PROCESS PIPING STRESS ANALYSIS with

CAESAR II
1|Page
HOW TO USE THIS BOOK

The idea behind this book is to discuss the technical details of the CAESAR II stress input and output as
well as get the user acquainted with the modelling and analysis procedure in the software.

The user is advised to follow the modelling procedure we used, especially if this is the user’s first
attempt at using the software.

Some solutions where exaggerated to drive certain technical points.

Some example models were reused albeit with modifications in later examples.

TABLE OF CONTENTS
INTRODUCTION TO CAESAR II 1

PROCESS PIPING MATERIALS SELECTION 6

STRESSES AND STRESS ANALYSIS TYPES 15

PIPING INPUT PARAMETERS 28

2|Page
THERMAL STRESSES 47

SUSTAINED STRESSES 72

FATIGUE 86

PUMP PIPING 98

MODAL ANALYSIS 132

HARMONIC ANALYSIS 152

WATER HAMMER ANALYSIS 162

3|Page

CAESAR II
Due to the expensive, hazardous and annoyingly destructive approach of crash test methods, it was
required to have a reliable method of designing against primary, secondary and peak loads that requires
far less time and money than the crash method.

Process Piping stress analysis is the process of questioning our piping system in the face of prevalent
loads acting on them, in order to keep the system in equilibrium or the stresses resulting from these
loads below the code defined allowable stress so we can achieve an energy-efficient and structurally
satisfying piping system while using the least possible cost method throughout the system’s design life.
From the above definition, the following words are underlining:

a. Loads
b. Equilibrium or below limit
c. Code
d. Energy-efficient
e. Structurally-satisfying
f. Least cost

Throughout this book, we will come across these terms and see what role they play.

Before we proceed further, a process piping system is any above ground system, comprising of
mechanical supports, pipes, in-line items like valves, bends and connected equipment nozzles, governed
by ASME B31.3

We have said that we want a system that is energy-efficient while upholding structural values. This
means not only analyzing the system against different loadings, but the selection of appropriate
materials and components and also with very good measures that will eliminate fouling of the pipe
internal and external surfaces in or out of service with the least possible cost. As the fouling of the

4|Page
internal surface increases, the energy of the fluid drops and as surface fouling (internal or external)
increases, the likelihood of a fatigue failure increases.

Stress analysis now employs the use of computer FEA programs because of the less amount of time it
takes to get accurate and reliable results.

CAESAR II is a Finite Element Analysis (FEA) software from Intergraph (formerly COADE) which makes
stress analysis easy. It is a better and a faster way of knowing the stress state of a piping system
nomatter the size and complexity. It is a computer-based FEA program which has built-in code
requirements and algorithms for special executions with little user input. It is the complete FEA
software.

One of the advantages of the program is in its wide usage, especially in the Oil and Gas and Power
industries. In the hands of an experienced user, it eliminates the chances of error, common with the
hand calculations, reduces design time and introduces cost effectiveness into piping stress resolutions.

The program can be used for both new and old systems. Like every other FEA package, it ensures the
structural integrity of every piping system and troubleshoots old ones.

CAESAR II’s approach to piping stress resolutions is the traditional finite element analysis approach. It
involves the discretization of the system, which results in a system of several finite parts where Hooke’s
law applies to every element except the piping system is non-linear.

The idea is, rather than consider the system consistently, the system is treated like a discontinuous
one defined by at least one boundary. These elements are assumed to be homogeneous and defects-
free.

Unlike in dynamics, in its static analysis, the software allows for non-linear conditions like support
friction, pipe liftoff and gaps. These non-linear conditions help describe systems to be analyzed,
especially in existing systems and forensic engineering (knowing the root cause of failure in failed
systems). The software reports failure if the cumulative stress, called code stress transcends the code
defined allowable stress for the material and process conditions. The allowable stress defined by the
code is to prevent failure which includes plastic deformation (applied stress greater than yield strength
of the material) and consequently fracture.

5|Page
Codes and standards
Codes usually set forth requirements for design, materials, fabrication, erection, test, and inspection of
piping systems, whereas standards contain design and construction rules and requirements for
individual piping components such as elbows, tees, returns, flanges, valves, and other in-line items.
Each code has limits on its jurisdiction, which are precisely defined in the code. Similarly, the scope of
application for each standard is defined in the standard.
Therefore, users must become familiar with limits of application of a code or standard before invoking
their requirements in design and construction documents of a piping system.
Without codes and standards, allowable stresses will not be calculated and piping systems’ stress states
become indeterminate.
Presently, the following sections of ASME B31, Code for Pressure Piping are published:
ASME B31.1 Power Piping
USAS B31.2 Fuel Gas Piping
ASME B31.3 Process Piping
ASME B31.4 Liquid Transportation Systems for Hydrocarbons, Liquid Petroleum Gas, Anhydrous
Ammonia, and Alcohol
ASME B31.5 Refrigeration Piping
ASME B31.8 Gas Transmission and Distribution Piping Systems
ASME B31.9 Building Services Piping
ASME B31.11 Slurry Transportation Piping Systems

ASME B31.3: PROCESS PIPING

This code prescribes requirements for the materials, design, fabrication, assembly, erection,
examination, inspection, and testing of piping within the property limits of facilities engaged in the
processing or handling of chemical, petroleum or related products. The requirements of ASME B31.3
apply to piping for all fluids, including raw, intermediate, and finished chemicals; petroleum products,
gas, steam, air, and water; fluidized solids; and refrigerants.

6|Page
In case of packaged equipment, the interconnecting piping with the exception of refrigeration piping
shall be in compliance with the requirements of ASME B31.3.

Due to the variety of process fluids and conditions within the jurisdiction of this code, ASME B31.3
defines the following fluid services:

Category D Fluid Service

This is a service in which the fluid handled is non-flammable, nontoxic, and not damaging to human
tissue, and the design pressure does not exceed 150 psi (1035 kPa) and the design temperature is
between 20F and 366F (29 and 186C).

Category M Fluid Service

This is a service in which the potential for personnel exposure is judged to be significant and in which a
single exposure to a very small quantity of a toxic fluid, caused by leakage, can produce serious
irreversible harm to persons on breathing or bodily contact, even when prompt restorative measures
are taken.

High-Pressure Fluid Service

This is a service in which the user specifies the application of Chapter IX of B31.3 for piping design and
construction for fluid services with high design pressures. High pressure is considered by the Code to be
pressure in excess of that allowed by the ASME B16.5 Class 2500 rating for the specified design
temperature and material group. However, there are no specified pressure limitations for the
application of these rules.

Normal Fluid Service

This is a fluid service pertaining to most piping covered by B31.3, i.e., not subject to the rules for
Category D, Category M, or High Pressure Fluid service, and not subject to severe cyclic conditions.

From the above fluid services, it is clear that the two most critical services are the category M fluid
service and the High-pressure Fluid service and as such, requires strict compliance with the rules
stated in the code.
The user should also note that whenever CAESAR II reports no code stress check processed, it is an
indication that a material not included in the code was selected; therefore a failure or a pass cannot
be reported because code and allowable stresses were not calculated.

7|Page
MATERIALS SELECTION
Piping materials selection or process piping materials selection as we will consider, is a very important
step performed during the design phase of a process piping system.

Complete discussions on piping materials considerations will not be achieved here, only the factors that
will lead to better understanding of the analysis section will be considered.

No material possesses 100% physical, mechanical, metallurgical and chemical properties. The selection
of a process piping material is not the process of selecting a material with 100% of the required
properties but the selection of a material with the best combination of required properties.

The best here implies that to every process service, there are numerous competing materials that can
perform or go close to performing well under the imposed service conditions. It also implies that
process piping material selection is a trade-off with many factors.

It is important that in stress analyzing a process piping system, that the microscopic properties of the
system materials be known and not just the macroscopic properties knowledge.

The microscopic properties as used here include the properties that define the material side of the
system such as physical, mechanical, metallurgical and chemical while the macroscopic properties as
used here include the physical configuration of the system, such as thickness. The performance of every
piping system is dependent on its micro and macro properties.

8|Page
We now know why it is impossible to talk process piping stress engineering without process piping
materials selection.

We will dedicate this chapter to process piping materials selection, while considering only the micro
properties of the piping material.

Before we proceed, a definition of process piping material selection follows:

It is the process of selecting available finished components based on their known material physical,
chemical, mechanical and metallurgical properties; and geometry for a process service condition which
will meet the expected design life at the least possible cost.

Code requirements have limited the use of certain materials for certain process conditions, thereby
reducing the workload on the process piping materials engineer, but making availability a constraint.
When certain materials are selected for certain pressures and temperatures that are above that
allowed, CAESAR II alerts the user by saying that the service conditions allowed for such a material has
been exceeded. An example of how to select valves and flanges based on pressure and temperature for
a particular material group was given in the chapter explaining Rigids.

From the above definition, we have included the three fundamental requirements in every material
selection process, which are:

1. Availability
2. Properties
3. Cost

AVAILABILITY

Availability does not seem a serious issue to so many engineers out there. This can be understandable
judging from the nature of a service requirement. In lines that qualify as process piping systems, but
that are regarded as Category D fluid service lines, availability becomes almost irrelevant as
nonconforming pipes and fitting materials can be selected for use.

Availability meanwhile becomes very important as the service conditions become more and more
stringent. In a fluid service denoted as Category M for example, the piping system is required to
comply with the requirements of ASME B31.3 code.

9|Page
Several materials are tagged not fit by the code for severe fluid services. It becomes important that the
fit materials are available in order to make feasible projections.

A very good example of the role availability plays in process piping materials selection is on the debate
between seamless pipes and seam welded pipes.

Pipe manufacture

10 | P a g e
The pipe manufacturing process is determined by the material, diameter, wall thickness and quality
desired for specific service. Steel pipes are classified according to their manufacturing process as
follows:

1. Seamless
2. Welded

The seamless category describes a pipe manufacturing method where the walls of the pipes produced
are without welds and are formed from hot, solid, round billets and in sizes up to 26 inches in diameter.
These pipes are often specified for high temperature services ranging from 7500F to 11000F.

The welded category of pipes has a weld in their walls and consists of the following:

i. Electric resistance welded ii.

Furnace-Butt welded iii.
Electric Fusion arc welded iv.
Spiral welded
v. Double submerged arc-welded vi.
Electric flash welded vii.
Furnace-Lap welded viii.
Submerged arc-welded

Mechanically speaking, due to the presence of the weld in the welded category, the seamless group is
better but more expensive. The welded group are also used albeit in less severe fluid services.

In every process piping systems, the number of welded joints is required to be at the minimum and this
can only be made possible when pipes of varied lengths are available. This is not often the case with
seamless pipes whose lengths and sizes are limited.

PROPERTIES

There is no substitute for properties. As properties are for a material, so is stiffness for a piping system.

A system with very good material properties will have a very good stiffness leading to high natural
frequencies which are very important in dynamic events, not to mention excellent response to static
loads. Material Properties determine a system’s performance in a particular service condition. Factors
which affect materials properties in or out of service are especially important to the materials engineer.

11 | P a g e
Consider the cases with the NDT (Nil Ductility Temperature) of a material and the manufacturing
process of a component.
It is known fact that as the temperature on a material decreases, the elastic modulus increases. At a
particular temperature called the NDT, a once ductile material could become brittle due to decreasing
toughness but increasing tensile strength. This phenomenon is a major contributor to fatigue failure.
The initial mechanical properties of a material before it is formed into a useful component might not
represent the mechanical properties of the finished product. Although advances in heat treatment
techniques have dealt considerably with the negative effects of some manufacturing processes on initial
materials mechanical properties, it is also important the materials engineer understands any likely
influence the final manufacturing process had or will have on the component.
It is very important that the materials engineer does not only consider initial material properties alone,
but factors that could increase or decrease them, especially in service.
As a rule, process piping materials engineering requires that materials with low NDT levels be selected,
especially for lines that can experience low temperature conditions.
Again, the process materials engineer does not only consider the materials properties and the loading
conditions, he/she relates these properties to that of the fluid.
We consider the following types of material properties:

Mechanical

Mechanical properties are critically important in the process piping materials selection. They are defined
as the characteristic response of a material to applied forces.
In the dynamic part of this book, during spectrum analysis of a water hammer event, CAESAR II will
generate response spectra for different inputted time history definitions which is the response
produced by a system to dynamic loads. The better the mechanical properties, the better the stiffness,
and the better is the natural frequency of the system leading to excellent spectrum which determines
the piping system’s response to dynamic loading. The time history of a system is its force-time response
dependent on its mechanical and macro properties. Common mechanical properties include:
i. Modulus of elasticity (Young’s modulus) ii.
Yield strength iii. Ultimate tensile
strength iv. Fatigue resistance

The primary reason we select ductile materials over their brittle counter-parts is due to the fact that the
ductile materials fail by necking. Every material in service will fail over time, but the fracture type
(brittle) of failure is undesirable.

