

Friends, welcome to lecture-57, in which we are going to learn more about the Ice Loads and
Ice Spectrum. In the last lecture, we understood that crushing ice failure is considered to be
the worst scenario as this induces the maximize ice force on the member. Let us say, what
would be the effect of ice on complaint platforms. In general, complaint platforms undergo
harmonic oscillation under higher ice-velocity. Under the crushing ice failure 3 distinct ice 4
modes occur in the response.

The main factor which governs this failure mode is ice-velocity. So, the response of the
structures and ice loads especially under crushing ice failure, can be done by 3 methods. One,
you can take this process of quasi-static process. You can also handle this as a steady
condition of loading. Further, you can also handle this as a phenomena of random vibration.

960
(Refer Slide Time: 03:14)

So, Karna et al, 2007 expressed the dynamic ice forced model. They studied the ice force
effects are narrow and wide members. And they said that member response decreases with
increase in diameter of the member. The study conducted by them postulated certain
limitations of the study. One, the study is not suitable for intermittent ice-crushing program.
Two, it does not include non-linear effects in the development. It also governs the approaches
to estimate ice force.

(Refer Slide Time: 05:49)

961
Let us quickly see what are various approaches to estimate ice forces. There are essentially 4
approaches. One is experimental studies on the scaled model. Now, this method uses scaling
laws to examine. This method shows a very large disagreement when ice is modeled through
the scaled value that is especially in case of CS because you know the large area, but a very
minimum thickness. So, the scaling law actually affects the correlation between the excellent
studies and the ideal behavior in the numerical model.

The second approach is using numerical studies. Numerical studies have a very serious
limitation in terms of validation of results with respect to experiment. The serious limitation
is validation of the study with experiments which is quietly big challenge in the present
scenario of research.

(Refer Slide Time: 08:05)

The third approach what generally people follow is field studies. And the fourth one is
addressed as data mining. In sense, one can examine the previously constructed platforms in
the locations where ice loads are occurring. They are called ice-infested areas.

The second approach could be estimate ice loads through field measures. Nor through the lab
scale, nor through the full numerical scale, but establish these measurements through free
observations. There are many codes of practice available to estimate ice loads like Canadian
code, European, Norway, and USA. There are many international codes which help you to
estimate ice forces.

962
(Refer Slide Time: 10:05)

Out of all of them, an approximate approach which is quite simple in frequency domain is as
below. The excitation force, caused by ice is modeled as sinusoidal, pseudo-excitation force.
Subsequently, the response of the structure is characterized using the appropriate transfer
function.

Friends, literature has developed an ice spectrum for, a narrow, conical structure. In the
absence of extensive field studies and experimental studies, one can use this ice spectrum to
compute the loads on narrow conical structure.

(Refer Slide Time: 12:10)

963
The ice spectrum S of f, one sided power specialist function is given by following equation 1.

Let us see what are the factors which govern this equation. Let us say A and B are called ice
spectrum constants, A is 10 and B is 5.47 obtained from the experimental studies.

So, A and B are known. A is known, B is known. F bar naught is the force amplitude on the
platform exerted by ice. So, 𝐹0 is also now known. 𝐹0 is given by an additional equation

which is

We call this equation number 2.

So, in this equation C is derived from experiments as 3.4. C is derived. σ𝑏 𝑜𝑟 𝑓 is called

bending strength of ice which is approximately equal to 0.7 megapascal. So, σ𝑏 is known. h is

the thickness of ice, so h is known. D, the diameter of the ice core, so D is known. 𝐿𝑐 is called

characteristic length of ice which is further given by

Call the equation number 3.

964
(Refer Slide Time: 15:26)

So, friends, in this equation E is modulus of elasticity of ice or ice-sheet to be very specific.
Usually, taken as 0.5 Gpa, which comes to 0.5 into 10 power 9 newton per mm square; it is a
newton per meter square. Of course, ρ𝑤 in this equation, E is known, h is of course, thickness

of ice, ρ𝑤 is density of sea water which is 1025 kg per cubic meters.

So, in this equation, we carefully look at of course, g is acceleration due to gravity which is
9.81 meter per second square. So, we know g. So, we know 𝐿𝑐. If you go back and check we

know 𝐿𝑐. So, equation 2 all values are described.

Let us go back to equation 1, we still have something called T bar. That is not explain.
Something called T bar, so let us do that. T bar is called period of ice which is given by

𝐿𝑏
𝑇= 𝑣

Where 𝐿𝑏 is a breaking length of ice-sheet and v is the velocity of ice-sheet.

Usually, the breaking length is approximately 4 to 10 times of ice-sheet thickness. Usually,

that is the usual practice. So, let us go back to this equation and see, I know now T bar, we
would like to know the values of alpha, beta, del, and gamma. So, let us see. Alpha is 0.64,
beta is also 0.64, gamma is 3.5 and del is 2.5 based on experiments. So, del, gamma, alpha,

965
beta are known. So, it is the frequency f, is a variable to be plotted. Let us quickly use the
MATLAB program, to plot the spectrum.

(Refer Slide Time: 19:22)

Equations are already given to you. Now, I will use the MATLAB program to plot the
spectrum.

(Refer Slide Time: 19:39)

So, this is a program being used for plotting ice spectrum.

966
(Refer Slide Time: 19:47)

This was a joint effort made by research scholar working under me.

(Refer Slide Time: 19:57)

That is a program which is being written available in the textbook.

967
(Refer Slide Time: 20:00)

You can download and run the program in MATLAB.

(Refer Slide Time: 20:05)

Let us run the program and see how typically an ice spectrum looks like. Ice spectrum looks
like I will see on the screen which is having a very sharp peak at a very low frequency and
then it dies down instantaneously. So, it is a narrow band spectrum which is meant only for
the Bohai Gulf which is a typical input used for calculating ice loads on offshore platforms in
arctic region.

968
So, the typical ice spectrum. So, this is my f, this is my spectral density which goes as high as
100, is an ice spectrum for Bohai Gulf.

(Refer Slide Time: 21:32)

The ice low which is calculated from the ice force spectrum is based on the field data. There
are 2 major factors in the ice load. The first factor is the ice-velocity. The second factor is the
thickness of the ice-sheet. Let us quickly see the probability density function of ice-velocity
and ice-thickness.

(Refer Slide Time: 22:45)

969
It is assumed that both follow Rayleigh distribution and logarithmic distribution. Probability
function of velocity v is given by

( )
2
𝑣 𝑣
𝑃𝑣(𝑣) = 826.55
𝑒𝑥𝑝⁡ − 1653.102

equation 1 a. Probability density function of thickness h is given by

𝑙𝑛(ℎ)−1.8671 2⎤
𝑒𝑥𝑝⁡⎡⎢−
1 1
𝑃ℎ(ℎ) = ⎡ ⎤⎥
0.55ℎ 2π ⎣ 2 ⎣ 0.5503 ⎦ ⎦

Let me put the square carefully here. So, one is a logarithmic distribution, other one is a
Rayleigh distribution.

(Refer Slide Time: 24:50)

Let us now quickly see what are the effects of ice load on offshore platforms. Platforms are
generally designed to be ice resistant when they are commissioned in arctic region, when
commissioned in arctic region. Ice loads are generally caused or ice loads induced,
ice-induced vibrations.

Further, ice-induced vibrations will cause fatigue in tubular joints. It also causes lot of human
discomfort to people on board. It also causes flange loosening of pipes. The effect of ice force
on an offshore platform is generally measured from deck acceleration.

970
(Refer Slide Time: 27:05)

So, while doing this analysis, it is assumed that the deck acceleration is a stationary process.
It can be also taken as a Gaussian process, and it is a narrow banded process. Rayleigh
distribution is used to determine the acceleration response of the deck.

Now, the probability condition of the acceleration is given by

𝑃𝑎(𝑎)=
𝑎
2
σ𝑎
𝑒𝑥𝑝⁡ −( )𝑎
2

2σ𝑎
2

equation 2. Where, σ𝑎 is given by the integral S u double dot spar distribution of df. S𝑢¨𝑢¨ is
2

called as the power spectral density function of the deck acceleration which is developed due
to ice loads on offshore platforms.

971
(Refer Slide Time: 29:30)

If one knows the power spectral density function of the deck acceleration, then if the
ice-velocity is divided into j groups and ice-thickness is divided into i groups. Then, the
environmental parameters that govern the ice load can be defined by N cases, where N will
be i cross j cases. So, hence, 𝑝𝑖𝑗 shall give the occurrence of probability of kth ice case, in

working on the ice load.

(Refer Slide Time: 31:12)

972
So, therefore, ℎ𝑖, ℎ𝑖+1, 𝑣𝑗, 𝑣𝑗+1 is given by the probability of occurrence of specific ice case

amongst the N cases. So, 𝑝𝑖𝑗 specific case, if you want to look at, it will be integral ℎ𝑖 to ℎ𝑖+1,

𝑝ℎ (h) dh into integral 𝑣𝑗 to 𝑣𝑗+1, 𝑝𝑣 (v)dv equation 3.

(Refer Slide Time: 32:20)

So, we can also explain this using a flowchart. So, let us discuss a flowchart being used to
assess the failure probability under ice loads. So, starting with the first factor which deals
with uncertainty of ice environment based on the uncertainty conditions and models one can
select the randomness of the parameters of ice load.

What parameters are we looking at? If you remember, thickness and velocity we are looking
only at these 2 parameters. So, once we know these, they can we can then choose the failure
probability of the platform deck.

On the other hand, if we have variability in the performance of ice resistant design, most ice
resistance platform, based on the variability we can try to compute ice-induced vibration
characteristics. From the ice-induced vibration characteristics, one can also assess the failure
probability. Once the failure probability is assessed, then one can look at short-term ice case
under dynamic loads acting on the platform.

973
(Refer Slide Time: 34:58)

One can also assess, the failure probability of the platform itself directly. So, first we have to
define a predetermined threshold value. Let us call this value as a L. Once this value is
defined, then one can calculate power spectral density function of the displacement under
pseudo-excitation which is 𝑆𝑢𝑢(f). Then, one can define a specific case of ice. Let us call this

Nth case.

For the Nth case one can generate the ice spectrum that is for specific velocity and specific
thickness, for a specific velocity and thickness. You can generate a spectrum which is 𝑆𝑘(𝑓).
𝑡ℎ
This is for the 𝑘 ice case. Once this is generated, from this one can estimate the probability
𝑡ℎ
of occurrence of 𝑘 ice case which is called as 𝑝𝑖𝑘.

974
(Refer Slide Time: 37:16)

Once this is assessed, then one can compute the power spectral density function of
acceleration response of the deck which we call as 𝑆𝑎𝑎, k(f). Having computed this one can
2
further also compute the distribution parameter which is σ𝑎 and k. We need to assume that a
𝑡ℎ
conditional exceedance occurs, the conditional exceedance of probability of failure given 𝑘
case for the predetermined cases. So, under this assumption one can now compute failure
probability which is at p f , k.

(Refer Slide Time: 39:01)

975
Further, if k is equal to N, then what happens?. If k is equal to N, if it is not equal then look
for the Nth ice case and say k is k plus 1 and run the program. If it is yes, then find the
maximum of p of k for k equals 1, 2, till N, to obtain the failure probability based on
short-term ice case. Friends, ice load can be estimated by 2 methods.

(Refer Slide Time: 40:06)

One, by failure probability approach which we just now discussed the flow chart. This
approach considers the probability of occurrence of failure event due to ice-induced
vibration. The second one is based on the expected loss. In this method, all this method can
account for failure probability and consequences due to ice-induced vibration.

976
(Refer Slide Time: 41:49)

We all know that ice loads can cause deck acceleration. It is accepted as a scenario of severe
consequence of the ice load. So, the occurrence of probability of ice case is 𝑃𝑖𝑘 and condition

of exceedance of this depends on ice-thickness.

So, very importantly, the occurrence of probability of ice case which is 𝑃𝑖𝑘 and the condition

of exceedance of probability of occurrence which is P f, k both depends on ice-thickness.

Ice-velocity plays no role on the other. That is a very important factor which you must
realize. When ice acts on the breaking ice cones, broken ice pieces are cleared up. Ice force is
periodic in nature which results in ice-induced vibration.