12 | P a g e
Those properties which enable materials to be classified as ductile or brittle are the mechanical
properties. Whether it is a ductile or a brittle material, every material possesses elastic modulus.
Materials with higher elastic modulus will have higher load carrying ability and will have longer design
lives.
Above the yield strength, failure occurs in ductile materials, but it is termed plastic deformation
(although it is not regarded as failure if the plastic deformation is due to a secondary load which does
not lead to incremental collapse).
Tougher materials exhibit wider plastic deformation ranges. The ultimate tensile strength determines
the highest load carrying ability of a stressed material. This strength determines the maximum stress a
material will carry before fracturing. Piping materials are required to be tough and ductile.
The ultimate tensile strength of a material depends on the percentage carbon present. The higher the
percentage carbon content, the higher the hardness and ultimate tensile strength and the lower is the
toughness and ductility except special elements are added to introduce ductility and toughness at the
expense of the high tensile strength. This is what alloy steels are known for.
Very hard materials exhibit low toughness and are often regarded as brittle. Therefore, processes which
harden a material in service, like strain hardening in process piping materials is undesirable because it
means a drop in toughness.
It is important to mention here that strength is desirable in process pipes, but not at the detriment of
toughness.
Under cyclic loading conditions, the endurance limit of the piping material is key. In the course of this
book, we will see that because a material has high elastic modulus plays little part in determining its
ability in a cyclic condition.
The best way to keep fatigue failure in check in all materials is to ensure that the load magnitude is low
(by low we mean lower than the endurance limit).
The endurance limit is the stress level below which a material can withstand an infinite number of load
cycles without failing.
In cases where the applied load is lower than the elastic limit (.i.e. in the elastic region) fatigue failure
can occur due to increasing load cycles, decreasing temperature and the presence of notches which
raises the fatigue stress above the endurance limit not necessarily past the yield strength (for systems
with considerable number of load cycles). Fatigue failure due to high load cycles of low magnitude loads
can be completely eliminated by materials selection, by ensuring that the predicted applied load and its
corresponding cycles are below that permitted by the S-N curve of the material. Fatigue failure due to
low load cycles of high magnitude loads can be eliminated by stress analysis by introducing techniques
that will reduce hot loads, although the failure process can be slowed when materials with high

13 | P a g e
endurance limit are selected. Also fatigue failure due to stress risers (notches) can be eliminated by
ensuring that pipe surfaces are not fouled by physical, mechanical or chemical processes in or out of
service.
Determining the endurance limit of materials is only done by tests and the result is dependent on the
test temperature, the material defects, materials’ surface condition etc.
CAESAR II only simulates the mechanical properties of materials.
Physical
Physical properties are those, other than mechanical properties, that pertain to the physics of a
material. Physical properties of importance to the materials engineer are material density, thermal
conductivity, thermal expansion, and specific heat.
In most cases, the materials engineer is interested in the extensive properties of the material which is a
ratio of its mechanical properties to its physical property-density.
It is not uncommon for materials with different physical properties to be joined, but care must be taken
to avoid several negative effects which include thermal cracking.
A lot has been written about the negative effects of self-weight of piping systems, that even several
advanced materials called composites have been developed. These advanced materials have better
extensive mechanical property called specific modulus than their widely used metal counter-parts.
Metals might have better elastic modulus, but when compared with density, the result favours
composites.
It is the heterogeneous and anisotropic nature of these materials that have limited their use in process
piping services not to mention other disadvantages including their huge costs and difficult (if not
impossible) repair methods and so they are referred to inexperienced materials by ASME B31.3.
Extensive mechanical properties are more important than pure mechanical properties in the materials
selection process because they relate pure mechanical properties with density.
It is the extensive mechanical properties that determine piping systems response to most loading
conditions.
Chemical
In ASTM material specification, this is what is regarded as grade. The chemical properties include those
that surface during chemical processes.
The chemical properties of components in a piping system are required to be similar, but in cases when
they should be unavoidably different, efforts should be made to ensure that the different components
do not make direct contact with one another.
CAESAR II does not simulate what happens when materials with differing chemical properties are joined.
The effects of directly joining two chemically different materials include galvanic corrosion.

14 | P a g e
The chemical properties of materials include:
Corrosion resistance
On the basis of chemical composition, process piping materials have three major classes:
1. Carbon steels
2. Low-alloy steels
3. High-alloy steels
Except Iron and carbon, every other elements found in steel (Iron + carbon) are referred to as alloying
elements.
The carbon steels according to ASTM is any steel having less than 1.65 percent manganese, 0.6 percent
silicon, and 0.6 percent copper, with the total of these other elements not exceeding 2 percent. By this,
carbon steels are regarded as unalloyed.
The low-alloy steels are regarded as steels with a total alloying content of up to 8 percent and any steel
with greater than 8 percent of its alloying content is regarded as high-alloy steel.
It is also worth noting that in the class of carbon steels, the following types are defined:
High carbon steels (carbon content greater than 0.5 percent)
Medium carbon steels (carbon content between 0.2 to 0.49 percent)
Low carbon steels (carbon content with 0.05 to 0.19 percent)
Extra-low carbon steels (carbon content between 0.015 to 0.05 percent)
Ultralow-carbon steels (carbon content less than 0.015 percent)
Whatever the steel type, all steels contain some percentage of alloying contents which include;
vanadium, nickel, chromium and cobalt. These alloying elements are then added intentionally but with
their quantities defined to get the alloyed steel category.
As the corrosion requirements of a particular process piping system increase, the degree of some
corrosion resistant alloys like nickel is required high and the selection might move from the carbon steel
category to the alloy steel category.
The other microscopic property called metallurgical properties includes grain size and hardenability.
These properties determine other properties, especially the mechanical properties. They are
structurerelated.

ASTM Material Designation

ASTM (American Society for Testing and Materials) developed a collection of documents for the piping
industry for the standardization of pipes and its attendant components.
Specifications starting with “A” are for ferrous metals; “B” is for non-ferrous metal; and “D” is for plastic.

15 | P a g e
The specification of a pipe or component by ASTM does not only include the material category and
chemical composition, but also includes the process in which the material was shaped into the final
product. This is so because, just like we pointed out in the materials selection considerations, the
properties of a pipe are affected by the shaping/manufacturing process.
This is why for any given material type; there are specifications for seamless, welded, wrought, cast and
forged. It should also be noted that due to the nature of these manufacturing or shaping methods, the
components that can be produced by each is limited. Pipes used in the chemical or refinery piping are
mostly seamless for critical and high temperature lines and welded for low temperature lines and lines
not conveying volatile fluids. The cast and forged are used for Rigids .i.e. valves and flanges, while the
wrought group is mainly used for pipe components.
Users can read up the different manufacturing methods in any good metallurgy book.
The following table shows different specifications for carbon steel material dependent on the shaping
process:
SEAMLESS PIPES ASTM A53 Gr. (A &B); A106 Gr.(A, B and C)

WELDED PIPES ASTM A134/A135/A139

WROUGHT FITTINGS ASTM A234 (WPA and WPB)

FORGED FITTINGS A105

CAST PARTS A216 (WCB)

As can be seen from the table, the material A106 Gr B used throughout this book is a seamless pipe. It
should also be noted that seamless specification for material A53 Gr A and B type S is becoming
outdated and replaced by A106 Gr B. Most A53 specifications are now welded.
The next time you hear A106 Gr B, know that a seamless carbon steel pipe is been referred to for
stringent process conditions.

A106 Gr. B EXAMPLE

A means ferrous metal

10 carbon steel
6 means the min percent carbon is 0.06
Gr. B means the steel grade which describes its chemical composition. The Grades A, B and C of this
material are only different in their percentage carbon content since it is a carbon steel. The max carbon
content for Grade A is 0.25; Grade B is 0.30; and Grade C is 0.35. Judging from the permissible carbon
range for each grade, it is obvious that the material, A 106 does not include the extra low, ultra-low and

16 | P a g e
high carbon steel types. This is because of the increased strength, very low ductility and low weldability
of the high carbon steel type, not to mention the high cost and the low strength of the extra and
ultralow carbon groups.
An increase in the grade letters for carbon steels indicates an increase in the tensile strength and a
reduction in ductility due to increased carbon content and insignificant percentages of other alloying
elements.
As the service environmental conditions become more and more stringent, the selection of a piping
material moves from the carbon steel category to the alloy steel category of which stainless steel is part.
This is because for a marine environment for example, the material does not only contend with the fluid
temperature and pressure; fluid properties and other conditions, but with corrosion from the
environment. Carbon steels have poor corrosion resistant properties due to the insignificant
percentages of other alloying elements, so they are not used as piping materials to convey volatile fluids
in adverse marine environments. If used, corrosion degrades the material, and the system can either fail
under hoop stress due to high pressure and reduced thickness (stress corrosion cracking) or from fatigue
(due to increasing surface roughness) under cyclic conditions, this is what is referred to as corrosion
induced fatigue failure.

COST
Cost becomes a factor to consider only when the component is available and the properties of the
materials or components have been fixed. Cost is important because in certain occasions, more than
one steel grade can meet a service requirement. As the material grade increases, the cost increases.

STRESS AND ANALYSIS TYPES Statics

17 | P a g e
Stress is the response given by a material when a force is applied on an area of the material. The
response given by a material is dependent on its micro and macro properties like earlier stated.

Consider the relationship below:

E = σ/ϵ ……(1) where;

E = material elastic modulus

Stress = σ F/A; F = applied force; A = material cross sectional area

Strain = ϵ = ΔL/L; ΔL = change in pipe length due to applied force; L = original length of pipe Rewriting

equ. (1) results:

E = (F/A) / (ΔL/L)…..(2)

From equ (2) above, we wish to make F subject of the relationship:

F = (E*A* ΔL) / L……(3)

K = (E*A) / L where k is the material stiffness

Rewriting equ (3):

F = k * ΔL……(4)

From equation (3), we see that the force required to deform a body must overcome the geometry of the
material (macro property) and the property of the material called the elastic modulus (micro property).
The greater the stiffness of a material, the greater is the force required to cause deformation (equation
4). Larger diameter pipes require greater applied force to cause deformation. The elastic modulus is
peculiar to every material. It is because of the response of the stiffness of a piping system to different
conditions, including the environment that process piping materials selection is hugely important and
the maintenance of the system is required so failure does not occur within the predicted design life. The
stiffness of a system responds to fluctuating process conditions which affects the elastic modulus of the
piping material and corrosion which degrades the system’s thickness.

Loads

Loads on piping systems arise from weight, temperature, pressure, man, animals, wind, connected
equipments etc.

18 | P a g e
The above list can go on, making the stress analyst feel correctly that stresses can develop from several
sources on his/her piping system.

There are basically two types of analysis:

1. Static analysis
2. Dynamic analysis

Static analysis involves loads that give the system time to distribute such loads and remain in
equilibrium. The rate at which those loads change with time is very small or zero.

Loads can be:

i. Primary ii.
Secondary iii. Peak

Primary stresses

Primary loads on a piping system are a combination of dead and live loads. They are also called primary
sustained loads or sustained loads.

A dead load is any load the system will experience throughout its operating life. They are constant.
They include:

i. Pipe weight ii.

Components weight iii.
Weight of insulation etc

A live load on the other hand is any load that process conditions could decrease or increase their
density (effects). They include: fluid density.

Primary sustained loads are not self-limiting in the fact that whenever these loads are a cause for
concern in the piping design, their effects should not be nullified by the system’s redress without
external help.

The two factors responsible for primary sustained loads are weight and pressure. These loads are only
collapsed by any one of the following:

a. Strain-hardening of the piping material

b. Use of supports or sufficient thickness

19 | P a g e
In the first case above, the material endures rather than relieve these loads. Strain hardening is a
strengthening phenomenon, which brings the system closer to a brittle failure (fracture). It is because of
the strain hardening of ductile materials that they are preferred over their brittle counter parts, but the
process is only a forerunner to fracture as we will see later. It is a common phenomenon which occurs in
ductile materials when applied stresses overcome the yield strength of the piping material. The primary
purpose of the piping codes ensures that the applied primary stress does not overcome the yield
strength of the material because any further strain hardening due to secondary stresses can cause a
collapse of the system. Strain hardening should not be counted upon to overcome primary loads for the
reason mentioned. CAESAR II ensures this!

The use of supports comes into play when the system experiences displacements arising from weight. It
is required therefore, that the supports introduced will make the displaced nodes experience
equilibrium by carrying some or the entire sustained load (weight) of the piping system at that particular
node. Such is not the case with pressure. The pressure forces on the pipe walls can only be absorbed by
tension in the pipe walls. The use of supports does very little when hoop stresses arising from pressure
forces cause displacements which could cause the pipe to burst in severe cases. In this situation, only a
matching thickness to the problematic pressure can solve resulting stresses.

Weight

As the density of pipe material increases, the self weight of the piping system also increases.

Weight is undesirable in structural designs but it is unavoidable. Even if we can make use of air as our
material of construction, it has also got weight. It is the responsibility of the stress engineer to manage
this undesirable quantity called weight by the excellent use of supports.

The negative effect of weight becomes pronounced when the support spacing does not comply with set
code and standard. In that case, weight causes bending of a straight run.

Support spacing is the distance between any two supports in the same line of action.

It is a very important thing to know in stress analysis of piping systems and support installations. We
didn’t discuss it in detail because we will rely on CAESAR II’s stress outputs to attach supports at
required locations.

20 | P a g e
Pressure

Pressure is the force per unit area on a body. The pressure of the fluid and the environment and
environmental activities must be contained by the pipe thickness and must result in zero force on pipe
supports.

The design pressure must not be lowered to solve circumferential stress problems .i.e. hoop stress by a
stress engineer if that pressure was certified by the process department as this will jeopardize the
energy consideration of the system/line. The design pressure must not be less than the pressure of the
fluid according to our code. One (if not the only) solution to pressure containment is the accurate
thickness sizing.

Even though we know and have said that pressure acts to expand the pipe laterally, it could also cause
significant contraction in length that could pose serious stress problems in the axial direction if non-zero
hoop stress values are part of our system.

The primary stress limits defined by the codes are to prevent plastic deformation and bursting.

Secondary stresses

Secondary stresses are developed by the constraint of displacements of a structure.

These displacements can be caused either by thermal expansion or by outwardly imposed restraint and
anchor point movements.
Under this loading condition, the piping system must satisfy an imposed strain pattern rather than be in
equilibrium with imposed forces. Local yielding and minor distortions of the piping system tend to
relieve these stresses. Therefore, secondary stresses are self-limiting.

Secondary stresses alone cannot cause failure. They result because there are primary loads.
Examples include: thermal expansion, outwardly imposed restraints, anchor point movement etc.

The primary plus secondary stress limits defined by the codes are to prevent excessive plastic deformation
leading to incremental collapse.

Pressure elongation

Pressure elongation is the Poisson-type effect or bourdon effect created by a pressure force. What
happens here is that as the hoop stress from the pressure forces causes an expansion of the pipe

21 | P a g e
diameter, it also causes a slight change in the axial direction of the pipe, especially when restrained in
that direction.

To avoid this poisson effect that hoop stress creates, pipe schedules should be specified to contain or
reduce resulting hoop stress from pressures far below the code stress allowable.

Peak stresses

These are highly localized stresses that exist at a discontinuity in the load path or stresses developed
due to fluctuating process conditions. If a perfect material system (continuum) experiences stress cycles,
these cycles will create a void leading to a crack initiation, growth and consequently failure by
fracturing. They require special analysis to limit their devastating consequences. A piping system might
be under the simultaneous attacks of primary, secondary and peak stresses.

When peak stresses are present, failure modes associated with them are catastrophic, but the
phenomenon of failure here takes a relatively higher time in ductile materials than in their brittle
counterparts. They are the result of discontinuities and thermal gradients through a pipe wall.

The peak stress limit defined by the code is to prevent fatigue failure resulting from cyclic loadings.

Fatigue
Fatigue is a failure phenomenon associated with tired engineering materials. We have used tired
because these materials or systems failed because they have been over-stressed by more than
predicted stresses. They could not have failed carrying only the predicted stresses. The stiffness of
piping systems responds to environmental conditions. This response implies a deteriorating material
property with respect to time. Under a high temperature and stress for example, creep deforms the
material over time. Changes in the environment and process conditions cause the material to withstand
stresses far lower than those initially predicted. There have been reported cases of failure of systems
not half way into their predicted design lives. Such systems failed while still in their initial elastic region.
There are other systems that failed not as a result of depreciating mechanical properties, but as a result
of increasing strain rates and from higher than predicted stresses.

22 | P a g e
An initially continuum material operating under a mild environment predicted to withstand an applied
stress of 100 mPa can fail under an applied stress of 50 mPa, as a result of surface flaws developed in
service. Surface flaws are stress risers. This is why processes that can lead to surface flaws should be
eliminated to prevent fatigue failures.
The fatigue analysis (and other analysis) performed in CAESAR II are for defects-free and polished
surfaces. As a cautionary measure, to account for uncertainties that might be developed in course of
operating a system, the fatigue or intensified stresses should be reduced by several possible techniques
so in the event of unpredicted surface fouling; the physical system does not fail.
Fatigue failure starts at stress concentrations initiated by square holes, sharp corners or discontinuities.
For every material, a fatigue curve, also called the S–N curve, can be generated by experimental tests
which correlates applied stress with the number of cycles to failure. For high-cycle fatigue, the analysis
is to determine the endurance limit, which is the stress level that can be applied an infinite number of
times without failure.
High-cycle fatigue is a fatigue type where the number of loading cycles is in the millions and can be
considered infinite for all practical purposes and the low-cycle fatigue is a fatigue type where the
loading cycles applied seldom exceed 105 and are frequently only a few thousand.

Piping systems however experience the low-cycle type of fatigue where the fatigue limit or endurance
limit determination shows a lot of scatter and uncertainty.
One reliable method of appraising the fatigue failures of piping systems with low-cycle fatigue is with
the use of cumulative usage factor which combines the effects of all the stresses associated with
number of cycles and predicts fatigue failure at a value greater than one.

Dynamic analysis
Dynamic analysis involves loads that do not give the system time to distribute such loads. Just like the
fatigue failure, the failure associated with dynamic loads involves little or no plastic action.
They result in local or in extreme cases, global failure in the piping system.
All load categories previously mentioned can also be dynamic provided they are moving or not
experienced by the entire piping system.

23 | P a g e
Dynamic loads are majorly from external sources (connected equipment) and on few occasions are self
imposed.
Dynamic analysis in CAESAR II can be performed by the following (each with its own objective):
1. Modal analysis
2. Harmonic analysis
3. Time history analysis
4. Response spectrum

Modal analysis
Modal analysis is performed to get the piping system’s natural frequencies. The natural frequencies of a
body are important to know so that the frequency of any applied periodic loading will not coincide with
it and cause resonance which leads to structural damage.
This analysis is the primary dynamic analysis that will be performed before other types of dynamic
analysis.
It extracts the system’s eigenvalues and eigenvectors.
To every eigenvalue, there is a corresponding eigenvector associated.
The eigenvector is a unique vibration mode associated with a particular eigenvalue.
The eigenvalue is also called natural frequency, while the eigenvector is also called mode shape.

Harmonic analysis
Any sustained cyclic load will produce a sustained cyclic response (a harmonic response) in a structural
system. Harmonic analysis or Harmonic response analysis gives the ability to predict the sustained
dynamic behaviour of structures, thus enabling the analyst to verify whether or not the system will
successfully overcome resonance, fatigue, and other harmful effects of forced vibrations.
With this type of profile, the load changes direction and/or magnitude following a harmonic profile,
ranging from its minimum to its maximum over a fixed time period.

Example of harmonic loads is the pulsation associated with a reciprocating pump.

Time history analysis

A time history analysis gives the response of a structure over time during and after the application of a
load. It involves solving the equation motion of the structure. It is not used in random system’s
response.

24 | P a g e
The structure’s equation of motion is a mathematical equation describing the motion of a structure over
time.
It is the method employed in solving loads with a rectangular profile called impulse.
Example of an impulse load is given below:
When system pressure reaches a dangerous level, relief valves are set to open in order to vent fluid and
reduce the internal pressure. Venting through the valve causes a jet force to act on the piping system;
this force ramps up to its full value, from zero, over the opening time of the valve. The relief valve
remains open (and the jet force remains relatively constant) until sufficient fluid is vented to relief the
over-pressure situation. The valve then closes, ramping down the jet force over the closing time of the
valve.

Spectrum Analysis
A spectrum analysis represents an attempt to estimate the maximum response developed in a system
during a transient load. The results are a statistical summation of the maximum displacements, forces,
reactions, stresses, etc; the individual responses do not represent an actual physical loading case in that
the maxima may all occur at different times. Spectrum analyses are especially useful when the loading
profile is random, or otherwise not known exactly, such as with seismic loads. CAESAR II provides the
ability to perform two types of spectrum analyses (which may be combined): seismic and force loadings.
Seismic loadings may be evaluated either uniformly over the entire system, or applied through
individual support groups (with corresponding anchor movements). Force spectra analyses may be used
to analyze impulse loadings (like time history), such as those due to relief valve, fluid hammer, or slug
flow.

CAESAR II Load Cases

Operating
In the operating case, the simultaneous influence of weight, pressure and temperature on the piping
system is investigated. We have already considered the influence of weight, temperature and pressure
leading to self-limiting and non-self-limiting loads.
However in CAESAR II, failure does not occur in this case. Not that these three (W, T, AND P) together
does not cause failure but other load cases will specially investigate these factors and report stresses as

25 | P a g e
necessary. The primary effect to note here is pipe lift where the pipe lifts up a support due to secondary
loads making the support(s) inactive.
Whenever there is a pipe lift, the use of supports that can carry pipe sustained loads while allowing for
thermal movements is called for.
Other cases like the expansion, sustain, and occasional loads are therefore investigated.

Sustained
This case reports failure if there is one in the piping system. This case reports stresses based on primary
sustained loads. Since this load case reports stresses from weight and pressure, it takes the ingenuity of
the stress engineer to resolve failure from this case.
The actual cause of any failure in this case is a result of pressure alone, weight alone or the combination
of the two.
The different stress types associated with these load cases will be explained and a better understanding
of how to know the root cause of failure in this case will be revealed.

Expansion
This case basically tells the extent of secondary effects on the system. Temperature here is the only
factor to be investigated, although sometimes in the presence of other loads like displacement loads. If
failure occurs in this case, it is only due to temperature or temperature and other defined loads. It is a
difference case in the sense that the magnitude of load found here is the difference of the previously
explained cases.

Occasional
This load case is used to investigate occasional loads. Occasional loads are either dynamic or static like
snow on the piping system, loads transferred to the piping system from connected equipment, although
it is mostly used in dynamics.

Fatigue
This case is used to analyze cyclic peak stresses. It can be used in either statics or dynamics.

Stress types

26 | P a g e
Bending stress
The bending stress on a piping system is the normal stress that is induced at a point in the piping system
subjected to loads that cause it to bend. When a load is applied perpendicular to the length of a piping
system (with two supports on each end), bending moments are induced in the system.
It is very important to note that bending can occur in both the sustained and expansion load cases. The
two primary causes of bending in the sustained case are too much weight and inadequate support
spacing.
Whenever these two are high, the chances of inducing bending moments in the system are high.
Remember, the sustained loads require that the system be stiff and in equilibrium. Therefore, whenever
there is a bending stress failure in the sustained case, the weight and distance between two supports
are to be reviewed or the system is very flexible and requires stiffness from the installation of supports
at appropriate locations.
If the bending stress is in the expansion case, flexibility should be added and SIF reviewed.

Torsion stress
This is the stress or deformation caused when one end of an object is twisted in one direction and the
other end is held motionless or twisted in the opposite direction.
Failure like this in a piping system results from wrong piping route. Whenever present, the following
should be considered:
i. A change in the piping route ii. Increase
in pipe diameter iii. Reduction in the
number of elbows iv. Resolve
moments at pipe boundaries

Axial stress
These stresses are developed in a piping system due to long lengths and excessive restraint. Their values
can rise to cause bending stress failure in the expansion case. They are due to secondary loads.

CAESAR II always reports bending stress failure when the axial stress is high. Whenever there is a
bending stress failure in the expansion case, the solutions outlined in bending stress above is actually
geared towards reducing the axial stress.

27 | P a g e
Hoop stress
Hoop stress is the result of the no-equilibrium condition that results from the interaction between fluid
pressure and the pipe wall. The result of this stress when severe leads to pipe bursting and it also
contributes to axial stress.
The solution to a hoop stress problem in a piping system is to specify appropriate pipe thickness, use
fittings of appropriate pressure rating as defined by our code and manage very well the resulting
increase in weight. When pressure is high, the thickness specified is geared towards reducing the
resulting high hoop stress to a value below the code defined allowable.

Code stress
The code stress in CAESAR II is the cumulative stress acting on the piping system. The program reports
failure if this value is greater than the allowable stress permitted by the code for the material of
construction and the design conditions.

Materials failure
Failure is not only the breakage of a piping component or system. It includes any slight increase in
applied stress which overcomes only the yield strength of the piping material and takes the material
into the plastic region. This is the basis of linear elastic stress theory and is referred to as the yielding. All
fractures are failures, but not all failures are fractures.
There are basically two types of failures as it concerns engineering materials:
i. Ductile failure ii.
Brittle failure
The brittle form of failure is catastrophic in nature and will not be discussed because all piping materials
are ductile, although not for normal fluid services where there was no compliance to code. It is because
of linear stress analysis that we appropriately use fracture for brittle materials and rarely use fracture
for ductile materials.
Ductile failure

Ductile failure in ductile materials is desirable in piping system designs because of the time and extra
stress it takes the material to deform plastically. During plastic deformation or strain hardening, the
material displays signals of failure rather than fail unexpectedly. It is an announced failure.