So, friends we know that estimation of ice force has a basic assumption on a narrow conical
structure. An ice is formed by breaking of ice. Ice force a random process which is controlled
by 2 parameters, thick classifies and velocity of motion of ice break.

977
(Refer Slide Time: 44:02)

𝑁
𝐹(𝑡) = ∑ 𝑓𝑖 (𝑡 − 𝑡˙𝑖)
𝑖=1

So, the ice force, F of t can be given by summation of i equals 1 to N, 𝑓𝑖 (t-𝑡˙𝑖). Equation, will

call this 3 a. Where in this above equation N is the length of the loading function and 𝑡˙𝑖 is 0,

whereas, t 1 0 is summation of j equals 1 to i equals 1, J j for i greater than 1; 3 b. The

function 𝑓𝑖t minus i, this function is ice force model for any chosen event of ice failure. How

do you define this?

978
(Refer Slide Time: 45:32)

The ice force model which is 𝑓𝑖(t) is chosen for a specific event of ice failure. So, 𝑓𝑖(t) is

given by 6 𝐹𝑜𝑖 of t by 𝑇𝑖, if 0 less than t less than 𝑇𝑖 by 6. It is equal to 2 times 𝐹𝑜𝑖 minus 6

times 𝐹𝑜𝑖(t) by 𝑇𝑖. Now, not into t, 𝐹𝑜𝑖 into t is also into t for 𝑇𝑖 by 6 less than t, 𝑇𝑖 by 3.

Otherwise, it is 0 for 𝑇𝑖 by 3, t, 𝑇𝑖. We call this equation as 4.

(Refer Slide Time: 46:58)

979
So, 𝐹𝑜𝑖 for i equals 1, 2 etcetera denotes the random ice force amplitude. 𝑇𝑖 denotes the

random ice force periods. Ice force amplitude, ice force periods are weak correlation. These 2
parameters have a very weak correlation which are experimentally verified. So, they can be
considered as an independent parameter in the model.

(Refer Slide Time: 48:01)

So, variance of ice force is defined as expected value of

2 2
[
𝐷[𝐹(𝑡)] = 𝐸 𝐹 (𝑡) − 𝐸 [𝐹(𝑡)]]
Equation 4 b. Where, E denotes expected operator. Further,

2 2
2 𝐹0 σ𝐹𝑜
[
𝐸 𝐹 (𝑡) =] 9
+ 9

2
𝐹0 and σ𝐹𝑜 are mean and variance of the force amplitude F naught.

980
(Refer Slide Time: 49:24)

2 2
𝐹0 σ𝐹𝑜
𝐷[𝐹(𝑡)] = 12
+ 9

That is the standard deviation of ice force amplitude is about 0.4 of the mean ice force
amplitude, is approximately 0.4 times of mean ice force amplitude. That is a correlation
computed by Qu et al, in the year 2000.

(Refer Slide Time: 50:36)

2
𝑚0 = 𝐷[𝐹(𝑡)]≅0. 1 𝐹0

981
(Refer Slide Time: 50:52)

So, friends, in this lecture, we learnt how to compute ice spectrum, how to handle ice as a
random process which is controlled by the thickness and velocity of ice movement. We have
also learnt how to compute the spectrum with ice load by characterizing this as a power
spectral density function.

So, kindly go through these lectures again. And try to understand the parameters involved in
ice load generation, which will be helpful for learning ice load as a process, which excites
force or load on an offshore platform constructed in arctic region.

Thank you very much. And have a good day.

982
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 58
Blast resistant design - 1

(Refer Slide Time: 00:19)

Friends welcome to lecture-58 of the course Advanced Steel Design. In this lecture we are
going to learn more about Blast resistant design.

983
(Refer Slide Time: 00:53)

Blast resistant design is a part of structural strengthening of buildings. It is one of the

measures to minimize the risk to people and facilities from hazards of accident explosions
happening in the industrial units. So, the primary objective of the blast resistant design is
essentially to minimize risk to people. Now, let us ask a question what are the steps involved
in Blast resistant design.

The 1st step is quantifying the blast over pressure that could result from accident explosions,
step number 2 establish the design blast load. So, from the blast over pressure compute the
design blast load. The 3rd step would be set up the structural performance requirements and
the 4th step could be design the building to withstand these load very important within the
preset performance requirements that is very important there is no blast proof structures.

For a blast resistant design, you set up the performance requirements of the structure. It can
be immediate occupancy it can be life safety and so on. There are collapse prevention like
earthquake engineering. There are many preset performance requirements which you can set
as a designer. Now you have to design the building under the blast loads, so that only those
preset performance requirements are met not beyond that.

So, you have to lay down the conditions of requirements of the building under the blast loads
which is pre laid. So, there is no post investigation of requirements after the blast test
happened. So, it is very important that the blast resistant design has a prerequisite, you must

984
set the standards the deflection requirements, the strength requirements, the performance
criteria of the structure which you are going to design for a specific blast load.

Once it is preset then you will design the building. For that computed blast load to meet these
preset requirements.

(Refer Slide Time: 06:20)

Now, the question comes what is a necessity for a blast resistant design. So, what is the
necessity? Now different kind of structures which are exposed to accidental cases need to be
designed for this. So, we have to say here that industrial structures, chemical plants, nuclear
power plants, offshore systems or offshore structures and other strategic structures which are
generally exposed or probably to accidental explosions need to be designed for blast loads.

𝑅𝑖𝑠𝑘 = (𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑜𝑐𝑐𝑢𝑟𝑟𝑒𝑛𝑐𝑒)𝑎𝑐𝑐𝑖𝑑𝑒𝑛𝑡 × (𝑐𝑜𝑛𝑠𝑒𝑞𝑢𝑒𝑛𝑐𝑒𝑠)𝑎𝑐𝑐𝑖𝑑𝑒𝑛𝑡

One may ask a question what is the probability of this blast or this explosion happening that
is a good question. So, let me answer this. Even though the probability of occurrence of
explosions or accidents are very low, the consequences are severe. Friends risk is actually the
product of probability of occurrence of an accident and the consequences if it happen. Even
though the probability is low, but the consequences are very severe.

So, it has got a very high-risk factor. So, accidental explosions since they have very high-risk
factor they can result in heavy financial loss to the asset. It can cause potential loss to the

985
environment and public. So therefore, there is a necessity for blast resistant design of certain
class of structures you do not have to design all structures for blast resistant design. You are
going to see what is the probability of that kind of system which can or which will be
exposed to accidental explosions.

If you have a probability of occurrence of these accidents happening in that particular

building or structural system, then it should be designed for blast tested or blast loads. We can
give a very classical example of two well-known accidents especially in offshore structures.

(Refer Slide Time: 10:38)

Offshore platforms which caused a very severe consequence and environmental impact: 1 is
piper alpha happened on 6th June 1988, the 2nd accident happened on the platform deep
water horizon which is 20th April 2010. Though the frequency of these accidents is very
large they are not very close, maybe it is about 25 years.

So, they occur very severe consequences at a very large interval of time. there is no
recurrence of these accidents, because offshore platforms are example, are well protected in
safety perspective and the firefighting systems are very effective and the design is very
comprehensive. So, accidents are generally avoided.

But still if there is a probability even though the occurrence of this accidents are very very
low, but the consequence are very severe. Therefore, one emphasizes the fact that such
structures must be designed for blast loads to avoid fire and explosion which can be very

986
devastating. Furthermore, the concentration of such fatality points prompt towards the need
for design under blast loads.

So, it does not pose any hazard by itself. So therefore, friends structural strengthening or
design to resist blast loads is set as an important objective of design of strategic structures.
This is to achieve the appropriate level of blast protection. Let us do this way. The next
question obviously comes what are the objectives of blast resistant design.

(Refer Slide Time: 14:56)

What are the objectives? The primary objective of providing blast resistant design for
buildings are as follows: 1, personal safety; 2, one can initiate what is called as controlled
shutdown; 3, financial or economic considerations. So, these are the three factors which
control the objective of blast resistant design. Let us see one by one in detail. Let us talk
about personal safety. Blast resistant design should provide a level of safety for persons
inside the building.

The previous accidents, which has happened for example in offshore platforms show clearly
that there were serious injuries caused because of collapse of buildings onto the persons. So,
previous incidents or previous accidents show that there were severe injuries on personnel
due to collapse of buildings under the accident explosions. So therefore, the primary
objective is to reduce the probability that the building itself becomes a hazard that is a very
important statement here.

987
So, the primary objective is to reduce the probability that the building itself becomes a hazard
under an explosion, that is why we say blast resistant design is also said as structural
strengthening under blast loads, the concept is same. Because in many incidents it has been
reported that the building collapsed and personnel inside the building were severely and
fatally injured.

To reduce the probability of building itself becoming a hazard when subject to explosive
accidents, we are going to strengthen the building under these kind of blast loads.

(Refer Slide Time: 18:38)

The second one is what we say as controlled shutdown. The control shutdown refers to the
objective of avoiding the loss of the control of process units that are not involved in the event.
The focus here is to avoid the loss of control of process units. Let us say process units which
are not involved in the event. At least those should be protected. Those are involved there can
be a mechanical a collapse there can be a complete collapse of the equipment’s and plants
fine that is a devastating loss.

But those which are not involved in the event should be at least protected no? So, an incident
in one unit should not have a cascading effect on the other. So, we need to have a stepwise
shutdown of the process units such that the units which are not involved in the event should
be protected. The third one is economic consideration presenting the economic factors help us
to prevent or minimize the financial loss.

988
So, here the objective is to minimize of course to prevent which may not be possible, but to
minimize the financial loss due to accidental explosions. Facilities with critical or essential
equipment’s which are expensive and those require long lead time to procure should be
protected. Equipment’s which are expensive those require a long lead time for
commissioning. Those are critical, should be protected. Let us now go ahead and discuss
what are the blast resistant requirements.

(Refer Slide Time: 22:55)

Blast resistant requirements are generally established through standard practice. There are
specific guidelines issued by the factories prevention or protection act which are legal which
should be followed. There are recommended practices for satisfactory design of structures
under blast loads. For example, offshore structures under blast loads are governed by API
recommended practice 2 FB, B stands for blast loads.

Let us ask a question what are those factors which will govern or influence the blast resistant
design requirements. What are the factors that govern the blast resistant design requirements,
if not we know the codal provision let us at least try to understand the factors. The factors are
distance from the blast source where is the blast source probably located and where is the
building.

Two, what is the criticality of the function of the building is it a strategic building, is it a
normal structure etcetera. So, what is the strategic requirement of this? Thirdly and most

989
importantly, what is the expected occupancy and what is the rate of occupancy? That is an
any given point of time how many people will be present to the building.

Because the occupancy rate can vary with time if it is an office building there can be specific
working hours beyond which the occupancy will be minimum or during which the occupancy
will be maximum and so on. There are structures where there are no occupancy, but the
storage of valuable equipment’s. For example, server units, where computers are being kept
and so on, which are very highly strategic important.

So, criticality of function of the building, expected occupancy in terms of personal life and
safety and of course, the distance or location of the building from the probable epicenter of
the blast source. I am sure people will identify and know how to identify what are the
probable blast sources, what are the reasons for industrial explosion which we will also
discuss in detail in the coming part of this lecture. But blast sources can be easily identified in
a given industrial plant.

So, what is the distance of the building which you are going to design which needs the design
for blast resistance? One of the important factors is the distance. If it is located much far
away from the probable blast source, then that unit or that part of the building need not be
designed for blast loads. So, we are not looking here the recommended guidelines specifically
for any or by any code, we are looking the factors that govern these guidelines.

Now friends, there is a big confusion in the glossary of terms blast proof and blast resistant.
Let us see what it is.

990
(Refer Slide Time: 28:38)

Friends, every structure which is being designed be it strategic non-strategic has some level
of blast resistance, because blast resistance is a part of strengthens structural requirement and
buildings constructed by steel essentially do possess this inherent characteristic. So, I can
now say very clearly here every steel structure possess a certain level of blast resistance.

Now the blast resistant design is sometimes referred as blast proof, but it is very important to
know that blast proof is a mis concept. It is nonrealistic to provide an absolute level of blast
protection. There is always some probability that a design basis event can be exceeded.

Since there is a probability or there will be a probability of exceedance of design basis events,
it is not possible to make the structure blast proof. So, one can only do blast resistant design.

991
(Refer Slide Time: 31:48)

In the context of offshore structures or heavy industrial units blast load can lead to partial or
total collapse. It can also result in huge loss of life and it can also cause serious
environmental impact. This is true in case of offshore structures. So therefore friends, critical
issue in blast resistant design in blast resistant design is the fact that the available space
should be protected from complete destruction.

Because available space in offshore facilities very limited and very expensive, therefore we
need to protect that space. So, we are looking for a significant mitigation of blast effects;
significant mitigation of blast effects. So, we are not looking for complete blast proof concept
by looking for blast effect mitigations on the design.

So, one of the important factor as a designer which you can consider is by keeping the
strategic units away or locate them well away far away from the probable blast source, both
horizontally and vertical we call this as horizontal vertical zoning. So, one can also ensure or
one can create a vertical or horizontal zoning to in improvise safety against blast loads. Of
course, you cannot do this for the entire structure, you can do it for part of the structure.

Consequentially, the following parts of the structure should be however protected. what are
they? So I will ask a question.

992
(Refer Slide Time: 36:04)

What are the critical units that need to be blast resistant design? Ask this question. What are
these units? One, control rooms both electrical and mechanical. Two, living quarters where
the occupancy rate is very high. Escape routes, evacuation facilities, critical structural support
systems and safety critical items.

Such as for example, fire water lines, supporting units of these supporting units of fire mains
etcetera. They should all be designed to remain as blast resistant. As we discussed we realize
that one need to mitigate the effect of blasts.

(Refer Slide Time: 38:03)

993
Now, the question comes how to mitigate the effect of blasts. We can cite again an example
offshore facility. Generally, in strategic structures such as nuclear power plants, offshore
facilities, dams’ etcetera, blast walls are constructed integrally connected with the existing
structure that is one option we have.

The second option we have is people use lightweight manufactured walls I should say
prefabricated walls which are added to the existing old constructions. You may ask me a
question why we are adding light weight manufactured walls. During blast friends when it
starts blasting the walls, it results in flying off the construction particles.

If the wall is made of steel or made of concrete of large jellies etcetera which is very heavy in
shape and size the blown off particles during blast load can also cause cascading injuries to
the public, to the personnel located around the epicenter of the blast flow as well as to the
plants and equipment’s. So, generally people recommend try to use light weight units please
understand lightweight units does not mean we compromise strength against safety.

It is only a partition wall which is going to cause or which is created which is added later to
create a cite of a blast wall. The blast walls are generally constructed with light weight, but
with strong integrity and strength. Having said this the blast resistant design has got 2 stages,
generally it is carried out in 2 stages. One is of course the preliminary design, the next one is
the detailed design. Let us quickly see what do we cover in preliminary design.

(Refer Slide Time: 41:52)

994
In preliminary design following parameters are concerned. 1 - nominal over pressure, 2 -
impulse loads. In detailed design one has to do a CFD analysis to solve the equations
describing the gas flow, turbulence and the combustion process. Let us ask a question what is
the design philosophy of blast resistant design, we will discuss this in next lecture in detail.
Now we will take up a summary and quickly summarize what we have learnt in this lecture.

(Refer Slide Time: 43:41)

In this lecture we understood about blast loads, the necessity of design structure for blast
loads, what kind of structures are designed for blast resistance. We have also learnt the
methodologies of preliminary and detailed design for blast resistance. We are in the process
of learning the design philosophy for blast resistant design.

Thank you very much and have a good day.

995
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 59
Blast - resistant design - 2

(Refer Slide Time: 00:15)

Friends, welcome to the 59th lecture on Advanced Steel Design course. In this lecture, we are
going to learn about Blast Resistant Design, let us say 2 because in the earlier lecture we
discussed about Blast Loads.

So, when you talk about the design philosophy of blast resistant design, the starting point for
the design of a structure of blast test and design is to consider the layout and arrangements.
So, the first step is to choose or to design the layout of the building units and the
arrangements of plants and equipment’s in the unit. That is a first step.

So, there is a specific objective which you have to keep in mind when you choose the layout
of the building which you want to design for blast resistance. The objectives are, one need to
design what needs protection. One need also imagine how damage or injury will be caused.
One need to also consider how the structure can be arranged or I should say how the units of
the structure can be arranged both vertically and horizontally to get the best blast resistant
protection.

996
(Refer Slide Time: 03:29)

Having said this, when we talk about blast loading, let us define what is a blast load. Blast
loading actually is caused by an explosion. An explosion is a very rapid chemical process
producing transient air pressure waves. These transient air pressure waves are called as blast
waves.

Now, these blast waves are governed by various parameters, parameters that govern these
blast waves are, one, peak over pressure; two, duration of the over pressure. Now, both these
parameters may vary with respect to the distance of the unit from the source of explosion.
Therefore, in blast load computation, it is important to recognize the scenarios that could
have severe outcome and take into account of these in the design. Therefore, friends, the blast
pressure waves will also be reflected and refracted by the buildings.

997
(Refer Slide Time: 06:46)

Blast waves are very intrusive and all sides of the structure can be subjected to over pressure.
So, if we have a structure, all the 4 sides can be subjected to over pressure. The peak over
pressure decreases as the wave moves further. So, as the blast wave progresses in the forward
direction, the peak over pressure decreases. Therefore, in simple terms, if the unit is located
far away from the blast source effect of peak over pressure on the unit will be minimum.

(Refer Slide Time: 09:08)

Further, reflection or let us say reflected waves, from the building generates multiple wave
fronts which can also cause further damage. So, let us say I have a building, there is a blast

998
wave source, so the fore is affected and there is a reflected wave. Even these waves can also
severe effect on the structures.

Further, when the blast wave impinges directly onto the face of the building it is reflected
from it. And these reflected waves will also have consequences and they can also cause
damage.

(Refer Slide Time: 11:12)

The effective pressure applied on to the face of the building is magnified when this occurs.
Under this reflected wave, the effective pressure applied on to the surface gets magnified. If
the facade is to survive without breaching, it must be designed to resist this pressure. So,
therefore, building facades should be also designed to counteract these reflected waves.

999
(Refer Slide Time: 12:44)

Let us quickly see qualitatively how the estimate the response of a building to a blast load.
So, blast loads are extremely intense, but acts for a very short duration. The energy
absorption capacity and the dynamic characteristics of the structure governs the response of
the structure under such explosions.

In general, a tall building, whose frequency or let us say natural frequency is low and has a
long response time in relation to the load. So, if the frequency is low as in the case of tall
buildings, it takes longer time to respond to the blast load.

(Refer Slide Time: 15:29)

1000
Especially, steel structures can absorb lot of strain energy. This is because of its ductility
property. Further, flexible components can absorb higher energy contents. So, friends, in blast
resistant design, flexible components are helpful because they resist blast loads through
elastic and plastic strain. Further, this is the first point we have. The second point, a higher
mass causes less energy to be imparted to the system.

In case of larger explosions, the structure as a whole becomes affected by the blast waves. So,
in such systems, the whole structure gets affected by the blast waves. In blast resistant design,
we have a concept called protected space. Let us see what is that.

(Refer Slide Time: 18:13)

What do we understand by protected space? Protected spaces or areas within the structure
which are hardened to protect the occupants and equipment’s against the effect of external
explosion. So, these are spaces within the building which are strengthened to withstand the
building unit from external explosions. For example, protected spaces could be control
rooms, units of critical functioning, etcetera.

Protected spaces should be located utilize the benefit offended by the offer with geometry.
So, protected spaces should be located by taking care or taking the advantage of the building
geometry and structural characteristics of the building. Having said this, in the design of
protected spaces, one need to quantify the blast capacity. But please note, the blast capacity of
protected spaces is highly limited and confined only to that space, not for the entire building.
That is important to remember.

1001
(Refer Slide Time: 21:28)

So, protected spaces therefore, should be enclosed on all sides by walls and door openings. It
should not become extensively hot during occupation. So, there should be enough ventilation
and air circulation within the space.

It should possess sufficient cooling capacity to remove the heat loads of the space. So, one
should be able to maintain both the temperature and the humidity conditions of this protected
space. So, protected spaces should be designed by the extreme care considering the whole
facts. Having said this, let us see what are the principal parameters which will govern the
blast resistant design.

1002
(Refer Slide Time: 24:18)

The principal parameters to be defined for blast resistant design are as follows. One is peak
side-on positive over pressure, indicated as 𝑃𝑠𝑜. So, the second is positive phase-duration

indicated as 𝑡𝑑 and the corresponding positive impulse indicated as 𝐼𝑜. This is one set of data.

The second set of data is peak side-on negative pressure, which is suction pressure which is
called as again 𝑃𝑠𝑜 with a negative sign. Negative phase duration is also the 𝑡𝑑 and the

associated impulse which is also 𝐼𝑜. In addition, other parameters also important.

(Refer Slide Time: 26:36)

1003
The third set of parameters is peak reflected pressure called as 𝑝𝑟, then peak dynamic

pressure called as 𝑞𝑜, shock front velocity which is u, and blast wave length which is 𝐿𝑤. The

secondary parameters can be always derived from the primary blast wave parameter. So,
these are called as secondary parameters.

Let us talk about peak reflected pressure, indicated as 𝑝𝑟. When the blast wave hits the

surface, it is reflected back because the building is causing obstruction for progress of the
blast wave. So, when the building opposes progress of the blast wave, it is reflected.

The effect of this reflection is that the surface will experience a pressure more than the
incident side-on value. So, the pressure will now increase. The pressure intensity will be
more.

(Refer Slide Time: 29:23)

So, therefore, the magnitude of reflected pressure is amplified using a ratio. So, we should
say the reflected pressure is given by reflection coefficient of 𝑝𝑠𝑜. Let us say equation 2.

𝑝𝑟 = 𝐶𝑟𝑝𝑠𝑜

Let us say equation 1, no problem, equation 1. Where, 𝑝𝑟 is reflected pressure wave, 𝐶𝑟 is the

reflection coefficient.

1004
Which depends on, which depends on the peak over pressure and angle of incidence of the
wave front related to the reflecting surface. It also depends on the type of the blast wave.

(Refer Slide Time: 31:41)

𝑝𝑟
𝐶𝑟≅ 𝑝𝑠𝑜

Having said this, let us say for peak over pressure up to 138 kilopascal following equation
provided by Newmark helps to compute the reflection coefficient. So, 𝐶𝑟 is given by 𝑝𝑟 by p

so, is approximately equal to, you can see from this equation, will call say equation number 2,
which is equation 2 (a).

𝐶𝑟≅2 + 0. 0073𝑝𝑠𝑜

Alternatively, for different type of blast waves 𝐶𝑟 can be also obtained from the TNO Green

Book. The duration of the reflected pressure depends on, the dimensions of the reflecting
surface, surface roughness, up to a maximum time approximately equal to the positive phase
duration of the wave.

1005
(Refer Slide Time: 34:30)

So, therefore, we can say time duration of the reflected wave is less than equal to time
duration of the blast wave, and depends on the O parameters. Upper limit of this corresponds
to total reflection of the entire blast wave without any diffraction.

Let us talk about the next parameter which is peak dynamic pressure, talk about this
parameter peak dynamic pressure which is given as 𝑞𝑜. The peak dynamic pressure causes a

blast effect, and it is caused due to the air movement around the building.

(Refer Slide Time: 35:51)

1006
Friends, in this lecture, we learned about various parameters that govern the blast wave
characteristics. We also learned about the design philosophy of blast resistant design and how
the flexible units are better. We have also learnt something called protected space, and where
and how a protected space should be used in a blast resistant design structure.

We will see further details in the next lecture about the Blast Resistant Design.

Thank you very much. And have a good day.

1007
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 60
Blast - resistant design - 3

(Refer Slide Time: 00:21)

Friends, welcome to lecture 60 of Advanced Steel Design course wherein we are going to
discuss about Blast resistant design I put this as lecture-3 in Blast resistant design. We are
talking about another parameter which is peak dynamic pressure given as 𝑞𝑜 this is essentially

caused due to the air movement around the building blast waves when propagate get mixed
with airspace around the building and that is resulting in peak dynamic pressure.

Wind pressure also varies and influenced by the intensity of peak over pressure.

1008
(Refer Slide Time: 02:35)

In case of low over pressure range with normal atmospheric conditions, the peak dynamic
pressure can be calculated as below.