28 | P a g e
The part on the stress-strain relationship for a ductile material above labelled a is called plastic
deformation. The higher the range of a, the tougher is the material.
CAESAR II reports failure in a piping system if any applied stress overcomes the Yield Strength (YS), even
though this applied stress greater than the yield strength is lower than the Ultimate Tensile Strength
(UTS). Above the YS, a non-linear relationship between the applied stress and strain sets in.
Materials with a no or very small a and a high UTS are referred to as brittle. Brittle materials failure can
be completely eliminated by selecting materials with a high a and compliance with the set design code.
There is a special type of failure however, where the initial material selected is a ductile one, but the
failure mode is associated to a brittle material. This kind of failure mechanism can be accelerated by the
following propagators; decreasing temperature (into the sub-zero region), increasing strain rates and
the presences of stress concentrations due to notches, defects and SIFs at pipe joints. This failure
mechanism is referred to as fatigue.
What these fatigue failure propagators do is that they reduce the a of the material in service. As a
decreases, the toughness lowers, and the UTS increases. The process of increasing the UTS at the
expense of a is called strain hardening. When strain hardening occurs in a piping material, we say the
material has deteriorated.

29 | P a g e
It is because of the possible brittle failure of ductile materials even after piping stress analysis that
process piping stress engineering ensures that:
1. Materials selected for a particular process does not experience very low temperatures not
permitted by the materials’ ductile regime.
2. The materials engineer must ensure that the surfaces of the materials or components are
polished or defects free during design and in service.
3. The load cycles associated with a particular stress amplitude are not greater than those
permitted by the material’s S-N curve.
4. In the design and analysis of the system, that components used in critical lines have higher fillet
radii as in the case between short and long radius elbows.
5. Stress amplitudes are low, by selecting materials with good elastic modulus and by reducing
SIF’s at joints.

30 | P a g e
Input parameters

Node numbers
The essence of node numbering a piping system is to easily identify culprit elements during analysis. It
also makes modelling easy. Away from those reasons though, node numbering is the act of meshing.

Earlier, we said CAESAR II’s approach to stress analysis is the same finite element analysis (FEA)
approach that we are probably familiar with. In theoretical FEA, meshing is the discretization of the
system to form separate elements all connected at appropriate nodes. It is the act which makes our
assumption of the fact that the material is perfect and a continuum true. It basically allows Hooke’s law
to find application at the element level.

You will probably notice that in the event of analysing our model, the program places extra nodes
between our already defined Nodes. That is because, the finer the elements (ie smaller), the better the
results (it approaches the real solution). So, to the nodes we specify during modelling, CAESAR II adds
extra nodes to make our already meshed model finer.

During or after stress calculations, the user is bound to say somethings about his/her system. It becomes
necessary that he/she does not describe the system in whole as this will be cumbersome and lacking in
information. Nodes are the only means of solving and describing our system in detail.

Dx, Dy, Dz (piping vectors)

Elements like those of the piping system are vectors. If you must need reminding that a vector does not
only posses magnitude but also a specified direction. The x, y and z attached to the “D” are the three
directions in the physical space. The “D” before these letters notifies the user that there are magnitudes
associated with these directions.
The mass of an element should not be placed in these cells (they are for geometric quantities alone).
The program automatically adds the weight of every element except they are not piping elements.

31 | P a g e
Diameter
The diameter required here is the size of the pipe. This corresponds to the outside diameter of the pipe.
The internal diameter of the pipe is the area fluid exits or enters a pipe. It is the difference between the
outside diameter and two times the pipe thickness. Notice that whenever a diameter that is less than or
equal to 12inches is entered, the program converts it to the actual size. The mean diameter of a pipe is
the internal diameter plus the thickness of the pipe.

Wt/Sch
The program checks the integrity of the piping design. Therefore, just like every FEA tool, it is not
interested in the diameter of the system alone, but also on the thickness. The thickness of the system
contributes to the dead weight on the system. The thickness of the system also increases cost. The
thicker the piping system gets; the more the weight and cost and the lower is the piping materials
extensive mechanical property-specific modulus. Remember, weight is an undesirable quantity in the
design of efficient structures. It should also be emphasized that the system’s ability to withstand
pressure lies in the accurate call made in this cell. A lot has been written about pipe minimum thickness
calculation in relation to pressure integrity and a very good understanding of what this cell does is the
single most important means of getting an accurate and fitting pipe thickness.
Pipe thicknesses are standardized. That is, a pipe size of 2 inches with a standard schedule has the same
thickness worldwide irrespective of material of construction and process conditions (ASME B36.10 M).
ASME B36.10 M is a standard which covers welded and seamless wrought steel pipes.
Schedule is only a term that enables standardization of pipe thickness. Rather than say, “this is a 10inch
pipe with a thickness of 0.5inch”, say “this is a 10inch pipe with a schedule of 60”. Schedules are a way
of universally and easily describing the thickness of pipes. The “wt” means weight and the “sch” means
schedule.
Of course you can explicitly enter the pipe thickness if it is known, but schedules are a better way of
specifying the thickness especially for systems about to be installed.
It is also important that we bring to your notice that regardless of thickness or internal diameter, all
pipes with the same size have the same outside diameter. For example, all pipe size of 6-inches in the
world has the same outside diameter, but the internal diameter could vary due to varying schedule .i.e.
6 inches or 168.2750 mm (actual value in mm) sch 10 pipe has a thickness of 3.4036 mm (internal
diameter = 168.2750 - 2*3.4036=161.4678 mm) while a 6-inch or 168.2750 mm sch 20 pipe has a
thickness of 7.1120 mm (internal diameter=168.2750 – 2*7.1120= 154.051 mm).

32 | P a g e
Caution: the user is warned that whenever a different schedule is requested by the stress department,
the internal diameter of the pipe which has a direct relationship with the energy of the system is
affected. In cases where there is an increase in schedule, pipe internal diameter decreases and energy
drops which might not be tolerated, therefore, in increasing schedule, it has to be the only means of
reducing stress and will therefore need a review from the process department which could lead to a
delay in project execution.

Corrosion allowance
Corrosion allowance is the deliberate increase in the weight of a system as a result of added mass
arising from corrosion inhibitors. It could also be a deliberate oversize of the system’s required
minimum thickness so as to prevent degradation of the actual system’s minimum thickness during the
system’s design life.
It enables the system not to fail under a circumferential stress type called hoop in the compulsory event
of corrosion in some fluid services.

When corrosion is anticipated to occur at a slow, regular rate and this rate can be reliably predicted, it
may be accommodated by the provision of excess material in the pipe wall which is a calculated value
dependent on corrosion rate known as the corrosion allowance (typical value is 1/16-inch in low
corrosion rates and ¼-inch in severe corrosion rates). This excess material will be consumed over the
design life (the total expected years of service of the system) of the piping system and therefore cannot
be counted upon to serve any other purpose such as pressure integrity, or structural strength. The
corrosion allowance specified no matter how small reduces extensive mechanical properties and can
cause sustained failure due to extra weight and also lowers natural frequency thereby increasing the
ease of the piping system to dynamic excitation.
In cases in which the corrosion rate is prohibitively high or would result in unacceptable contamination
of the process fluid, a pipe lining or corrosion-resistant piping material should be specified. In cases
where the fluid velocity is high, washing away of the lining could result which can cause dynamic events
in the line.
It should not be taken as erosion allowance.

33 | P a g e
Insulation thickness
Insulation is geared towards reducing the temperature loss to the environment. CAESAR II requires the
thickness of insulation because it contributes to the gravity load (weight) of the system which increases
the dead weight. An insulation thickness will have an insulation density peculiar to the insulation
material alone.
The following materials are known insulating materials commonly used in the commercial piping
industry:

● Calcium silicate insulation

● Cellular glass insulation
● Elastomeric foam insulation

If an insulation thickness is specified without the insulation density specified, CAESAR II calculates and
uses the resulting density of its default insulation material-calcium silicate.

Temperature
CAESAR II allows for operating temperatures of up to nine. This is because; very few systems will
operate at a steady temperature range throughout their operating life. If only one temperature is
specified here, it tells the program that the system is steady and it will reflect itself in the recommended
load cases with respect to the applicable code. Meanwhile, the recommended load cases are those
suggested by the applicable code depending on the set process conditions. Otherwise, it should be
treated or analyzed under the fatigue stress type.
The operating temperatures the system will experience should be used here.
If for whatever reason, the analysis should be done with just a single temperature, then the extreme
operating temperature must be used. In most of the cases, and according to the standards of the code,
it is required that all other temperatures (offset temperatures) the system will experience be specified.
This is particularly important to prevent fatigue failure due to the usage contributions of the low
temperatures which could result in a cumulative usage ratio greater than 1 (fatigue failure).

Pressure
As the internal diameter of a pipe increases, the hoop stress acting on the system with regards to
pressure increases at a constant thickness. As the pressure increases there is a need for a material with
greater strength, because it might combine with the temperature of the system and cause failure.

34 | P a g e
Appropriate pressure ratings of pipe components must be done with respect to the pressure, material
group and temperature of the system.

Just like temperature, the program permits for offset pressures from the normal of up to nine. The last
cell is the hydrostatic pressure of the system which is necessary in leakage examinations. With hydro
pressure, the effect of pressure exerted by a fluid at rest is analyzed on an idle piping system.

Bend
Pipe bends are one of the most used pipe components. The pipe bend is any piping component which
changes the direction of the system and consequently the flow. There are standard elbow angles, which
are:
i. 30 ii. 45 iii.
90 iv. 180 (return)
It should be noted that whenever a fluid changes direction is a piping system, there is a minor increment
in energy, which introduces a downward force on subsequent elbows. If appropriate radius of the elbow
is not used, a stress problem could arise. In its draft form, a pipe elbow is an arc of definite radius, angle
and length.
In other words, the radius, length and angle of the elbow are what define it. The radius of the elbow
could either be short or long.
Short radius elbows are pipe elbows which have the same radius as the outside diameter of the pipe.
These elbows should not be used in primary lines. By primary we mean, in the case of a jacket piping,
they should not be used as the core piping components. Moreover, a jacket piping like the name implies
is a kind of piping system where a very hot fluid is transported via a pipe (core), but shielded by another
piping system (jacket). This kind of system is often done to avoid heat loss to the environment. In this
case the jacket, often made up of an insulating material retains the heat from the core and reheats the
fluid inside the core.
The long radius elbow has the same radius as the pipe outside diameter multiplied by 1.5. The long
radius elbow because of the increased radius improves fatigue life. The long radius elbow is always
chosen in critical lines.
The length of the elbow is a very important physical property of the elbow. The length has its formula,
but, there is a “standard pipe and fitting chart” available. This chart has the lengths of the
aforementioned pipe elbow standards. If the elbow in question is a standard one, then it is required
that to ease the load on the analyst, the appropriate elbow length should be extracted from this
chart. The formula to calculate the elbow length is: tan Ө /2 *1.5*Ps*25.4

35 | P a g e
Where: Ө = elbow angle

Ps = pipe size (outside diameter)

The 25.4 here is only a conversion factor from the inch to millimetre.
The angle of the elbow only directs the system to a new direction. It is the direction of the suction
length measured in degrees from the discharge position.
Flanges can also be attached to elbow ends to give added stiffness which reduces SIF at the expense of
flexibility.

Rigid
Rigid includes flanges and valves. They are very stiff and heavy. They don’t respond to temperature
changes. The factors that determine rigid selection is the design pressure, design temperature, and
material of construction. They have their pressure ratings and should be selected only on those bases.
Considering cost while complying with the ASME standard, the materials engineer will work with the
maximum operating pressure a class of rigid can carry. Consider the example below:
It is required to transport fluid from location A to location B at a temperature of 6500F, a pressure of
1060psig (pounds per square inch gage) and the piping flanges are made of ASTM A105. What class of
flange will be most appropriate?

Pressure-Temperature Ratings for ASME B16.5 Flanges Made from Material Group 1.1 Materials
Class
Temp., oF
150 300 400 600 900 1500 2500

-20-100 285 740 990 1480 2220 3705 6170

200 260 675 900 1350 2025 3375 5625

300 230 655 875 1315 1970 3280 5470

400 200 635 845 1270 1900 3170 5280

500 170 600 800 1200 1795 2995 4990

36 | P a g e
 600 140 550 730 1095 1640 2735

4560

650 125 535 715 1075 1610 2685 4475

700* 110 535 710 1065 1600 2665 4440

750* 95 505 670 1010 1510 2520 4200

800* 80 410 550 825 1235 2060 3430

850* 65 270 355 535 805 1340 2230

900* 50 170 230 345 515 860 1430

950* 35 105 140 205 310 515 860

1000* 20 50 70 105 155 260 430

* For notes concerning the use of carbon-steel materials covered by this table at temperatures over 700 0F
(3710C), see ASME B16.5, Table 2-1.1.
Pressure values in the table are given in psig and are the maximum working pressures.
To convert pressures to kPa gage, multiply tabular value by 6.9.
Source: Adapted from ASME B16.5, Pipe Flanges and Flanged Fittings, American Society of Mechanical Engineers,
New York, 1996, Table 2-1.1, p. 15.