2
2.5 (𝑝𝑠𝑜) 2
𝑞𝑜 = (7𝑝𝑜+𝑝𝑠𝑜)
≅0. 0032(𝑝𝑠𝑜)

calls 3 a, where p 0 is called ambient atmospheric pressure.

Having said this the net dynamic pressure is therefore given by

𝑞𝑛𝑒𝑡 = 𝐶𝑑. 𝑞𝑜

where 𝐶𝑑 is called the drag coefficient which is influenced by the wind pressure present

around the building. The drag coefficient depends on shape and size of the obstructing
surface.

1009
(Refer Slide Time: 05:16)

For example, in case of a rectangular building, the drag coefficient is given by for front walls
it is plus 1.0 for sidewalls rear walls and roof it is given as minus 0.4. Let us talk about the
3rd parameter which is shock front velocity which is indicated as (u). In a free field the blast
wave originated from an explosion travel either at the acoustics speed or above the acoustic
speed in a propagating medium.

So, in general blast waves will travel the acoustic speed of the medium or sometimes even at
a higher speed.

(Refer Slide Time: 07:03)

1010
In a low-pressure range and for normal atmospheric conditions the shock front velocity u is
given by

0.5 𝑚
𝑢≅345(1 + 0. 0083𝑝𝑠𝑜) 𝑠

equation number 5. Let us talk about the 4th parameter which is blast wavelength given by 𝐿𝑤

the propagating blast wave at any instant of time extends over a limited radial distance.

This can be extended as shock pressure waves and travels outward from the explosion source.
The pressure is largest at the front and reduces over distance.

(Refer Slide Time: 10:01)

In low pressure range and normal atmospheric conditions as given by Newmark.

𝐿𝑤 = 𝑢 𝑡𝑑

Now let us talk about idealized blast wave parameters. A blast wave profile is generally
idealized as below to simplify the blast resistant design procedure.

If this is my time axis and this indicates the pressure axis, then idealized shockwave where let
me extend this line. This is shockwave. This value is called 𝑝𝑜 then this distance is what we

call as pso and this duration is 𝑡𝑑. So, an idealized shockwave for design purposes.

1011
(Refer Slide Time: 12:55)

When you talk about the pressure wave time scale and pressure, so, again called as 𝑡𝑑 this is

𝑝𝑜 and this is called pso and this is idealized pressure wave. A pressure is also simplified

using equivalent shock load. So, let us do that. We know idealized pressure wave and looking
at this peak and considering this value an equivalent load is established and this is shifted at
this junction.

This is the blast load the equivalent shock load. So, this is what we say as idealized
equivalent pressure load. Now, let us talk about determination of vapor cloud design over
pressures.

1012
(Refer Slide Time: 15:37)

Design blast loads are usually supplied by the owner input to the designer by the owner of the
building. It is custom designed or custom developed blast load. It is based on the desired
hazard level that the building should sustain. These hazard levels depend upon the material
and the process used for the purpose. So, let us talk about what is design over pressure.

The design over pressure can be stated in two ways. The simplest is all buildings should be
designed for a peak reflected over pressure which is over pressure of x kilo pascal a peak side
on over pressure y intensity in kilo pascal and for a duration of z milliseconds. It is a general
statement is one way of prescribing the design over pressure.

1013
(Refer Slide Time: 18:48)

The 2nd way of prescribing design over pressure is to specify over pressure intensity and
duration based on the expected or probable explosion that could occur in the structure from
the potential source. Additionally, one can also conduct site specific studies to define the
design over pressure then if you ask me a question what should be the basis for developing
such site-specific studies, what should be the steps involved in developing such approach.

The site-specific study approach should focus on the hazard parameters and quantify the
design pressure based on explosion hazards. There are various steps involved in this.

(Refer Slide Time: 21:17)

1014
Let us see what are the steps involved in developing site specific design over pressure. Define
the worst release scenario that is step number 1 based on that form the explosion cloud using
emission models and dispersion models. Friends, there are different models available to
define the explosion cloud. Please see the book written by me on Health Safety and
Environmental Management to Petroleum Industries.

So, HSE book by Wiley and authored by me. You can have more details about the book in a
home page. This will give you more information about various emission and its dispersion
models that can be helpful to define the explosion cloud. In step number 3 calculate the
amount of energy contributing to explosion. In step number 4 calculate the blast overpressure
parameters.

(Refer Slide Time: 24:21)

Having said this, let us talk about the methods to find the blast over pressure parameters.
There are two methods. One is the Strehlow curves suggested by Baker. The second one is
multi-energy method based on TNO 1985. Both methods provide family of curves based on
the explosion strength, the flame speed and r explosion strength. These curves are used to
select the dimensionless parameters which are then unscaled to determine the actual over
pressure.

So, these curves are helpful to determine the actual over pressure.

1015
(Refer Slide Time: 26:39)

Over pressure can be determined at any point on the structure on the structure based on its
distance from the source and then can be applied to the entire structure. The structure is very
large an average over pressure can be applied to the surface. Normally a building should be
designed considering the potential blast waves from any horizontal direction but not all
directions simultaneously. Please understand that. Building should be designed to withstand
blast waves considering the waves as horizontal.

If you consider directional waves then one should not include waves from all directions
simultaneously that is the point, one at a time. There is something called a common criterion
which is helpful to determine this over pressure. Commonly used criteria are actually based
on SG-22 and CIA approach. I will give you this reference for the benefit of the learners.

1016
(Refer Slide Time: 29:42)

SG-22 is siting and construction of new control houses for chemical manufacturing plants,
Safety guideline Manufacturing Chemists Association Washington DC 1978. The second
could be CIA approach which is an approach to the categorization of process plants hazard
and control building design published by Safety Committee of the Chemical Industry Safety
and Control Council, Chemical Industries Association London, 1992.

Both documents specify at least two blast over pressures for the buildings spaced at 30 meters
from the vapour cloud.

(Refer Slide Time: 33:18)

1017
It says that high pressure, short duration, triangular shock loading side on over pressure of 69
kilo pascal and duration 20 milliseconds. As far as low pressure is concerned it says a low
pressure, wave of long duration, triangular loading side on over pressure of 21 kilo pascal for
100 milliseconds duration. (Refer Slide Time: 35:04)

Let us now talk about blast loads on buildings. To design the blast resistant building loads on
the building as a whole or on individual elements should determine. So, therefore, we need to
understand the blast wave interaction to obtain these loads. So, blast load on buildings
depends on blast load structure interaction. Now let us talk about blast wave interaction.
When a blast wave strikes the building, the building is loaded by over pressure or drag forces
by the blast load.

So, let us say when the building is strike by the blast wave the building is loaded by over
pressure or drag forces caused by the blast wave. The interaction between the wave and the
structure is quite complex. When the blast wave encounters a solid surface, it starts reflecting
from the surface and as we discussed this reflecting wave depends on the geometry of the
structure, shape and size and the diffraction surface and it diffracts around the building.

Further so, when blast wave hits the surface, it is reflected. During the process of reflection
energy is transferred between the wave and the object. The incident blast wave is reflected
from the building and it starts producing further compression of air in the vicinity of the
structure on the structure. On a molecular level the surface applies an external force to each

1018
air molecule which is sufficient to give it equal momentum in the opposite direction.
Therefore, there applies the same external force to the surface.

(Refer Slide Time: 39:11)

Due to the change of momentum, pressure is locally increased above the incident pressure.
This is termed as reflected pressure. Diagrammatically let us say if this is my shock front
which is approaching the structure, it passes around the structure. So, the shock front
approaches the structure that is the first step we have. So, what happens is after it approaches
let us copy this figure, put it here.

(Refer Slide Time: 41:04)

1019
After it approaches it moves ahead, but it reflects, its reflected wave. Similarly, around the
building it moves ahead, but I will remove this because this is moved. So, I will remove this
there is only a wall the source this arrow is removed. It reflects and it creates a vertex. So,
this picture depicts shockwave reflected from the front surface and diffract over the structure.

(Refer Slide Time: 43:18)

As the wave progresses further let us copy this figure. As the wave progresses further remove
this it creates vertex and the shock wave becomes here. We call this as diffracted shock wave.
Similarly in this case vertex will be created. So, this figure indicates diffraction across the
rear surface.

Now once the shockwave moves further when the diffraction is complete, I will just copy this
figure and put it here. The diffraction is complete. I should say this figure indicates a
complete diffraction. So, the whole set of figures a, b, c, d indicate shockwave propagating
over the structure. So, this is available at TNO Green Book.

1020
(Refer Slide Time: 45:52)

So, you can see for additional reading. Please see TNO Green Book which talks about
method for determination of possible damage to people and objects resulting from release of
hazard materials which is CPR 16 E, committee for disaster prevention due to dangerous
substances, the Director General of Labor, the Hague Netherlands published in the year 1992.
Now, for design purposes we idealize this blast waves.

(Refer Slide Time: 47:41)

We are now talking about idealized blast waves. So, if you have a building, this is the
elevation. The building is subjected to some blast wave. Now, the blast wave will have three

1021
components. One will be in the fore side, other will be on the rear side and one will be in the
top. In addition, this follows pso distribution.

So, this is what I call as 𝑝𝑟, this is 𝑝𝑏 and this is pso where pso is the incident side on over

pressure, 𝑝𝑟 is the reflected pressure, 𝑝𝑎 is the average over pressure which is indicated here

and 𝑝𝑏 is the back face over pressure. So, the building is receiving pressure from all the sides.

So, these can be also marked as 𝐵𝑤 and 𝐵𝐿 where 𝐵𝑤 and 𝐵𝐿 are physical dimensions of the

building in the direction of propagation of the wave. So, one can see as time progresses, the
pressure intensity keeps on decreasing. With respect to time the peak over pressure intensity
keeps on decreasing. This is blast loading arrangement for rectangular building of size 𝐵𝐿 by

𝐵𝑤 given by Forbes in 1998.

(Refer Slide Time: 52:20)

So, friends, in this lecture we learned more details of blast over pressure. Parameters used
that govern the design and different ways or different steps to compute the blast design
pressure. We have also learnt the idealized design load for blast waves. We will see more
details in the next lecture and work out couple of examples to estimate the blast load on
buildings using these calculated values.

Thank you very much and have a good day, bye.

1022
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 61
Blast-resistant design - 4

(Refer Slide Time: 00:19)

Friends, welcome to lecture 61 on Advanced Steel Design course. We will continue discuss
Blast resistant design. I will say this is IV. In the last lecture, we discussed that the blast
loading general arrangements of rectangular building. And we understood that the blast wave
travels horizontally from left to right. And when it travels it starts imposing reflected and
back phase over pressure waves all around the circumference of the building.

Having said this, depending upon the location of the potential explosion hazards related to the
building site, the blast load strikes the building from any direction. It cannot necessarily be
horizontal it can be inclined to the building major axis. It all depends upon the source of
explosion. Therefore, an ideal assumption could be the blast wave travels horizontally, but
not necessarily all the time.

So, in this specific discussion, we are assuming that the blast wave hitting the surface of the
building is normal to its surface and it propagates horizontally along the axis of the building.
If it is so, it creates a variety of loading. Let us see one by one, how a front wall loading is

1023
created. Front wall in sense is the wall which faces the explosion source. Therefore, the wall
facing this explosion source will receive reflected over pressure. The reflected over pressure
gets amplified and that depends on various factors.

Let us see, what are these factors. It depends on angle of incidence of the blast wave which is
indicated as alpha. It also depends on what is called as the rise time which is t r. It also
depends on the side on over pressure pulse. For design purposes, the normal shock refraction
can be considered by considering or by taking.

(Refer Slide Time: 04:18)

We will write it here. So, for design purposes, the normal shock reflection condition could be
alpha 0 and t r 0 ok, they are generally assumed. However, in some cases when there is an
oblique reflection which can become critical to overall building, then the reflective pressure
from two adjacent sides should be considered after resolving it into the respective
components.

Now, the reflected over pressure decays to a stagnation level, we call this stagnation pressure
as p s over a period of time. And this time is called clearing time indicated as t c. So, p s is
given by p so + C d of q naught. We call this equation number 8.

We are continuing the equation number from the previous derivation. And the clearance time
t c is given by 3S by u which is always less than t d ok, where S is the clearing distance. It

1024
should be smaller value of B H or B W where B H is the height of the building, B W is the
width of the building.

(Refer Slide Time: 07:46)

The duration of the reflected over pressure t c should not exceed the free field positive over
pressure t d. So, mathematically t c cannot be greater than t d. Having said this, let us see the
simplified equivalent triangle of a bilinear pressure time curve. So, now, I am looking for a
simplified equivalent triangle of the bilinear pressure time curve, which I am drawing here.
So, the horizontal axis is indicated as time and the vertical axis is the pressure. We are now
talking about the loading over the front wall. So, ideally speaking, the pressure variation.

So, this is what I call as t d. And this particular time is called the clearing time. Let us project
this and this is what we call as p s, and this is p r . Now, I am looking for an equivalent
loading. Equivalent loading starts from there and I get a time equivalent, and I call this as an
equivalent loading of blast wave for the front wall. This equivalent load is computed by
equating the impulse for each load shape and the same peak pressure p r.

The impulse I w under the bilinear pressure time curve is given by the following relationship.
Half of p r - p s into t c + half of p s into t d, let me call this equation number 10. Now, by
equating the impulse for each load shape and use the same peak pressure p r, I can find we
get t equivalent time should be twice I w by p r, which is t d - t c by t r by p r of p s + t c, call
this equation number 11.

1025
(Refer Slide Time: 13:28)

Now, let us talk about the side wall loading due to the blast load. The side walls will
experience less blast load than the front wall. This is due to lack of over pressure reflection
and attenuation effect of the blast wave. You know friends, attenuation is a property of
variation of any subject with respect to distance. So, attenuation effect is decay of the blast
wave with distance propagated by the blast wave from the explosion source.

As the blast wave travels along the length of the structural member, the peak over side on
pressure will not be applied uniformly ok as blast wave progresses along the length of the
building peak side on over pressure changes. It does not remain same. We have seen that
yesterday if I have a building with respect to time it changes. So, it is not applied uniformly.
It varies both with time and with space in fact, it gets reduced.

If the length of the sidewall is equal to the length of the blast wave for a specific condition,
length of the sidewall equal to length of the blast wave.

1026
(Refer Slide Time: 16:50)

Then the peak side on over pressure reaches the far end of the wall. In that case the over
pressure at the near end turns to be ambient. So, now, this behaviour should be accounted the
decay of the side on over pressure with respect to the attenuation should be accounted.

So, to account this effect a reduction factor is used. A reduction factor is being used which
we call as C d is used. Now, this depends on the length of the structural element and the
direction of propagation of the blast wave. So, for the side walls, blast wave load can be
computed as below; p a will be equal to ok this is not C d, this is C e. We can say ok C e of p
so, reduction factor + C d of q naught. We call this equation number 12. We know p a is
effective side on over pressure.

Let us plot this if this is time versus pressure. This is indicated as t r this is what we say as p
a. And this is called as t o. This is the figure for roof and side wall load. Having said this, the
sidewall load has a rise time equal to the top, as shown in the figure.

1027
(Refer Slide Time: 21:05)

The sidewall load has a rise time equal to the top, as shown in the figure. Equal to the time it
takes for the blast wave to travel across the element. So, t r is given by B L by u equation 13.
Now, the overall duration t o is equal to the rise time of the duration + the duration of the free
field side on over pressure. Mathematically, t o will be t r + t d, equation 14. This is for the
sidewall loading.

(Refer Slide Time: 23:22)

1028
Let us look for roof load for a building with blast roof flat roof. Let us say let us consider a
building with a flat roof. So, in such cases reflection of the blast wave does not occur as the
blast wave travels horizontally. This horizontal travel was an assumption made by us.

Therefore, roof will experience side on over pressure combined with the dynamic wind
pressure which is same as that of the sidewalls. The dynamic wind force on the roof acts in
the opposite direction to the over pressure that is it is going to be acting normal upward. The
roof loading depends on ratio of blast wavelength to the span of the roof element that is the
first factor. The second factor orientation of the roof with respect to the direction of the blast
wave. Ideally speaking, we assume it is horizontal, but not necessarily, it needs to be
horizontal.

(Refer Slide Time: 27:08)

Let us now talk about rear wall load. Friends, we have seen the front wall, side wall, the roof;
now the rear wall. The rear wall load is used only to determine the net overall loading on the
frame on the frame. It tends to reduce the overall lateral blast force. The rear wall effects are
neglected for buildings where the blast loads could not occur from any direction. If blast
loads occur from different directions, then rear wall load is neglected. The rear wall blast load
lags that of the front wall.

So, the blast load on the rear wall lags with respect to the front wall by the following
relationship which is equal to B L by u. Time for the blast wave to travel the length of B L of
the building. Essentially this will be the time for the blast wave travel, the length of the

1029
building, is it not? Because distance by speed will give you the lag. The effect of peak over
pressure is similar to that of side walls, on rear walls is similar to that of side walls.

(Refer Slide Time: 30:23)

Now, according to TNO Green book, we have this figure is t r and this is B L by u and this is
p o . Now, the rise time of the positive phase that is this, is it not? The rise time is given by 4S
by u. This is t naught, the overall duration is t d and t r is given by B L by u + 4S by u further
to is t r + t d equation 15.

(Refer Slide Time: 32:11)

1030
So, now according to UFC, 3- 340 - 02 quo. This is t r this is still B L by u. And this is p o,
the positive phase rise time is S by u. Total duration is td and tr are B L by u + S by u. And
overall time is t r + t d equation 16. This got very interesting reference friends. UFC- 3- 340 -
02 is the quo for structures to resist the effects of accidents caused by explosions.

Unified facilities criteria 3 – 340 - 02 US, Department of Defense, Washington D.C 2008.

(Refer Slide Time: 34:58)

Then, let us talk about the overall frame loading under the blast wave. Frame of the entire
building will be experiencing diffraction load, which is the net load on the front, rear walls
after accounting for the time lag.

During the time lag, when the blast wave travels from the front to the back side of the
building, the structural frame will be subject to logged horizontal imbalance pressure on the
front wall. After this passes away the front wall loading is partially offset by the rear wall
loading. So, this can be experimentally I mean represented by a figure ok let us see how. So,
this is time is p.

Let us say, we have a loading for a specific time. We call this as p r and this as t c. Let us
extend this slope and call this as p s + C d of q naught. So, now, I can say this represents load
on the front wall. This is what we have discussed. When we talk about the load on the rear
wall, this was the rear wall loading.

1031
So, friends, in the rear wall loading this curve does not start from 0. Please be very careful
this starts from B L by u. Let me change this it starts from B L by u. So, we call this as time t
2, there is load on the rear wall. And we know very well this is going to be B L by u. So, now,
the net force is going to be.

So, I can say, this is the net lateral load on the rectangular building as per TNO Green book.
There will be in addition negative and rebound loading as well, let us see what that is.

(Refer Slide Time: 39:52)

Due to the negative phase, the building components will experience blast load effects
opposite to the direction of primary load. It also occurs due to the rebound of the structural
members and its components from the inertial effects.

So, negative pressure forces are generally ignored since they are negligible; however, the
structural component should be adequately detailed to perform satisfactorily for the rebound
effects. So, the negative effects, negative pressure is created due to phase lag as they are
negligible. They are ignored. Rebound effects ok, rebound loads are created due to the
structural characteristics. And they should be designed to take care of these rebound loads.

1032
(Refer Slide Time: 41:50)

There is something called leakage pressure. The potential expansion of blast loads into the
building through the opening is called as leakage pressure. As the blast wave expands
through an opening, the pressure level drops and sudden expansion into the building volume.

(Refer Slide Time: 43:34)

Let us take an example and illustrate this load estimates for a design. So, we say calculate the
blast load on the building shown below which is subjected to a blast wave travelling
horizontally. Let us say this is the building block. This is B L which is 20 meters, this is B W

1033
which is taken as 30 meters and this B H which is taken as 5 meters. The blast wave is hitting
the building like this assumed to travel horizontally.

Now, you know in this example, the blast wave is applied normal to the longer side of the
building, am I right. Peak side on over pressure is taken as 40 kilo pascal and the duration is
0.05 seconds. So, that is the loading wave we have. This is t and this is p and this value is 40
kilo pascal and this time is 0.05 seconds.

(Refer Slide Time: 46:50)

Let us work out the shock wave parameter. Shock front velocity which is u is given by 345 1
+ 0.083 of p so to the power 0.5. So, u in our case is going to be 345 1 + point zero sorry
0083 into 40 to the power 0.5 which comes to 398.17 meter per second. Let us work out the
length of the pressure wave.

So, the length of pressure wave is u into t d which is 398.17 into 0.05 which comes to 19.909
meters peak dynamic wind pressure; q naught is given by 0.0032 times of p so square 0032
into 40 square which is 5.12 kilo pascals.

1034
(Refer Slide Time: 48:56)

Now, let us do the front wall loading. Let us calculate the reflected over pressure which is
given by C r is equal to 2 + 0.0073 of p so which is 2 + 0073 of 40 which comes to 2 point
292 and reflected over pressure is C r times of p so which is 2.292 into 40 which is 91.68 kilo
pascal. Let us calculate the clearing distance.

This should be the least of B H and B W by 2 that is least of 5 meters comma 15 by 2 sorry.
So, B, B W is 30, is it not? That is 15 meters. So, therefore, S is 5 meters the minimum of this
two.

(Refer Slide Time: 50:46)

1035
Therefore, reflected over pressure clearing time t c is given by 3S by u which is 3 into 5 by
398.172 which comes to be 0.0376S which is less than t d which is seconds.

It has got to be less than t d that is a condition, is it not? Take the drag coefficient C d as
unity. So, now, let us calculate the stagnation pressure. The stagnation pressure p s is p so + C
d of q naught which is 40 + 1 into 5.12. That is not q naught we calculated, is it not? That is
the q naught we have right. Yeah, ok 5.12 which comes to be 45.12 kilo pascal.

(Refer Slide Time: 52:41)

Front wall impulse, Iw will be ½ (pr - ps)* tc + ½ of ps * td. Let us substitute them ½ of (91.68
- 45.12) of 0.0376 + ½ of 45.12 * 0.05 which gives me 2 kilo pascal seconds. So, if I try to
plot this, this is time in seconds and this is p in kilo pascal.

So, I get this as 45.12 for 0.043 seconds that is the effective duration; t e is given by 2 a I w
by p r, 2 into 2 by 91.68 which comes to 0.0436 seconds which is here.

1036
(Refer Slide Time: 54:19)

Let us do the sidewall loading. The drag coefficient for the sidewall is - 0.4. The equivalent
load coefficient is going to be Lw / L1 which is 20 / 0.3 which is 66.667.

So, equivalent over pressure is given by Ce * Pso + Cd*qo which is 1 * 40 + (-0.4* 5.12)
which comes to 37.952 kilo pascal. The rise time is calculated as L 1 by u which is 0.3 by
398.172 which comes to 0.0007 seconds, we can plot this.

So, t and p this value is 0.05 and this is 37.952. The total duration td will be now 0.05
seconds.

1037
(Refer Slide Time: 56:13)

(Refer Slide Time: 56:34)

Let us calculate the roof load. So, we will continue this problem in the next lecture. So, we
want to make this understand for all the components. So, let us do this. So, I will not do this
here we look at the summary. So, friends in this lecture, we learnt how to estimate blast loads
for different components of the building.

We have learnt how this TNO Green book is helpful to estimate these (Refer Slide Time:
56:59) and we have learnt diagrammatically how the reflected side over pressure creates

1038
additional loading on the surface of the building and as well as in the rear side of the
building. In the whole derivation, we assume that the blast wave is travelling horizontally.

Thank you very much and have a good day.

1039
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 62
Fire - resistant design - 1

(Refer Slide Time: 00:19)

Friends, welcome to lecture 62, where we are going to talk about Fire Load Estimate. Before
that, we will complete the numerical problem of blast load estimate on a rectangular building.
We will continue with that.

So, we worked out previously the sidewall load, now let us work out the roof load for the
problem. For design of roof, let us consider a strip of 0.3 meter wide and 2.4 meter long. So,
let us take L1 as 2.4 meter, and let the drag coefficient be 0.4 negative for the roof.