First, the duty of the materials piping engineer is to look through the table above. For material group 1.1,
the following materials are part (grouped under their final processing methods):
FORGINGS CASTING
A 105, and A 350 Gr. LF2 A 216 Gr. WCB

37 | P a g e
In the table, there are different pressure classes. There are classes: 150, 300, and 600 etc. These class
names are not the pressure designations in psi.
A look into the table shows that the maximum pressure a class 600, A105 flange can carry at 650 F is
1075 psig. This is an upgrade on the design pressure of 1060 psig.

How did we arrive at class 600? Look in the temperature column, and locate the temperature (650 F in
our case). Move along that row until you see a pressure equal to or immediately greater than your
design pressure (1075 psig is immediately greater than 1060 psig). Look up to locate the pressure class.
That class is the appropriate class. Material A105 meanwhile is a forged carbon steel. It does not have
different grades and was normalized after the forging process to eliminate internal stresses. It is
sometimes designated as ASTM A105N. The N represents normalized.
The analyst should note that as the design temperature increases, the pressure containing ability of a
particular class decreases. Also, as the quality (elastic modulus) of the material drops, the pressure
containing capacity drops.
Leaks can develop in a flanged joint if appropriate temperature-pressure rating is not done because this
is the weakest link in a piping system.
The flange material we used throughout this book is ASTM A105.

Valves are of many types. They are used for fluid regulation and throttling purposes. The uses of each
type of valve will not be discussed here as the only reason a stress analyst needs them is to incorporate
their masses into the system. For large valve bodies, the cast material, A 216 Gr. WCB is used.

Reducer
The reducer is a pipe component that changes the diameter of a pipe. It enables two different pipe sizes
to be joined. The reducer is of two types. They are:
i. Concentric ii.
Eccentric
The concentric reducer is the reducer type that maintains the same centre line of flow. It is used to
maintain a particular fluid velocity. It thus promotes steady flow. This reducer type should only be used
in all pipe sections where steady flow is needed.
Eccentric reducers are reducer types with an offset in the y-direction (vertical direction). These reducers
do not maintain the centreline of flow. They rather offset the system’s centreline and consequently the
flow.
For any reducer, due to the small change in the centres of the two connected pipe sizes, a small change
in velocity is given birth to and consequently a change in pressure which reduces the energy of the fluid.

38 | P a g e
To calculate the length of the reduction, the following equation is used:
Lr= 4*(Dl-Ds) where; Lr = length of reducer
Dl = diameter of large pipe
Ds =diameter of small pipe

Expansion Joint
According to the name, it is a joint type with respect to thermal cases. This is a component which helps
relieve the system of expansion problems by providing an allowance for the system to expand and
contract. This component is becoming obsolete and even when used; the system must be a low
pressure one. The expansion joint has the draw-backs that it disengages at the joint where it attaches
itself to the system in a high pressure line and it is very expensive.
In situations where it is used, other measures that can perform similar function must have been
explored and certified not good enough. It is therefore presently the system’s last resort to dealing with
thermal problems.

Expansion Loop
Expansion loop is a technique rather than a component. It is the present means of solving a thermal
problem which involves the inclusion of elbows forming a loop around the node(s) requiring expansion
solution. Expansion loops are only used if the system can tolerate minor energy losses associated with
the inclusion of elbows or where other techniques can compensate for the energy drop given by the
loop.

Stiff piping
Consider the illustration below; the two nozzles are perfectly opposite each other in the physical space.

39 | P a g e
It is required for the piping engineer to connect these two hot systems, A and B.

The engineer should not say because it will be cheaper and easier for the nozzles to be connected with a
straight run of pipe, therefore he connects them as such. What will happen is that as the system starts
operating, it will seek redress as the operating temperature rises to the design temperature, since this is
a system of no allowance (stiff system), the system will crack. The possible solutions to this include:
i. The use of expansion joint ii.
Creating an expansion loop
The first possible solution above fails if the system is a high pressure one. High pressure according to
ASME B31.3 is the pressure in excess of that allowed by ASME B16.5 class 2500 for the specified design
temperature and material group. For example, for material group 1.1, at a temperature of 3000C, a class
2500 flange cannot withstand a pressure of 5600 psig (from the table given in input parameters for
Rigids). If a flange of material 1.1 must be used for a pressure of 5600 psig and at a temperature of
3000C for example, then ASME B31.3 pressure piping section (Chapter IX of B31.3) for high-pressure
fluid service must be referred to.
ASME B16.5 is a standard which covers pressure-temperature ratings, materials, dimensions, tolerances,
marking, testing, and methods of designating openings for pipe flanges and flanged fittings.
The second option like discussed previously only fits in if the system can tolerate minor energy losses
from the inclusion of elbows. This method can still be used even when the system cannot tolerate the
minor energy losses from the elbow by doing any of these:
i. By attaching a pump
ii. By over-sizing the required internal diameter of the pipe. (Continuity says area and energy
are proportional)

SIF’s and Tees

SIFs are not components. It means stress intensification factor. It is a nominal stress multiplier due to
imperfections at intersecting geometries.

40 | P a g e
It is a term which describes the intensified stress introduced by pipe joints. It is particularly important in
expansion or secondary and fatigue or cyclic situations. SIFs do not affect primary loads.

Meanwhile, the stress concentration factor (SCF) should not be taken as SIF even though they both
amplify nominal stresses. SCF is as a result of material discontinuities e.g. notches.

Tees
These are components but in the tee database in CAESAR II, not all tee types are components. The
program considers the mode of connection of the branch pipe to the run pipe in some cases. Therefore,
not all tees the user will find in the tees section are components. The mode of connecting the branch
line to the run line or the mode of connecting one piping component to another is the single overriding
factor for the names given here. Description of some of the available options (tee types) follows:

Unreinforced

Branch pipes are sometimes connected to the run pipe directly without introducing a tee. This is what is
referred to as a fabricated tee. This is often the case with a normal fluid service and other fluid services
where the size of the branch must be less than 0.8 times the run pipe with huge consideration of cost (it
is the cheapest tee). Having said this, it is important the user knows that whenever this tee type is
selected, the branch introduced is called a stub.

A stub is a branch line connected directly to the run pipe. There are two types of stub connections. They
are:

i. Stub-in ii. Stub-

on

When the branch line is inserted a bit into the internal diameter of the run line, it is called a stub-in. This
kind of connection by virtue of the small insertion is stronger than the stub-on but obstructs fluid flow in
the run/header, whereas the stub-on is the connection type where the branch pipe is terminated at the
end of the thickness of the run. There is therefore, no flow obstruction.

Welding the branch pipe directly to the run pipe should be avoided when the branch pipe size exceeds
0.8 times the run pipe size.

Reinforced

Reinforcement is not needed when:

i. The branch connection is made using a fitting made to an ASME B16.9 standard (e.g.

41 | P a g e
 Welding tee) ii. The branch connection is made using a threaded or socket-weld
coupling not exceeding
a. Nominal Pipe Size (NPS) of 2-inches
b. One-fourth the nominal diameter of the main/run
iii. An integrally reinforced outlet fitting is used (O-let fittings)

In critical systems the weakness created by the hole on the run pipe (for the branch to be inserted) will
have to be compensated and restoration of a matching strength to the original strength is achieved by
the use of a reinforcing pad. The reinforcing pad strengthens the run.

ASME B16.9 is a standard which covers overall dimensions, tolerances, ratings, testing, and markings for
wrought factory-made butt-welding fittings in size NPS ½ through 48.

To select appropriate dimensions for the reinforcing pad, the user can handle the following as a rule:

i. The pad material must be the same as the run material ii. The
thickness should be the same as the thickness of the run pipe

O-let: any pipe component whose name ends in olet is an O-let fitting. The weldolet is the most
common of all O-let branch connections, and is welded onto the branch pipe. The ends are bevelled to
facilitate this process, and therefore the weldolet is considered a butt-weld fitting. O-lets are designed
to minimize stress intensification and provide integral reinforcements.

Having given little descriptions of the most widely used tee types in CAESAR II, it is important to take
into account the factors to consider when selecting a joint type:

1. SIF
2. Ease of assembly
3. Pipe size
4. Corrosion
5. Code

SIF: It is a huge factor to consider when selecting a joint type especially when more than one joint type
can fit a particular service and the process conditions are cyclic in nature.
Whenever the user selects a tee type, the software calculates SIF based on the code, the geometry of
the selected component and the sizes of the branch and main lines.

42 | P a g e
SIF to a large extent determines the fatigue life of a piping system. When loads are to be reduced,
especially in other to extend fatigue life, then reducing the SIF of joints should be the first interest.

Ease of assembly: How easy is it to fabricate the joint? If a tee type is easy to fabricate, the installation
labour cost will be reduced. Every business will want to make profit. If a suitably cheap alternative exist
for a tee type, it should be selected at the expense of its difficult to assemble counter-part although not
at the expense of the joint integrity.
Joint integrity is a term used in piping engineering to describe a joint that is leak-proof.

Pipe size: The size of the main line also influences the selection of tee types. Most tee types are not
permitted to be used by the code due to size restrictions. For example, threaded joints are limited to
NPS (Nominal Pipe Size) 2 and lower.

Corrosion: Corrosion is almost inevitable in process piping, but the rate can be reduced by good material
selection as mentioned previously. In the case of tee type selection, the mode of connection also plays a
part. The tee type material must be the same with that of the main line to avoid galvanic corrosion and
consequently leakage. Cast-iron is usually limited to category D water services due to corrosion
concerns.

Threaded joints are known to pose huge crevice corrosion problems for large diameter pipes and
consequently leakage in the field due to their intersection pattern which allows for fluid retention. It is
therefore also important that the analyst has an eye on corrosion in selecting an appropriate tee type.

Code and standard: This guides the stress engineer in making code-compliant choices as regards the tee
type material and the type of welding. Every other factor mentioned previously is answerable to the
code, because the code ensures integrity.

Some inputs germane to SIF’s and TEEs which greatly affect SIF are:

PAD THK

Thickness of the reinforcing pad for reinforced fabricated or full encirclement tees, intersection type #1
and #17 respectively. The pad thickness is only valid for these intersection types. We have already
discussed the dimensions of a reinforcing pad.

43 | P a g e
CROTCH R

The crotch radius of the formed lip on an extruded welding tee, intersection type 6. Basically, if the user
makes an attempt to reduce the stress riser (SIF) at a fabricated intersection, by guaranteeing that there
will be a smooth transition radius from the header to the branch pipe, then he may reduce the resulting
stress intensification by a factor of 2.0.

If this value is left blank, a value of zero will be used. This indicates no crotch, i.e. a corner.

Restraints
A pipe support or restraint is a designed element that transfers the load from the piping systems to
supporting structures, although, a restraint restricts pipe movement in all/some directions, but supports
carry the dead weight of the pipe. Restraints and supports have been used interchangeably in many
documents.
The program offers a wide a range of restraint types in its restraints database. Before describing some
types, a brief note on piping restraints.
Restraints are piping components that support the cold and hot loads of piping systems. They are good
to use but the piping engineer should not go crazy with their use. Sound knowledge of the piping system
is required before appropriate restraints can be called for. If not properly installed or used, restraints
can increase plant cost and overstress the system. It is then highly important that the piping engineer
installs his restraints at only appropriate locations where they are needed.
The above conditions make it to compulsory for the piping engineer to know the behaviour of his system
both statically and dynamically before calling for support (s).

Note: All translational support types allow for thermal expression of the system, although not in their
line (axis) of action. Also, any restraint type mentioned below can be modelled as a rigid or explicitly
given stiffness values if they are known.

Anchor (ANC): this restraint type eliminates every degree of freedom (six degrees of freedom). In the
CAESAR II’ piping analysis, it represents a nozzle. The user should use the ANC to represent nozzles of
the physical system. In the cause of solving displacement problems in stress analysis, an ANC can be
used but the user should have at the back of his mind that at nodes where ANCs are installed, the pipe
cannot freely respond to thermal stresses in all degrees of freedom.

44 | P a g e
X: This is a double acting x-restraint. It eliminates freedom in the x-axis. This support type is a
translational one. It eliminates freedom in the x-axis (both positive and negative).

Y: This is a double acting y-restraint. It eliminates freedom in the y-axis. This support type is a
translational one. It eliminates freedom in the y-axis (both positive and negative).

Z: This is a double acting z-restraint. It eliminates freedom in the z-axis. This support type is a
translational one. It eliminates freedom in the z-axis (both positive and negative).

+X: similar to the X-restraint type discussed above, but the difference is that this support type precludes
freedom in the direction opposite to that in its name (i.e. in the -x direction) and the converse applies to
the –X type.

+Y: similar to the Y-restraint type discussed above, but the difference is that this support type precludes
freedom in the direction opposite to that in its name (i.e. in the -y direction) and the converse applies to
the –Y type.

+Z: similar to the Z-restraint type discussed above, but the difference is that this support type precludes
freedom in the direction opposite to that in its name (i.e. in the -z direction) and the converse applies to
the –Z type.

The RX, RY, RZ, +RX, +RY, +RZ, -RX, -RY, and –RZ: These are all rotational support types as designated by
the “R” letter before them. They preclude freedom in their line of action designated by the letter
attached to their names but allow the piping system rotate in other axes at that particular node where
they are installed.