1040
(Refer Slide Time: 01:46)

Equivalent load coefficient will be now equal to Lw/L1 which is 20/2.4 which is 8.333. From
the UFC chart, one can calculate Ce as 0.9. Therefore, the equivalent over pressure is given
by pa = Ce pso + Cd qo which is (0.9 x 40) +(-0.4 x 5.12), that is what we had as qo, which comes
to be 33.952 kpa. The rise time for this, tr will be L1/u which is 2.4 by 398.172 which comes
to 0.006 seconds. So, let us plot this. Further time of duration td is 0.05.

(Refer Slide Time: 03:38)

1041
So, the total positive phase duration will be equal to tr plus td which is 0.006 plus 0.05, 0.056
seconds. Let us plot this. So, this is 0.006 seconds, is the positive rise phase. This is 0.056
seconds.

(Refer Slide Time: 04:24)

Let us look at the rear wall loading. Let us consider the drag coefficient Cd as -0.4. So, the
equivalent load coefficient will be actually given by Lw/s which is 20 / 5 which is 4 meters.
And from the UFC chart, one can find Ce as 0.88. So, the equivalent overload pressure is
given by pa = Ce pso + Cd qo which will be (0.88 x 40) + (- 0.4 x 5.12) which will be 33.152
kilopascal.

1042
(Refer Slide Time: 05:47)

Time of arrival is ta which is BL/u which is 20/398.172 which is 0.0502 seconds. Now, the
rise time of the positive phase tr is given by S/u which is 5/398.172, which is 0.0125 seconds.
Duration, t d is 0.05 seconds.

Therefore, the total positive phase duration will equal to t naught which is tr plus td which is
0.0125 plus 0.05 which is 0.0625 seconds, which can be plotted as t and pa 0.0125, this is
0.0625, and this value is 33.125 kilopascal. So, friends, this example illustrated how to
compute the blast loads on different segments of a rectangular building taken for the study
using UFC code.

Now, let us discuss about estimate of fire loads on buildings.

1043
(Refer Slide Time: 08:01)

(Refer Slide Time: 08:11)

So, friends, now we will talk about fire loads. These are some interesting references which
you would like to see back when you use fire load computation on structures. Design Aids of
Offshore Structures Including Fire Resistance, this book covers concepts of fire-resistant
design as well as the special loads. Please have a look at these textbooks, which are
recommended for this course.

In addition, if you want to know about the properties of functionally graded material
compared to steel, this is a very nice reference book which is recommended for this course.

1044
(Refer Slide Time: 09:03)

Now, let us say steel is the most favorite material for construction. The design procedure for
steel structures is well laid and well established. For conventional loads such as live load,
dead load, wind load, wave load, current and earthquake loads, we have recommendations
given by the codes. But when you talk about fire load, blast load, impact load, etcetera, they
need a special attention, and the analysis methods are different for them.

For example, there are varieties of Indian codes which discuss about the design procedures.
For example, let us say, IS 456 this is for concrete structures, IS 800 for steel structures, IS
1893 for earthquake loads. But there are no explicit design procedures available and
recommended by the codes for fire loads.

1045
(Refer Slide Time: 11:21)

So, let us see what are the basics of fire load. So, we would like to learn, what is a fire load?
How fire can occur in buildings? What are consequences of fire load? And how material
behaves under fire? Because there are no special design and repair procedures of structures
under post fire accidents. And the steel what we generally use for construction is a high
carbon steel. So, there are issues related to these two concerns, so we need to address that.

But, one thing we need to realize is that, if fire loads occur, causes catastrophic damage, very
importantly no post-fire repair is possible, only rebuilding is done, no repair, it is amounting
to a complete renovation because fire causes a total devastation. Very importantly no
retrofitting is possible for structures subjected to severe fire loads.

1046
(Refer Slide Time: 13:49)

Having said this, let us ask a question which structures are exposed to fire loads? Do we have
to design every building for a fire load? The answer is no. Heavy industrial structures,
process industries, nuclear reactor power plants, offshore structures, chemical process plants,
can cause fire. We can also include electric substations, because electric fire is an important
aspect for design.

(Refer Slide Time: 14:58)

Now, let us talk about what do you understand by fire resistant design. Structures can be only
designed to withstand fire for the specific duration. This is called fire rating. So, please

1047
understand friends, there is no fire proof building, there is no fire-resistant building, but there
are fire safety buildings for a specific duration only.

Then, in that case, what is important in fire resistant design? Structural design of members for
fire load is only secondary. The primary is public safety. The primary is public safety.
Structural design of members for fire is necessary, but that is only a secondary case.

(Refer Slide Time: 17:03)

Now, let us see what are the factors that govern the geometric design of fire-resistant
buildings. The first one is the firefighting system to access the building. That is the building
should have enough circulation space around the building. The building should have
something called a common assembly point, which is also otherwise called as meeting point
which should be known to all the occupants of the building.

The building should have a very good evacuation facility. It should follow public safety under
fire accidents. It should have well-lit escape routes; it should follow fire evacuation procedure
as instructed by the course. Moreover, and most importantly, all critical items such as fire
hydrants, control rooms, fire water tank, should be in proper working condition. It should be
frequently inspected and certified for its proper working.

1048
(Refer Slide Time: 19:32)

Let us quickly see some basics of fire. Friends, fire is a rapid exothermal oxidation of ignition
fuel. The fuel can be in the form of solid, liquid or gas. Fire releases energy which is an
exothermal reaction. Now, within short time, the released energy reaches its peak amplitude
or peak intensity. That is the most serious concern as far as fire is concerned.

(Refer Slide Time: 21:11)

Fire can also result from explosion. Explosion is rapid expansion of gases caused by pressure
waves or shockwaves. In the previous set of lectures, we understood how to calculate the

1049
shockwave and overpressure waves. A rapid expansion of gases caused by these waves can
result in explosion.

The interesting part is these waves propagate very fast. In the numerical example you see that
they were only in milliseconds. What is the concern of this? It can cause adiabatic expansion.

Explosion resulting from fire can be a mechanical process or a chemical process.

(Refer Slide Time: 23:13)

Let us see what are categories of fire as per DNV code. So, DNV code classifies fire as
follows. No fire risk, so the building in that case has no energy resources to catch fire. The
example could be home or house.

The next is low fire risk, this will occur in buildings which stores non-flammable materials.
Examples of academic blocks, schools, colleges, apartments, which stores non-flammable
materials.

Next could be medium fire risk, which is classified as buildings that has or that have electric
and major testing equipment. For example, it can be workshop. Lastly, high fire risk, these
can be buildings that stores flammable liquid. Example could be process industries, offshore
platforms, as well as it also classifies and high-power mechanical equipment. So, in that case,
I can also include electric substation.

1050
(Refer Slide Time: 26:17)

Let us ask a question how can we avoid fire accidents. The fluid used in such industries
which are prone to fire should be above the flash point. It should be above the fire point
temperature. It should not be above the auto ignition temperature.

Next, fluid can be used up to the maximum bulk temperature. So, the maximum bulk
temperature of any fluid is always much higher than the flash point temperature. It is a
common characteristic of a fluid. Finally, no confined space near the ignition source, it
should be open from all sides.

(Refer Slide Time: 28:31)

1051
Further, one should ensure good oxygen supply or the space should be well-ventilated. But
importantly no oxygen supply near the heat source, electric heater, the heat exchanger, the
drilling controller. Further, fluid should be contained within the system. There should be no
direct contact with the external ignition sources.

(Refer Slide Time: 30:18)

So, let us see some of the important points for fire resistant design.

​ Avoid congested layout of workflow.

​ Design the building with good ventilation.
​ There should be no congestion of pipelines, electric cables, or water mains in any
floor.

1052
(Refer Slide Time: 31:37)

Now, let us see what are important characteristics of fire.

First one, Auto ignition temperature, AIT. It is also called as kindling point of materials. It is
the lowest temperature above which the material may not require any external source for
combustion. So, ignition takes place at normal atmospheric conditions when AIT of a fluid is
reached.

(Refer Slide Time: 33:29)

1053
AIT is called also called as self-ignition temperature. It is also the minimum temperature
require to supply activation energy required for combustion. Let us say for example, diesel
has got 200 degrees Celsius, methane is 580 degrees Celsius, gasoline is in the range of 247
to 280, butane is about 405 degree Celsius.

Let us also look at something called a flashpoint. This is the lowest temperature at which the
liquid gives enough vapour to maintain a continuous flame. It is not the temperature at which
vapours are produced, for easy ignition. It is the temperature at which vapours are produced
for easy ignition.

So, friends, an important point here is at flashpoint, fluid will not burn. That is important.

(Refer Slide Time: 36:49)

The next one is called as a fire point. It is a temperature at which the fluid will sustain fire, if
ignited by an external source. But interestingly, fire point of a fluid is always much lesser
than the auto ignition temperature. Since, we are writing about the fire-resistant design, let us
talk quickly about fire protection systems as well for our understanding.

1054
(Refer Slide Time: 37:56)

There are many fire protection systems available. One is called as the frame arrestor. Flame
arrestor, it is a passive device which prevents propagation of fire through pipelines.

The second one is called fire barriers, fire insulators, water shower, foams, water spray lines
etc. The third set of fire protection system involves spark preventers. It helps to prevent
spark. This is one of the highest priorities in firefighting protection systems. Of course, the
last one as we all know is fire extinguishers which are firefighting systems.

(Refer Slide Time: 40:08)

1055
And let us have a brief look at the components of firefighting system. Firefighting system
should be simple in design. It should be easy to operate. It should be free from high tech
components. It should be maintained periodically. Most importantly, should be available on
demand.

Friends, the recent research is now focusing on reliability and maintainability. So, you are
looking for reliability and maintainability to check whether any specific system is reliable
and available on demand. So, to ensure this, people conduct fire drills at periodic intervals in
all industrial structural systems to ensure that the firefighting system is essentially works
maintained in a proper manner.

(Refer Slide Time: 42:13)

Now, friends as we understand, steel is seen as a very advantageous material. The curve on
the screen shows the temperature is a reduction factor which is a strength reduction of steel.
So, steel is the most preferred material, but steel has got properties which are dependent on
temperature.

Look at the effective yield strength and proportional limit as well as the slope of linear elastic
range. So, the yield strength, the stiffness, and modulus of elasticity, all decrease with
increase in temperature. Especially, the yield strength after 400-degree Celsius decreases
significantly.

1056
The proportionality limit also decreases beyond 100 degree Celsius. So, the ductility
characteristics, strength degradation, happens in steel as you proceed with the temperature.

(Refer Slide Time: 44:04)

Furthermore, if you look at thermal conductivity of steel, it also decreases with increase in
temperature beyond 800 degrees Celsius. It remains constant. Furthermore, the specific heat
capacity of steel remains almost constant except at about 700 degrees Celsius, it reaches a
specific peak where the specific heat capacity touches about 5000 joules per kg kelvin close
to range of about 750 degrees Celsius in the temperature.

So, these properties of steel which were discussed in the earlier lectures are only helping you
to re-brush the constant characteristics of steel with respect to time and temperature.

1057
(Refer Slide Time: 45:13)

Furthermore, if we look at the thermal strain, it keeps on increasing with increase in

temperature. So, therefore, friends, thermal conductivity, specific heat, elastic constants,
specific weight, thermal expansion, and plastic properties of steel all become function of
temperature.

So, when you are using steel for a fire-resistant design, where the fire is going to happen at a
higher temperature because of explosion or hydrocarbon fire et. You got to be very careful in
understanding the properties of steel which is being used for fire resistant design.

(Refer Slide Time: 46:28)

1058
So, friends, in this lecture, we learnt about characteristics of fire. We learned briefly about
tips for fire resistant design. We also completed a problem on blast loads analysis. And we
learned that steel has got properties which are function of temperature. And one has got to be
very careful, in using these properties appropriately in the design, if you are enabling fire
resistant design of steel structures.

Thank you very much. And have a good day.

1059
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 63
Fire - resistant design - 2

(Refer Slide Time: 00:19)

Friends, welcome to lecture 63 of Advanced Steel Design, where we are continuing to discuss
the procedures for Fire resistant design as a part of this course. In the last lecture, we
discussed about some essential characteristics of fire, what are the tips for good fire-resistant
design buildings, and so on. Now, we will focus more on design methods for fire resistance.
This is what specific reference to API-RP 2A, American Petroleum Institute Recommended
Practice 2A.

1060
(Refer Slide Time: 00:56)

This specific course says there are 3 possible ways by which you can design fire resistance
structures. 1 is by what is called as a zone method, 2, linear-elastic method and the 3rd one is
elastic-plastic method.

(Refer Slide Time: 01:57)

In the zone method, the maximum allowable temperature is assigned to the steel member.
This method has no reference to the stress level in the member before fire.

1061
Let us say, in a given building, there are zones where there is a possibility of fire, let us say
this member, and let us say in this member, column and this beam. These are the possible
locations where a fire could outburst.

So, now these steel members are assigned to a maximum allowable temperature up to which
it has to sustain. When you do that, you do not have any cross reference to the stress level in
these members before fire. So, the design is more focused on member utilization ratio. The
basic allowable stress will remain unchanged even under fire.

That is the reduction in stress level due to high temperature is not accounted in the design as
the allowable value. They remain unchanged. So, this method is focusing only on certain
areas, on certain members which has got high proximity of fire, probability of fire occurrence
is higher. So, that is why this method is called zone method.

(Refer Slide Time: 05:13)

In this method, the yield strength σ𝑦 is reduced by a factor of 0.6. Now, there are

disadvantages of this method. 1, it is not applicable with the unmatched reduction in both σ𝑦

and E, because you have a constant number. So, this affects the design criteria.

1062
(Refer Slide Time: 06:26)

The 2nd method is linear elastic method. In this method, the maximum allowable stress is
assigned to the steel member based on the stress level before fire. So, here the main focus is
to maintain the member utilization ratio close to unity. So, effective utilization is ensured and
this is done even at increase in the temperature.

In this design method, the stress level corresponds to the assigned temperature level. So, you
have to pick up the yield value with respect to the corresponding temperature level of steel
during fire. So, when you talk about the temperature variation during fire, then you may ask
me a question what temperature should I consider for estimating σ𝑦 from the material

property? That is a good question.

You can have a maximum temperature; you can have a minimum temperature. So, what we
generally consider is, your core average temperature is considered and the equivalent σ𝑦 at

that temperature is taken for design.

1063
(Refer Slide Time: 09:04)

This method has some disadvantages. This method also proposes an unmatched value of σ𝑦

and E, which is a serious concern because the degradation in young’s modulus and the yield
strength are not same at a specific temperature. It considers an average core temperature
value for which σ𝑦 is considered, but E is not accounted with respect to the temperature

variation.

(Refer Slide Time: 10:02)

1064
The 3rd method is elastic plastic method, where the maximum allowable temperature is
assigned to the member, based on the stress level before fire. Friends, this is not assigned to
all the members. It is only assigned to critical members. What do you mean by critical
members? The members whose stress level is maximum before fire.

As a result, the member utilization in this design procedure is lesser than 1.0 with the increase
in temperature.

So, all members are not subjected to uniform stress range, only those critical members
subjected to higher temperature during fire is assigned a specific stress value. So, therefore,
one need to verify the structure under collapse. So, you need to verify the structure under
collapse loads, but the design should ensure serviceability under fire loads. So, that is the
design criteria what this method ensures.

(Refer Slide Time: 12:30)

This method also has couple of disadvantages. Commonly used strain value in the design is
0.2 which matches with the reduction in σ𝑦 with increase in temperature. But selection of

higher strain, will result in higher temperature. See, you are looking for an average
temperature for which the strain value may not be 0.2. In the last lecture, we discussed about
the strain variation in steel with respect to temperature.

So, either you pick up the strain value for a specific temperature or for a specific strain pick
up the temperature. So, the procedure says you pick up the stress value for a specific

1065
temperature for the critical members, but when it assigned the strain value it is 0.2. So, a
higher strain value assigned to the members will result in higher temperature, which will not
match with the stress and young’s modulus for the design. So, that is a concern in this
method.

(Refer Slide Time: 14:38)

Having said this let us talk about energy released during fire. If the ignition fluid is present in
the building, then the total energy released during fire can be given by what is called as fire
load. So, fire load is the total energy which is sum of 𝑀𝑖𝐶𝑖, where M i is the mass of

combustible materials present in the building in kgs.

𝑛
𝐸 = ∑ 𝑀𝑖𝐶𝑖
𝑖=1

𝐶𝑖 is the calorific value of each material which is either in joules per kg or of course,

megajoules per kg. For common materials being used in buildings, national building core part
4 gives the value of 𝑀𝑖 and 𝐶𝑖 value of different materials.

1066
(Refer Slide Time: 16:44)

Look at this table which is extracted from the national building core for different material.
Let us say solids, hydrocarbons, common solids, even foodstuffs.

A calorie value and the wood equivalence are given in the core. Let us do a small example.
Let us say there is a presence of 200 kg of benzene hydrocarbon. If you look at benzene
hydrocarbon, the calorific value is 39.6, I can say megajoules per kg; plus 10 power 3 I have
taken as megajoules.

Now, the average release rate which is called as HRR can last for 1 hour. So, we can say E is
going to be 200 into 39.6 that is 7920 megajoules. So, the energy released Q can be 79; 7920
by 60 into 60, which is 2200 kilowatts.

1067
(Refer Slide Time: 18:50)

Let us do one more example of a mixed mass. You can see on the screen the NBC table
extracted. Let us say the building contains 100 kg of benzene whose C value is 39.6, 50 kg of
butane whose C value is 47.1, 200 kg of ethane C value is 49.1.

So, the average energy released during 1 hour is called as HRR, heat release rate that is 60
into 60 seconds is given by sum of 𝑀𝑖𝐶𝑖. In our case going to be,

𝑛
𝐸 = ∑ 𝑀𝑖𝐶𝑖 = (100×39. 6) + (50×47. 1) + (200×49. 1) = 16135 𝑀𝐽
𝑖=1

3
16135×10
𝑄 = 60×10
= 4481. 9𝑘𝑊

1068
(Refer Slide Time: 20:55)

Let us look into the t square fire model which is a very common time temperature curve,
which is being used is also called as T-t model or a t square fire model. This can be used to
quantify the energy released.

Furthermore friends, you can also see the time temperature or time heat release rate curve
where, the α g t square which is the heat release rate depends on a factor which is there for
different kind of fire. Let us say, if it is a slow fire growth α g is different compared to
ultra-fast fire growth.

Now, if the fire growth rate is known to us, we can estimate the time of burning. We can
easily find out how long the fire will be sustained. Furthermore, there are different phases of
fire growth initiation, growth phase, decay phase and end of combustion. So, in the growth
phase, there is a significant rise in the heat release rate with reaches its peak and then it gets
decay as you see here in the figure.

So, now let us say for example, I have 100 kg of fuel oil. The fuel oil has got C as 41.6
megajoule per kg from the table. Let us assume an ultra-fast fire to take place. The peak
HRR, that heat release rate is 9 megawatts. We want to estimate the time of burning. Let us
do that.

1069
(Refer Slide Time: 24:27)

𝐸 = (100×41. 6) = 4160 𝑀𝐽

The growth factor for fire or ultra fast fire that is α g from this table, you can see, it is 0.1874,
0.1874 kilowatt per second square. Therefore, the time taken to reach the peak HRR that is Q
p is 9000 because we already said the HRR is about 9 megawatts divided by 0.1874 root
which will be 219 seconds.

𝑘𝑁
α𝑔 = 0. 1874 2
𝑠

9000
𝑡= 0.1874
= 219 𝑠𝑒𝑐

Now, the total energy released in the growth stage let us call this E 1 can be computed as

𝑄𝑝𝑡𝑝 9×219
𝐸1 = 3
= 3
= 657 𝑀𝐽 < 4160 𝑀𝐽

Since, the 657 megajoule in the growth stage is much lesser than the heat release which is the
energy present because of the oil or because of the fuel oil. We can say the total burning time
will be

𝑡𝑏 = 𝑡𝑝 + ( 4160−657
θ𝑝 ) = 10. 1 𝑚𝑖𝑛

1070
(Refer Slide Time: 27:20)

Let us discuss about the time temperature behavior of fire. Friends, there is a significant
difference between the real fire and the standard fire. Fire has a starting phase which is
initially small about 20 minutes. So, it is during this time, one can activate the firefighting
equipment’s like fire alarms, sprinklers, extinguishers, call to fire station, evacuation plan
etcetera. All can be activated during this time. It is always good to provide a compartment
design. So, the fire does not spread to the places where device and equipment’s are kept. So,
friends, we must understand that firefighting is initiated during the fire growth phase during
the fire growth phase.

(Refer Slide Time: 30:00)

1071
(Refer Slide Time: 30:05)

Following by this, let us have something called pre-flash over stage. This is true up to a
temperature less than 600 degree Celsius. The post-flash over stage is called as the burning
stage. It is interesting friends for you to know that hydrocarbon fires which happen in
offshore platforms, can reach the temperature as high as 1200 degree Celsius.

Therefore friends, in such cases fire reaches its peak temperature approximately within 90
minutes and then starts decaying. Usually, the decay period of fire is about 4 to 6 hours. But it
contains and depends on the unburned inventory. Let us look at this curve which is the
pre-flash over and post-flash over stage. The pre-flash over stage is about approximately 20

1072
minutes on initial ignition and smoldering. Then, the burning takes for 90 minutes, then the
decay is about 240 minutes. The natural fire curve is what you see here.

Whereas, the standard fire curve always starts after the pre-flash over stage, after this stage.
So, the philosophy for design is that the structural system should not collapse even at high
intensity of fire. There are various codes of practices which ensure this fire safety which
shows the standard fire curve. For example, IS 3809, ISO834, ASTM E119, helps you by
providing the standard fire curve.

(Refer Slide Time: 33:37)

(Refer Slide Time: 33:54)

1073
Furthermore, we can also estimate the fire load energy density which is called as FLED. The
fire load energy density depends on different code provisions which are also carried and
revised by different countries. For example, in London, the mean FLED energy density is
about 330 megajoules per square meter, whereas in India it is about 348. So, it keeps on
varying and these effects are reflected in the respective national standards of the countries.

A recent study carried out by IIT, Gandhinagar in 2015 shows that the mean value of fire load
energy density is much higher in reality. This was tested in the laboratory and people have
reported that the FLED value for design which is referred as different codal provisions are far
lesser than the actual value. Now, for the standard fire curve T is given by T 0 plus 345 log of
8 t m plus 1, for standard fire. We call this equation number 2.

(Refer Slide Time: 35:54)

Now, look at this fire curve back again it says that the natural fire curve has got a cooling
phase. The standard fire curve closes here, whereas the parametric fire curve, which indicate
in Euro code which is EN 1992-1-2 indicates a cooling phase for about short period.

So, now, there is a cooling phase which is shown in the red color here. So, this is for the
natural fire which also indicates a cooling phase whereas, in standard fire curve this cooling
phase is missing. Whereas, if you look at the Swedish fire code in Annexa A, FLED is
revised based on the room shape and size, and openings present in the room, and inventory
materials present in the room.

1074
So, there has been a revision in the Swedish fire code to compute the fire load energy density,
not at the standard value, but is dependent on realistic parameters as you see here. The
standard fire curve only indicates the growth phase, but does not talk about the cooling phase.
Whereas, a parametric fire curve used in euro code talks about the cooling phase as well.

(Refer Slide Time: 38:05)

So, therefore, friends in fire resistance design, the 1st step is to compute fire load energy
density, which is given by

𝑞𝑓,𝑑 = 𝑞𝑓,𝑘(𝑚)δ𝑞1δ𝑞2δ𝑞𝑛

Call this equation number 3 where m is the combustion factor, which is approximately 0.8.
δ𝑞1 is a fire activation risk which depends on the shape and size.

δ𝑞2 depends on the type of occupancy. And del n is a combined effect of the above factors.

Now, 𝑞𝑓,𝑑 is called the design FLED, 𝑞𝑓,𝑘 is called the characteristic fire load density per area

of flow which will be expressed in megajoules per square meter.

1075
(Refer Slide Time: 40:25)

For example, in national building code of Indian practice, it is said as 25 kg per square meter
of wood, equivalent of wood. So, we now say, wood equivalent we can say, which is 25 into;
the calorific value of wood which is giving me 425 megajoules per square meter.