Guides (GUI): These are pipe supports that guide against the movement of the pipe along its minor axis.
They restrict the longitudinal movements of the piping system but allow for rotational and axial
movements.

Limits (LIM): These are supports that restrict the axial movements of the pipe but allow for free
movements in the rotational and longitudinal directions. They are double acting, while the +LIM and –
LIM are directional.

45 | P a g e
Translational double acting snubber (XSNB, YSNB, and ZSNB) AND Directional snubber (+XSNB, +YSNB,
+ZSNB, -XSNB, -YSNB, -ZSNB): Seismic design of high-temperature piping systems represents an
especially challenging exercise. Thermal expansion effects require that the piping system be flexibly
supported to allow for free thermal growth. The dynamic aspects of design usually require that the
piping system be rigidly supported. The simultaneous consideration of these diametrically opposite
requirements results in the need to use a significant number of specialized (and therefore expensive)
pipe support devices called snubbers. Snubbers lock up and carry load when subjected to the rapidly
varying vibratory loads, yet remain free to permit thermal movement of the piping system during the
relatively slow expansion or contraction caused by temperature changes.
Static snubbers may be directional, i.e. may be preceded by a plus or minus sign.

Having gotten an idea on what some of the listed restraints types are, it is high time we considered
some of the factors affecting restraint selection:
1. Corrosion: If the support will be joined to the pipe directly, it should be made up of a material
that does not cause galvanic corrosion.
2. Piping load: Of what use is the support if it fails under the load of the piping system it is
supposed to carry? This is where the stiffness of the support plays the highest role.
3. Thermal forces: Not all support types are designed to transfer hot loads. Some are designed for
cold loads. Example of a support suitable for operating or hot load is the spring hanger.
4. Occasional loads: Not all supports are suitable for dynamic events. In fact, in the case of
vibration, only the snubber can efficiently transfer such loads, while also satisfying thermal
requirements of the static analysis.

Spring hanger
The spring hanger is a pipe support, using a spring or springs to permit vertical movement in a piping
system during thermal expansion or contraction while carrying cold load.
Spring hangers should be used where expansion can cause vertical movement of a pipeline. For
example, a piping system might expand 1-nch during hot environmental conditions and contract 1-inch
during cold environmental conditions. A spring hanger is used in such conditions as the expansion and
contraction abilities of its spring can efficiently handle such conditions.
The more the load the spring hanger carries, the greater the size and the more the cost.
There are two categories of spring hanger namely:
Variable spring hanger (VSH)

46 | P a g e
Constant spring hanger (CSH)

CAESAR II has a built-in hanger sizing algorithm where if the user specifies a spring hanger at a particular
node, the program considers the overriding conditions at the node, including the extent of displacement
to select an appropriate hanger size depending on the hanger table selected.

Some Hanger sizing parameters and their meaning:

Before explaining what some hanger field means. It is important the user knows that the inputs in these
fields are not prerequisites to using the hanger selection but a good knowledge will help perfectly
describe the spring hanger of choice.

Cnode: The CNode, or connecting node number, is used only when the other end of the hanger is to be
connected to another point in the system, such as another pipe node.

Hanger table: it is not economical to design custom springs, therefore, the program has made available
a commercial catalog of spring hangers from different piping support manufacturers.
This has no significance on stress output. It is purely a commercial selection.

Operating Load
To override the operating load that CAESAR II is calculating, enter the desired value in the Operating
Load field. This value is normally entered when the user thinks that loads on a piece of equipment will
be reduced if a hanger in the vicinity of the equipment is artificially caused to carry a proportionately
larger part of the total load. This operating load is the hot load the hanger is designed to support after it
undergoes any travel due to the thermal expansion of the piping.

Free Anchor/Restraint at Node: this is an important approach towards relieving equipment nozzle of
piping loads. In situations when the piping system passes, the system boundaries (nozzles, modelled as
ANC) can fail from excessive piping loads that have been transferred from the piping system. The analyst
is bound to add some hangers to carry loads arising from process conditions. The particular nozzle to be
freed is specified here (node number). Bigger springs are sized here and increased cost results.

Free Code
Whenever an anchor or restraint should be released for the restrained weight run, that anchor’s node
number should be put in the Free Anchor/Restraint at Node field, and the Free Code describing the
directions to be released should be put in the Free Code field on the same hanger spreadsheet. Free

47 | P a g e
Codes are
1. Free the anchor or restraint in the Y direction only.
2. Free the anchor or restraint in the Y and X directions only.
3. Free the anchor or restraint in the Y and Z directions only.
4. Free all translational degrees of freedom for the anchor or restraint. (X,Y and Z)
5. Free all translational and rotational degrees of freedom for the anchor or restraint. (X, Y, Z, RX,
RY, and RZ)
The last option usually results in the highest adjacent hanger loads, but should only be used when the
horizontal distance between the hanger and the anchor is within about 4 pipe diameters.

Materials
Any material chosen here represents the material of construction for the element on that particular
spreadsheet. CAESAR II has a large material database and even allows the user to define his own
material properties if they are different from what the program offers.
Allowable stress: This is an auxiliary field. It is distributive just like most CAESAR II elements, unless the
material is changed i.e. a material on element 10-20 automatically becomes the material of subsequent
elements except manually overwritten.
Elastic Modulus (c): The elastic Modulus of the piping material is an intrinsic property of the material
which answers the design conditions. If the elastic modulus of the material is not sufficient to carry the
design conditions, a new material is called for. With the allowable stress auxiliary field active, the
program notifies the user when the design temperature goes against that allowed for such materials.
We have earlier said that the intrinsic property called Elastic Modulus is important that it is high for
temperature and pressure considerations, but the ratio of elastic modulus to density called specific
elastic modulus is more importantly required to be high with due consideration given to cold loads. It
should also be known by the user that the elastic modulus specified here is cold (i.e. it is a room
temperature elastic modulus). It varies with the design temperature.

Poisson’s ratio: The ratio of unit lateral contraction to unit axial elongation. This Pipe material property
is provided by CAESAR II.

Fluid density
The density of a fluid contributes to the weight of the piping system. It is a live load. CAESAR II assumes
that the pipe is always full of the fluid. It is another very important field to enter for accurate analysis.
The value varies from one fluid to another and it’s dependent on process conditions.

48 | P a g e
Pipe density
The density of the every pipe material is required low, but it is often difficult to find a material with high
mechanical properties and low density (FRP is an exception). It is the ratio of mass to volume.

Insulation density
For systems with insulation, it increases the dead weight of the piping and decreases extensive
mechanical properties which lowers natural frequency and increases the ease of dynamic excitation. If
the system has insulation, the program must be aware.

Code
We will make use of B31.3 which is the code and standard governing process pipes.

THERMAL STRESSES

49 | P a g e
At the end of this example, the user should:
i. Remodel this system with minimum help ii.
Learn how to troubleshoot pipe expansion problems

50 | P a g e
 iii. Learn how to model expansion loops iv.
Proffer more thermal relief techniques

The following CAESAR II piping input-images and notes will model the piping system shown above.

Open CAESAR II click on file>new or click . Make sure your selection is on piping input and enter
any name in the Enter the name for the NEW job file .i.e.

Accept the unit system. CAESAR II by default works with the English unit system.

51 | P a g e
The piping system we are about to model and analyze is in the mm (millimetre) system. For convenience

in the modelling process, close the piping input screen and click on (configure)>database
definition>units file name and choose the mm.

52 | P a g e
Exit with save and click on you will be returned to the piping input spreadsheet.

If you have been following along this book from chapter one, you will be familiar with the terms on the
input screen below, else, kindly go back so you can follow along.

Also note that we started from the bottom right ANC labelled A (The user can start anywhere on the
model).

Make the element 10-20 spreadsheet look like the one below:

Pipe sizes are by default measured in inches, therefore, the user can type “10” in the diameter field and
the program will do the conversion to millimetres.

The user is not expected to enter the values in the allowable stress column.

CAESAR II calculates and fills the cells.

We used ANC to denote nozzles. At every location where there is a nozzle, use ANC restraint type. This
is so because a nozzle eliminates every degree of freedom of every adjoining pipe node and the ANC
restraint type also does same.

To enter the ANC restraint type, double click on restraints and choose ANC from the drop down.

53 | P a g e
54 | P a g e
The following screenshot shows how the spreadsheet 10-20 should look after inputting the diameter,
which is 10-in for the line where we started from and typing S and hitting the enter button on the
wt/sch field and selecting the material which is A106 B (pronounced as A106 Grade B). The pressure and
temperature (process conditions) are not left out.

55 | P a g e
17 Bar = 1700 kPa

Next :

To enter a rigid (the next element is a flange), on the title menu, click on model>valve>on the rigid type
column choose flange>on the end type column choose flg>on the class column choose 300 (remember,
this class was chosen considering the material group, design temperature and pressure). After doing
these, the program will fill the appropriate cells based on the selection made above.

Please refer to the chapter on flanges, if you are having any difficulty in understanding why we chose this
class.

The user should also change the material to ASTM A105 on every rigid spreadsheet and don’t forget to
re-change it if the next element is not a rigid .i.e.

56 | P a g e
 Next :

With the same procedure as above for a rigid, the only difference here is rigid type>gate valve

Next :

57 | P a g e
Flange: click on model>valve>on the rigid type column choose flange>on the end type column choose
flg>on the class column choose 300.

Next :

The length of 704.350 mm specified here includes the length of the elbow.

By inputting the elbow on the same spreadsheet as the pipe element tells CAESAR II that the pipe
element’s actual length is 704.350 minus elbow length (381 mm, as extracted from the pipe chart or
calculated from the formula given in the bend length calculation).

704.350 mm=825 mm-120.650 mm (flange length)

The user should also be aware of how the bend definition works. The suction length of the elbow goes in
the same direction as the element where the bend was attached. The discharge length goes in the same
direction of the succeeding element to the bend.

Notice how the program has added a radius of 381mm which is a long radius elbow corresponding to 1.5
times 254 mm (nominal pipe diameter). The actual diameter of the pipe is 273.05 mm which
corresponds to a size of 10.75-in. The nominal size is 254 mm or 10-in.

The value can be overwritten for a value of 254 mm (short radius elbow length) corresponding to the
nominal size of the pipe.

58 | P a g e
We will continue with the default of 381 mm because the piping is a category M one.

Next :

Next :

The length of the reducer was calculated from Lr= 4*(Dl-Ds). Please review the piping input chapter on
reducers for clarification.

Next :

The user was given the licence to select a fitting branch connection type. For now, let’s use the
fabricated tee-unreinforced considering cost at first.

Notice that the tee was placed at Node 90. This is to centre the branch line on the element 80-100.

For the sake of the tee, we included a node in between the straight run. This run should have had the
node number of 80-90 and length of 504 mm without the tee.

59 | P a g e
 Next :

Next :

Next :

Next :

60 | P a g e
 Next :

Next :

Next :

61 | P a g e
 Next :

Next :

Remember we have a tee without the three legs complete at node 90.

At this stage, we are telling CAESAR II, that instead of continuing the numbering (i.e. 170-180), we want
a complete definition of the tee at node 90. The following spreadsheet will create that long vertical
branch line.

Next :

62 | P a g e
Analysis of the system
The code used is ASME B31.3, because the piping is a chemical piping and an above ground one.

With the analysis initiated, the stress output from CAESAR II tells:

Click on view reports after selecting (SUS) W+P1 and Stresses.

The sustained code stress on the piping system is 52915.8 kPa with an allowable of 118203.7 kPa.

The first case is the operating case. If you remember, the stress engineer is only concerned, although at
this stage with the failure cases (sustained and expansion). The sustained stress report indicates that the
system is in order.

The percentage of the total code stress with regards to the sustained code allowable stress is 44.77 at
Node 170. The implication of this value points to the fact that if there was a chance of failure under
increased primary loads, the node 170 is susceptible to plastic deformation (failure).

The stress types with their highest values at particular nodes are also displayed.

Therefore, primary loads specified on this job (weight and pressure) have not raised any cause for
concern.

63 | P a g e
In the same manner, select the (EXP) L3=L1-L2 and Stresses and view reports.

The expansion case reports there is failure in the piping system. The code stress is 529245.0 kPa with an
allowable of 282078.3 kPa. The percentage code stress is 187.62 at Node 90. Obviously, the stress
permitted by the code has been topped.

Possible solutions are on the 80-90 spreadsheet.

Possible solutions
The failure occurs at node 90 where we have an unreinforced tee. Remember, the unreinforced tee
weakens the main line and introduces a greater load on the stub.

Let’s investigate the SIF of the tee type. Close the stress reports. Click on .

We are investigating the SIF of the tee because node 90 is a pipe joint. Recall that SIF plays a huge role in
the integrity of a pipe component in the expansion and fatigue cases.

If the piping input spreadsheet for the job is not open after closing the stress reports. Click on The

following screen results:

Recall that node 90 is the tee node .i.e. the tee centre.