(Refer Slide Time: 41:10)

Now, the other factors δ𝑞1 and δ𝑞2 are given here depending upon the type of occupancy and

the size and shape. If the compartment area is about 25 square meters δ𝑞1 is about 1.1 and for

art gallery museum etcetera is 0.78, whereas manufacturing fireworks is about 1.66. So, the

1076
example of occupancy is also listed here which helps me to compute the δ𝑞1 and δ𝑞2 factors

required to compute the FLED of the design value.

(Refer Slide Time: 42:04)

If you have firefighting systems which is useful for suppressing fire with water extinguisher
systems, independent water supply system, fire brigade located off site, or firefighting
devices available, then the del n value is also varied. So, it is a combined effect which is
useful to compute the FLED for the design.

(Refer Slide Time: 42:38)

1077
The table what you see on the screen is an average value of 𝑞𝑓,𝑘 in megajoules per square

meter is 𝑞𝑓,𝑘. So, for example, 𝑞𝑓,𝑘 for dwelling units is about 948 megajoules per square

meter which is equivalent value of wood. So, it is available for different kinds of public
spaces as well as private spaces which is used for the design, which is the characteristic
value. Now, with the help of this one can compute the FLED.

(Refer Slide Time: 43:29)

Now, the second step involved in fires resistant design or the fire load estimate is heat phase
calculation.

−0.2𝑡* −1.7𝑡* −19𝑡*

[
θ𝑔 = 20 + 1325 1 − 0. 324𝑒 − 0. 204𝑒 − 0. 472𝑒 ]
(Refer Time: 44:17) equation number 5.

Where θ𝑔 is expresses the gas temperature in degree Celsius which is based upon the time

delay. t star is expressed in hours. t star depends on opening factors that is the ventilation then
area of the vertical opening in walls. Total area of the compartment for which fires resistant
design is attempted.

1078
(Refer Slide Time: 45:23)

*
𝑡 = 𝑡. 𝑟

where t is the actual time in hours and gamma is opening factor by breadth b square. Let me
write down here. Gamma is opening area by breadth square which is 0.04 by 1160 square. Let
us not.

(Refer Slide Time: 46:27)

𝑂 2
γ= ( )
𝑏

1079
So, gamma is based on the ratio of opening to the whole area. This is referred as other factors
that influence the design. Now, gamma is O by b square, where O is called as the opening
factor which is

𝐴𝑣 ℎ𝑟𝑒𝑞
𝑂 = 𝐴𝑡

0. 02 < 𝑂 < 0. 2

where A v is ventilated opening, A t is the total area of enclosure, and O lies between 0.02 to
0.2. h equivalent is the average height of windows in the openings.

(Refer Slide Time: 48:11)

In this expression, b depends on the material property;

𝑏 = ρ𝑐λ

for 100 and 2200 joules as square meter per root s k where ρ, c, λ are density, specific heat,
and thermal conductivity of the material. ρ, c, λ product is called thermal inertia of the
compartment. Gamma, if it is unity, which is referred to which will refer to the standard time
temperature curve. We have got seen two curves, right, standard and actual the equation 8.

1080
(Refer Slide Time: 49:47)

Opening factor also depends on or also governs the fire behavior. In general, all fire are
ventilation controlled. Few of them could be fuel controlled. For example, in offshore
platforms, you have got lot of ventilation around. So, there is no problem of ventilation, but
inventory of fuel will control the fire.

* −3 𝑞𝑡,α
𝑡𝑚𝑎𝑥 = 𝑀𝑎𝑥(0. 2×10 𝑂
; 𝑡𝑙𝑖𝑚

Fire intensity depends on the fuel. where O is the opening area and t limit depends on
whether it is fuel controlled or ventilation controlled - equation 9.

1081
(Refer Slide Time: 51:48)

𝐴𝑓
𝑞𝑡,𝑑 = 𝑞𝑓,𝑑 𝐴𝑡

Now, for 50 less than or equal to 𝑞𝑡,𝑑 less than or equal to 1000 megajoules as square meter,

where 𝑞𝑓,𝑑 is called the design value for fire load density, which is computed based on the

total area of ventilation which is 𝐴𝑡.

(Refer Slide Time: 52:45)

1082
The 3rd step is the cooling phase. Now,

(*
)
* *
θ𝑔 = θ𝑚𝑎𝑥 − 625 𝑡 − 𝑡𝑚𝑎𝑥, 𝑥 𝑓𝑜𝑟 𝑡𝑚𝑎𝑥≤0. 5 ℎ𝑟𝑠

(
θ𝑔 = θ𝑚𝑎𝑥 − 250 3 − 𝑡𝑚𝑎𝑥
*
)(𝑡 − 𝑡 , 𝑥)𝑓𝑜𝑟 0. 5 ℎ𝑟𝑠 ≤ 𝑡
* *
𝑚𝑎𝑥
*
𝑚𝑎𝑥
≤2ℎ𝑟𝑠

(Refer Slide Time: 54:21)

So, friends, in this lecture, we learnt about the heat release rate of fire load. We also learnt
about the standard fire curve and the cooling phase of the fire curve. We understood the time,
temperature, behavior of fire, and understood how to compute the fire load for design.

So, more details of this fire-resistant design part can be seen in a separate course of NPTEL
conducted by IIT, Madras by me on Offshore Structures under Special Loads including Fire
Resistant Design. So, design examples are available. And I also authored a book on fire
resistant design which can be useful to you for working out more examples on this. Please
refer to them. And try to also refer the DNB code and the European code for learning more
about the parametric fire curves.

Thank you very much. And have a good day. Bye.

1083
 Advanced Design of Steel Structures
Prof. Srinivasan Chandrasekaran
Department of Ocean Engineering
Indian Institute of Technology, Madras

Lecture - 64
Analysis under impact loads

(Refer Slide Time: 00:21)

Friends, welcome to lecture 64 of Advanced Steel Design course. In this lecture, we are going
to learn a design a example of Analysis of an offshore topside under impact loads; so,
thorough design example where the excerpts of the design is being projected and discussed
for you. So, in this example we are going to discuss the analysis comparison between steel
and functionally graded materials (FGM).

1084
(Refer Slide Time: 01:28)

Now, we all know that impact loads happen on offshore structures or buildings due to fall of
objects. So, they affect the strength of the member. They also challenge the structural
integrity. Now, impact loads can occur during construction or operations. Falling of objects
from a considerable height may damage the structural components. It can also cause
excessive deformation to the deck. So, the impact loads generally arise from fall of crane
hooks itself, fall of objects dropped from a height, machinery components, equipment’s and
tools.

(Refer Slide Time: 03:35)

1085
Let us talk a classical example of an offshore platform where the top side is being shown on
the figure. So, this is a top side element. On this top side, the vulnerable region for impact
loads is being marked. This is a vulnerable region. Let us have a model of this type.

(Refer Slide Time: 04:09)

So, friends, severe impact loads can be caused on decks of the platform. It can damage the
stiffness; the secondary beams and it cannot also result collapse of the deck due to excessive
deformation. Therefore, it is necessary to investigate the strength capabilities of the topside
elements to support this kind of impact loads.

(Refer Slide Time: 05:12)

1086
Interestingly friends, we look at the accident statistics of offshore units developed by Britain
for the past 20 year’s period. You can see the maximum accidents, because of drop objects
have also caused fatalities and there are increased number of accidents because of crane
hooks and falling objects compared to explosion, collision and fire.

See this table and this has been a consistent mark over the years. So, drop objects impact
loads due to cranes are more frequent, which create accident statistics in at least offshore
platforms as per the table shown on the screen.

(Refer Slide Time: 06:20)

Now, the varieties of cranes being used in offshore structures are shown on the screen. So, the
fall of crane hooks of these heavy mass and other objects is frequent and fall of crane hooks
can cause potential damage to structure and people on board. Look at the typical crane hooks
being shown on the screen. They are very massive and they vary from 2 to about 32 tons that
is the range of weight these crane hooks have.

1087
(Refer Slide Time: 07:24)

Let us pick up a specific crane hook of that order and make a model. Let us say the crane
operational height ranges from 15 to 25 meters. So, let us take a 20-meter drop of the crane
hook falling freely with a high velocity. Typical time history of an impact force is what you
see on the screen here which has got a very shorter duration, you can see it is only about 0.07
seconds. But, a very large amplitude, very large amplitude about 50 tons etcetera. So, let us
say this kind of a model is being applied for this.

(Refer Slide Time: 08:22)

1088
When you talk about materials, which is being used for the top side in the design, material
selection plays a very crucial role particularly under impact load. During high velocity
impact, the rate of strain increases rapidly over a short duration of time. This can cause a
permanent deformation of member. So, the material chosen should be strong enough, should
have higher toughness, higher resistance to impact loads should possess high weld ability and
durability.

(Refer Slide Time: 09:55)

For comparison, we have taken three materials. We have taken steel which is X-52 steel
which is the structural steel commonly used for offshore topside is also the marine grade
steel. They have also tested with composite and FGM for comparing the performance of this
topside deck. When you talk about composites and their use in offshore structures, composite
plates are actually fabricated by hydraulic press.

So, they have a very serious problem of delamination under high pressure and temperature.
So, new materials have been tested, we have also examined FGM. In earlier lectures of this
particular course, you would know the combinations of functioning graded material and their
characteristics, which has been discussed very well, in the earlier part of the lectures. Please
refer them back and we have considered a specific FGM for this application for the time
being.

1089
(Refer Slide Time: 11:32)

So, for our convenience let us quickly recollect that FGM offers improved residual stress
distribution. It has enhanced mechanical properties. It has got higher resistance to fracture
and it has got reduced stress intensity factors. So, we are not discussing about the FGM
properties at this moment.

(Refer Slide Time: 12:22)

We will start directly going to the model of the crane hook. The model of the crane hook is
showing on the screen whose dimensions are given here. The velocity of the impactor

1090
considered in the analysis is calculated for 20-meter fall and the weight of the hook is taken
as 2.85 tons which is subjected to the CG of this hook.

(Refer Slide Time: 13:14)

The analyst carried out in abacus and the boundary conditions for the deck and crane hook
model are explained on the figure shown on the screen. The crane hook can have a rebound
effect only on one direction as you see here, remaining all are restrained.

(Refer Slide Time: 13:40)

1091
So, the deformation contour for all the three materials, I have been shown here. FGM has the
stress concentration only on a smaller area compared to X-52 and aluminum. Of course, this
is for t 0.02 seconds, t 0.04 seconds and this is for 0.06 seconds; over the rate of time increase
of course, the impact spreads. But, even at a higher period, you will see that the stress
concentration is far lesser in terms of its area of spread compared to that of X-52 and
aluminum under this impact load.

(Refer Slide Time: 14:35)

Now, the response of the deck is also plotted for contact force history, deflection time history,
central deflection and comparison of plastic strain time history, kinetic energy time history
and impactor displacement. In all the cases, you would very well see that FGM material
performs better compared to its competitor X-52.

If you look at the central deflection of the deck which is made of FGM, it is showing far
lesser values compared to the other two material taken for the study. Even the deflection time
history is lesser for FGM whereas, the contact force time history is much higher for FGM.
So, it has got a very good capacity of plastic strain absorption as you see here.

1092
(Refer Slide Time: 15:52)

Having said this, we can now summarize that FGM possesses reduced plastic strain. It also
has increased kinetic energy compared to X-52 steel. However, if you look at aluminum, it
experiences lesser kinetic energy, but undergoes excessive deformation. Hence, cannot be
advised for off shore deck. X-52 steel anyway performs better followed by FGM. So, FGM is
far better than all the three compared materials.

So, FGM has improved impact strength, reduced plastic strain, reduced deflection, for the
same thickness of the deck plate that is very important. So, there is no change in the geometry
dimension of the member, only the material is being changed. So, friends a simple analysis
example applied for impact loads explain how the impact analysis can be carried out using
abacus, by incorporating the material characteristics of FGM and steel which are discussed in
length in the previous lectures.

So, friends there are many references available for this particular course including textbooks
authored by me and other leading academicians of the world. Please go through them, try to
learn and create more interest towards the advanced steel design procedures, which has been
discussed extensively in this course.

Thank you very much and have a good day.

1093
 THIS BOOK
IS NOT FOR
SALE
NOR COMMERCIAL USE

(044) 2257 5905/08

[Link]
[Link]