The program shows the SIF for the tee type and the interacting pipe sizes as follows:

Header (i) = 4.625

Header (o) = 5.833

Branch (i) = 4.625

Branch (o) = 5.833

64 | P a g e
Looking at the SIF column, the user will discover that the SIF at node 90 is highest from the expansion
stress report earlier checked. Let’s reduce these values and see what good it will do to our piping
system.

Solution a: Let’s add a reinforcing pad to compensate for the weakness in the main line due to
the hole cut in it.

Recall, as a rule. The user should do the following in selecting a reinforcing pad:

i. The pad material must be the same as the run material ii. The
thickness should be the same as the thickness of the run pipe

Appropriate reinforcing thickness is chosen with a thickness of 9.5250 mm (same as main line).
Recalculating the SIF, the following SIFs are now obtained:

Header (i) = 2.476

Header (o) = 2.968

Branch (i) = 2.476

Branch (o) = 2.968

The introduction of a reinforcing pad shows an almost 50% drop in the value of SIF.

Click the cancel button and respond yes to the dialog asking whether you wish to transfer the changes
to the spreadsheet.

Re-run the analysis just like before:

The failure has left node 90 to node 190.

Let’s see if there are other solutions to the piping failure at node 90.

65 | P a g e
 Solution b:

Investigating the expansion displacement report shows a huge value of -9.166 mm displacement in the
Y-direction and a -9.559 mm displacement in the Z-direction at node 90. This is an indication of high
forces at the failing node (90).

Let’s see what the global force report tells. It shows an excessively high downward Y force of 136098 N
acting on the stub. This force is off from the consistent force of 68049 N experienced at preceding
nodes. Let us counter this displacement force by the addition of a positive Y support.

The addition of a positive Y support proves a solution too. The failure has also moved to node 190.

Going back to the displacement report, a zero displacement is seen at node 90. The global force report
also shows an increase in the magnitude of the force (-191473 N in the Y-direction). This support has not
reduced the force at that node but has eliminated the displacement in the Y-direction.

Solution c: A weldolet shows the following reduced SIFs:

With the same procedure used in adding a reinforcing pad, add a weldolet from the drop down of tee
types without a reinforcing pad. Remember, reinforcement is not needed when any o-let fitting is used.
Click the calculate button and notice the drop in SIF.

Header (i) = 2.632

Header (o) = 2.632

Branch (i) = 2.632

Branch (o) = 2.632

Solution d: similarly a welding tee shows the following SIFs:

Header (i) = 2.308

Header (o) = 2.744

66 | P a g e
Branch (i) = 2.308

Branch (o) = 2.744

Solution e: A sweepolet also shows the following SIFs:

Header (i) = 2.308

Header (o) = 2.744

Branch (i) = 2.308

Branch (o) = 2.744

Solution f: The other solution to the failing node could be the increase in the fitting (tee)
material cold elastic modulus. Material (120) A182 F21 is called from the material database on the
piping input spreadsheet with a cold elastic modulus of 2.1098E+008 kPa.

The failure again moves to node 190 (just like the previously discussed possible solutions).

Even though at this stage, our piping system has being solved at node 90, the system still reports Failure
at Node 190. The user can also consider other solutions that will make the element at node 90 pass. The
solutions above can be combined to achieve better results if needed.

We will continue solving our system due to the failure at Node 190 but will have to proceed with the
best possible solution from the above.

Comment
We know that the cause of failure at node 90 is due to thermal (secondary) stress because the failure
has occurred in the expansion stress category. This failure is not due to a restraint, so over-restraint
cannot be held responsible for the failure at that node (90). We started with an unreinforced tee
because it is the cheapest branch connection type but as we saw, this failed the system at node 90.

The o-let fittings show better reductions in SIF, they are relatively more expensive compared to the
welding tee and the reinforced tee.

The problem at node 90 is not the problem of imposed restraint but SIF, so just like is evident in the
global force report after the inclusion of a +Y support, the use of a +Y support only counters the large

67 | P a g e
vertical displacement but a high force still resides in the stub as a result of SIF. This solution is a
temporary one and should not be used for long term structural assurance of the system. The
unreinforced tee weakens the main pipe and subjects the stub to increased stress as shown by the
global force report.

The solution f is only used in the case where the tee made of that material is available, but the material
must be one that will not incur galvanic corrosion. The interface where this component meets the pipe
might also be subject to thermal cracking due to the difference in thermal response.

Resolving Node 190

We proceed to node 190 using solution d. At this node we have a restraint. Also notice that this is not
the only failing node but this node gave the highest bending stress (code stress) due to temperature
with a value of 443979.2 kPa whereas, the code defined allowable is 277438.1 kPa.

The analyst should have an idea why failure occurred at these identical nodes (with ANC). The secondary
load acting on the system requires that the system be able to redress, else failure will occur.
In this case, the ANC at node 190 has restricted the free response of the system resulting in failure.

It will be a very poor suggestion to eliminate the ANC (nozzles). Re-routing the system could be a
solution, but our solutions does not require that (our assumption).

All solutions are on the 180-190 spreadsheet.

Solution 1a (use of an expansion loop Y-direction):

Modelling an expansion Loop (on element 180-190)

Click on

68 | P a g e
69 | P a g e
70 | P a g e
The increase in the vector lengths came from 400 mm + 381 mm (elbow length from the pipe chart)

The use of an expansion loop in the element 180-190 shows that the excess thermal stress has been
absorbed by the elbows. Elbows are highly flexible.

This loop points in the Y-direction and by its inclusion the system now boasts a code stress of 240211.6
kPa with a code allowable stress of 256524.6 kPa. The percentage code stress here is 93.64

Solution 1b (expansion loop X-direction)

We use an expansion loop again, this time, pointing in the -X-direction. The procedure is the same as
outlined above with the exception being that the element 186-187 should have its length in the –
Xdirection and element 188-189 in the +X-direction. The piping system has a code stress of 245505.0
kPa with an allowable of 256900.4 kPa. The percentage code stress here is 95.56. Solution 1c
(expansion loop, reducing the length of the last run)

71 | P a g e
From solution 1a, by reducing the length of element 189-190 from 781 mm to 581 mm, and adding the
difference of 200 mm to the element 187-188 (from 781 mm to 981 mm), a considerable drop in the
stress level was obtained. The code stress is now 236302.3 kPa with an allowable of 255612.8 kPa. The
percentage code stress is now 92.45.

Solution 1d (expansion joint)

Let’s call a finite length expansion joint with an axial stiffness of 20000 N.m./deg and a transverse
stiffness of 7500 N.m./deg.

The expansion stress output shows a code expansion stress of 230412.4 kPa with an allowable of
255962.7 kPa. The percentage code stress is 90.02.

Comment:
All the possible solutions explored absorbed the excess thermal stress on the straight run 180-190 in the
piping system under review.

The use of an expansion joint gave the least percentage code stress (90.02), it becomes the best
solution only if the known drawbacks can be effectively managed in the system. The expansion

72 | P a g e
 joint cannot withstand cycling conditions; neither can it be used in a high pressure service. It is
also very expensive.
The system cannot be regarded as a high pressure one because the flange class (300) we used at
6500F permits 535 psi of pressure against the 246 psi we require. The highest class of 2500 at
that temperature of 6500F has its maximum pressure as 4475 psi. Remember, according to the
code, it is a high pressure system if our design pressure was greater than 4475 psi for this
particular system.
The other solution that came close to the result obtained with the expansion joint is the solution
1c. This solution has the drawback that because it is a loop introduced by elbows, there will be
minor energy losses associated with the system which may not be tolerated. It requires the use
of extra space which may not be available. It could also require extra supporting structures
which can demand re-routing and consequently a change in the system which may not be
convenient. It has the advantages that it is not susceptible to fatigue failure and it is not
expensive. Also, the direction of the loop contributed little advantage stress wise.
If the high cost of the expansion joint is not a design consideration, and cyclic conditions are not
present i.e. no temperature or pressure fluctuations, the solution 1d should be used, otherwise,
solution 1c should be used but with an oversize pipe (if minor energy losses cannot be tolerate).

In this example, we have shown that secondary stresses develop in a piping system not just due to
overrestraint, but also SIF due to improper selection of tees.

With the appropriate expansion solution(s) chosen, the system should be re-analyzed not just for the
expansion case, but for the sustained case too.

73 | P a g e
SUSTAINED STRESSES

74 | P a g e
At the end of this example, the user should:
i. Get used to the modelling process
ii. Learn how to troubleshoot pipe sustained problems iii. Proffer more
primary stress relief methods

Just like the previous example, the modelling procedure is presented below:

In the diameter field, type “6” and press the “enter” button, CAESAR II converts the 6” NPS to its actual
size in millimetres. A temperature of 200 C (392 F) is entered and a pressure of 8763 kPa is also entered,
which are the system’s operating conditions.

A schedule of 5S was initially specified for the pipes. The stress analyst is required to make appropriate
sizing with regards to thickness. It has been assumed that the system does not require an allowance for
corrosion, neither does it require insulation.

We started from the flange before the element with a length of 750 mm.

Just like before, accurate flange rating is performed by looking into table 1 for a class with equal or
higher pressure than the design pressure specified, at the design temperature specified and for the
material group also specified.

Class 600 is the appropriate class for our temperature and pressure.

On the title menu, click on model>valve>on the rigid type column choose flange>on the end type
column choose flg>on the class column choose 600, after doing these, the program will fill the
appropriate cells based on the selection made.

Move the value to the Y field if it is not there already. The screenshot below shows what is required.

The restraint is an ANC representing the fixed support (Nozzle) attached to the flange in the drawing
above.

75 | P a g e
 Next :

76 | P a g e
 Next :

Next :

77 | P a g e
 Next :

Next :

78 | P a g e
Discover that the straight vertical run is 8000 mm but our input here is 8458 mm. This is because, by
attaching bends to the ends of the element, the length would have been reduced by 458 mm (twice the
length of a 6” NPS long radius elbow .i.e. 8000-458 = 7542 mm). To avoid this, we added the 458 mm to
the straight vertical run .i.e. element 60-70 so 8458-458 will preserve our 8000 mm run.

The 458 mm is as a result of half the preceding elbow length (suction) and half the succeeding elbow
length (discharge).

Next :

Next :

79 | P a g e
 Next :

With the system modelled, we move to analysis.

80 | P a g e
Analysis of the system
The expansion case reports that the system is in order. The code stress is 22348.2 kPa with an allowable
of 206842.7 kPa. The highest percentage code stress is 10.80 at Node 59. This shows that the chance of
failing under temperature is just over 1 out of 10.

The sustained case however reports that the system does fail. The code stress here is 585719.9 kPa with
an allowable of 137895.1 kPa. The highest percentage code stress is 424.76 at Node 80.

Remember, for failure in the sustained case, two primary factors are responsible. They are pressure and
weight. The highest stress types in this case are also displayed:

BENDING STRESS: 462789.9 kPa at Node 80

TORSIONAL STRESS: 37094.7 kPa at Node 40

AXIAL STRESS: 129363.7 kPa at Node 68

HOOP STRESS: 257543.3 kPa at Node 28

Comparing these stress values to the allowable shows that although bending stress contributed the
highest to the code stress value, and as such is the chief stress that has failed the system, hoop stress
also has a value that transcends the allowable.

Notice that the value of the code stress is greater than the value of the highest bending stress, and as
such has another contributor which arises from the second in command here, hoop stress.

The failure of the system due to hoop stress is an indication that the specified schedule (thickness) is
insufficient to contain the design pressure.

Let’s make an upgrade on the schedule so the hoop stress can be reduced below the allowable
stress.

On element 10-20, we clear the value in the wt/sch field and type “20” (so the new thickness will apply
to all other elements). The program displays an error message meaning that the schedule does not exist
for the specified pipe size.

So we type 40 and press the enter button. Notice that the system still fails after running the analysis due
to bending at node 80 but the hoop stress has gone below the allowable and there is a significant
decrease in the bending and code stress values.

81 | P a g e
Now, what is causing this huge bending stress in the system? Pressure has been contained by thickness,
although without achieving equilibrium, which is why there is a small difference in the code and bending
stresses.

The code stress is 384701.6 kPa with an allowable of 137895.1 kPa.

Since hoop stress has been eliminated; the other factor that comes in is weight. Weight shows
pronounced negative effect without adequate number of supports or inadequate support spacing.

Bending stresses produce displacement from equilibrium positions. Now, looking at the displacement
report at Node 80 shows a very small displacement at that node but very high displacements at the
preceding nodes.

The question now is: why will the highest percentage code stress be at node 80 with a small
displacement?

Remember, the sustained case requires that every element be rigid, but the expansion case requires that
elements be flexible. We are required to apply caution in making elements rigid.

The elbow at node 80 has been overstressed by attached stressed elements. The straight runs are
stronger and hence were able to transmit these stresses (without failing) to node 80 which has a very
flexible element (bend) and consequently failed.

The solution here can follow two ways. They are:

i. Make the node 80 element very stiff so it can withstand the high bending stresses ii.
Support all bends that goes to a horizontal change in direction with a +Y restraint, which
should greatly reduce the magnitude of stress transferred to the elbow at node 80.

Solution one above might solve the problem only if the system has zero chance of dynamic excitation
and if cost is a huge design consideration. This is only a temporary treatment.

For now, we will assume that our system has no chance of being excited dynamically and that cost is a
huge design consideration.

Calculating the SIF of the bend in review shows that the bend has an SIF greater than 1 (i.e. 2.021 for
inplane and 1.684 for out-plane).

82 | P a g e
Click on

This is also an indication of high flexibility. The addition of double flanges in the bend type shows a
reduction in the SIF values of 1.274 for in-plane and 1.061 for out-plane. Notice also the drop in the
flexibility factors.

83 | P a g e
Click the cancel button and respond yes to the resulting dialog.

The addition of the double flanges which reduced bend SIF has not prevented failure at node 80 which is
in agreement with our earlier discussions on SIF that it is particularly important in secondary and peak
stress conditions, albeit, a small decrease in the code stress was experienced. The code stress shows a
value of 273706.3 kPa with an allowable of 137895.1 kPa which is a significant decrease in the initial
code stress.

84 | P a g e
We add a positive Y restraint at node 60 which will carry some of the sustained weight while making our
bend at node 80 stiffer.

The failure has moved to Node 28, which is a bend node again. To solve the problem at this node, the
analyst will have to go through the steps we used in resolving node 80, but failure might move to
another elbow node and the similar procedure taken and so on. Before complete solution will be
achieved using this method, the analyst must have spent a whole lot of time resolving stresses and as
well as incur huge and unnecessary cost which is a design consideration to begin with.

Let’s support the piping at every horizontal change in direction and see what good it offers.

A positive Y support is called at nodes 40, 50 and 60 and the system re-analyzed.

85 | P a g e
The sustained case now reports that the system is structurally ok. The code stress here is 59709.0 kPa
with an allowable of 137895.1 kPa. The highest percentage code stress is 43.30 at node 28.

Recall that the elbow at 80 is still having the double flanges. Significant savings might be achieved if the
system will remain ok without these flanges attached at the elbow ends.

Eliminating the double flanges result in a code stress of 64914.1 kPa with an allowable of 137895.1 kPa.
The highest percentage code stress is 47.08 at node 28. This is an increase of 4.5% in code stress as
compared to the code stress due to the added stiffness provided by the double flanges at node 30 and
80.

Since the system remained ok after the elimination of these flanges, their use is unnecessary. Also
eliminating the support at node 50 could also result in extra savings.

Let’s see. Eliminate the positive Y support at node 50. To obtain the screen below (just like before),
double click on the restraint checkbox.

86 | P a g e
The code stress after running the analysis again shows that the code stress is 86028.5 kPa with an
allowable of 137895.1 kPa and the highest percentage code stress is 62.39 at node 60. This is an
indication that the distance between the present two positive Y supports (support spacing) agrees with
code defined permissible value. This again shows an increase in the code stress of over 15%, but
significant savings have been achieved while maintaining structurally sustained integrity.

Comment
Remember, we have only considered weight, temperature and pressure. Statically, our system is ok, but
we are not sure of its dynamic integrity.

In the wt/sch field, we could have specified any schedule greater than the 40 we used, but we are
only interested in that schedule which will give the least thickness that will contain pressure
called, the minimum thickness.
Even though schedule 20 may have represented a better choice than the schedule 40 used in terms
of weight and cost, it is not available and cannot be considered.

This will not only save us from unnecessary cost, it will give us a reasonable weight to worry about.

We assumed that there is no corrosion, whether due to the environment or due to the fluid and
that insulation is not needed. If these two fields were given a non-zero value just like in example
1, the dead weight of the system would have been increased, which means a drop in extensive
mechanical properties and maybe we would have done more than we did to solve the stress
problems.
87 | P a g e

THERMAL FATIGUE

At the end of this example, the user should:

i. Learn how to model a 450 and 1800 elbow

88 | P a g e
 ii. Understand why fatigue poses a unique problem and learn how to design against it
iii. Understand the risk posed by SIF in cyclic conditions iv. Predict fatigue life
and remaining fatigue life of a piping system

Just like the previous examples, the following system is modelled with the same procedure.

The user will notice that even without a branch line like the example on thermal stresses, our node
numbering is not ordered. With this example, we show that node numbering must not be ordered to
produce correct models. The user is advised to remodel this system in line with the previous examples and
follow our modelling approach here to compare the two. We started on the element with length of 1’6
going in the positive X direction. For convenience just like we did in the example on thermal stresses,
change the unit to inches.

Click on and read from file. Select TD12ST.FAT and the screenshot below
results.

89 | P a g e
 Next :

90 | P a g e
 Next :

91 | P a g e
 Next :

Next :

92 | P a g e
 Next :

Next :

93 | P a g e
 Next :

Next :

Next :

94 | P a g e
 Error check the model
On the title menu, click on analysis > statics > click Recommend and set out the table like the
screen below:

Here we have assumed that the system will experience 100000 load cycles throughout it design life.

Run the analysis

95 | P a g e
Fatigue as you already know can be accelerated by several factors including decreasing temperature,
cracks due to discontinuities, increasing stress etc.

The example under review shows the effect of temperature fluctuations (cyclic loading) on the system.
It is also important the user knows that pressure fluctuations are also a fatigue concern.

The specified design temperatures are 500 F, 500F, and 450 F. the repetition of the 500 F is to enable us
investigate that temperature to know if there is a thermal failure when acting under the expansion case.

With a code stress of 138662.2 kPa and an allowable of 313720 kPa, the expansion case reports the
temperature of 500 F is not a problem to the system. The highest percentage code stress is 44.2 at Node
90.

Notice how the 3D max intensity value is picked up and displayed as the code stress. The usual
code stress on its own does not fully take into account the effects of SIF in cyclic situations. The
maximum 3D intensity is the intensified stress often referred to as a text book stress which fully
takes into account loads and SIFs that would have been dismissed by the code stress as
insignificant. This intensified stress value is important for use in fatigue evaluation because the
insignificant loads and SIF could greatly affect piping fatigue life.

The analyst might think that his system is in order and has a no chance of fatigue failure when the
individual FAT cases are selected singly with the usage factor, but as rightly pointed out on the chapter
where we discussed fatigue; low cycle fatigue is a combination effect that can only be appraised by the
cumulative usage ratio.

If it does not add up, it is not cumulative!

All the fatigue cases are selected at the same time (two in our case), the cumulative usage report
selected and reports viewed.

The report shows a possible fatigue failure as against when the FAT cases where selected singly.

96 | P a g e
The highest usage ratio is 1.11 at Node 90. The usage ratio shows a possible fatigue failure at Node 90
with the combination of these two fluctuating temperatures.

One very interesting thing here is the fact that failure has occurred with loads that if acting alone, will
not cause failure (under the expansion case, the 500 F temperature did not cause failure). It brings to
the reminder of the analyst that fatigue fails the system while the piping material might still be in the
elastic region.

Notice again how CAESAR II recommends an allowable number of cycles at every node. The value of
infinity in some nodes tells that the highest calculated intensified stress at that Node is below the
endurance limit of the S-N curve earlier entered.

In some documents, the stress level below which a material will withstand an infinite number of load
cycles without failing is called the fatigue limit while the stress level under which the material can
withstand a very high number of cycles without failing is called the endurance limit.

We have used the two terms to represent each other in this book because they are practically the same.
The difference is purely mathematical.

Remember, below the endurance limit, the material has zero chance of fatigue failure and can
consequently take an infinite number of load cycles.

Solution
At node 90, we are then required to reduce the excess load (the solution we seek) or reduce the number
of load cycles. A look at the displacement report shows that there is a nodal displacement of 2.178-
inches in the positive Y direction. This value is not a problem but points and agrees to the stress report
that the force at node 90 is high. We said the displacement is not a problem because expansion case did
not report failure. Therefore, by the standards of the code, it is permissible.

At Node 90, there is an elbow (very flexible element). Remember, we need flexibility to deal with
thermal problems. Inclusion of supporting systems at that node will not be the best as the loads
were permitted to begin with.
In the 1950s, Markl and his team of researchers discovered that SIF was a major contributor to
piping systems under cyclic conditions. This finding is the base of fatigue evaluation by our code.
We therefore investigate the bend SIFs.

97 | P a g e
 A look at the SIF column shows an SIF of 2.345 at node 90. This is a stress multiplier and as such
its value should be reduced, if possible to 1. We add double flanges to the bend element at
node 90 to make the SIF lower and consequently make the bend stiffer.

Click on on the job’s piping input spreadsheet title menu. The bend node number to be
reviewed is 90 and make the bend type double flanged.
Recalculating the SIF shows a value of 1.424 for in-plane and a value of 1.187 for out-of-plane.
Notice the drop in the flexibility factors. This is because the bend is now stiffer than it was.
Re-analyzing the system, the expansion case shows a 3D max intensity of 188050.2 kPa which was
246689 kPa. This shows a huge drop in the stress value and the highest stress has moved to
Node 100 showing that the huge stress value at node 90 was due to high SIF. By moving to node
100, the system might still report possible fatigue failure if the number of load cycles specified
(100000) is above the recommended allowable number of cycles at a stress of 1884050.2 kPa.
The two FAT cases together now report a highest usage ratio of 0.74 which is a no chance to
fatigue failure. A look at the allowable number of cycles at Node 100 shows that the node can
carry load cycles up to 225423 from the first FAT case of 500 F and a number of cycles of 340740
from the second FAT case of 450 F which agrees with our knowledge of the S-N curve, that as
the applied stress goes up, the allowable number of cycles should be dropped to avoid fatigue
failure or as the number of load cycles go up, the stress level should be dropped to avoid fatigue
failure.

Comment
We have seen that if the system has temperature fluctuations, it should be treated with the fatigue
stress type even though these temperatures are not a concern when they act alone.

The 500 F with the 100000 number of cycles didn’t cause fatigue failure alone, neither did the
450 F cause fatigue failure alone with its 100000 load cycles, but the two in combination
showed a cumulative usage ratio of 1.11 at node 90. This is because the load at this node does
not permit such high number of load cycles. The addition of the double flanges reduced the
stress by reducing the SIF. Recall that in the chapters leading to this, we said SIF is hugely
responsible for fatigue failures of process lines. When fatigue failure occurs in a new piping
system, especially at a pipe component, SIF is suspect.

98 | P a g e
The system was already statically balanced, but fatigue required that those permissible high loads at
node 90 be reduced to a value that can carry 100000 load cycles without failing.

The user should also know that this result is only peculiar with the fatigue data entered. Different
materials have different fatigue data and consequently different results. The presence of
Notches reduces fatigue life. Even the same materials of different surfaces produce different S-
N curves.
Fatigue failure can arise from different areas as previously mentioned, but we have only
considered how to design against that due to a cyclic condition, but other factors can cause
fatigue failure if not checked even after fatigue analysis like this. Intensified stress is SIF
multiplied by actual stress.

It is best to perform fatigue test on the piping components with the intended operating conditions.

If an ASME fatigue curve for a particular material is to be used like we did, then the usage ratio due to
any particular stress intensity should be reduced as much as possible to account for likely difference in
surfaces which will act as stress risers. All ASME fatigue data are based on polished surfaces.

99 | P a g e
PUMP NOZZLE LOADS

At the end of this example, the user should:

100 | P a g e
 i. Learn how to relief Nozzles of piping loads
ii. Optimize the use of supports by using accurate support spacing iii. Study API
610 for centrifugal pumps

Next :

101 | P a g e
 Next :

Next :

Next :

102 | P a g e
 Next :

Next :

Next :

Next :

103 | P a g e
 Next :

Next :

Next :

104 | P a g e
 Next :

Next :

Next :

105 | P a g e
 Next :

Next :

Next :

106 | P a g e
 Next :

Next :

Next :

107 | P a g e
 Next :

Next :

Next :

108 | P a g e
 Next :

Next :

Next :

109 | P a g e
 Next :

Next :

Next :

110 | P a g e
 Next :

Next :

111 | P a g e
 Next :

Next :

Next :

112 | P a g e
 Next :

Next :

Next :

113 | P a g e
 Next :

Next :

114 | P a g e
 Next :

Next :

Next :

115 | P a g e
 Next :

Next :

Next :

116 | P a g e
In this example, the piping is not the concern but the connected pumps discharge nozzles. When the
load on a pump is high, the following problems result:

• Excessive equipment vibration

• Premature bearing failure of the pump and driver
• Cavitation damage to the impeller and inlet of the casing, and possibly the
discharge

• Leakage at the flanges etc

The hot and cold loads on the pipe and its connected fittings are below limits defined by ASME B31.3.
The system is structurally ok? It is wrong for a stress engineer to say a system is ok, considering the
pipe and its connected fittings alone without the connected equipments.

The nozzles of equipments also serve as pipe supports, therefore they also form the piping system and
the effects on the nozzles from any connected piping should also be monitored.

Since the connected pumps are all centrifugal pumps, API 601 code investigates the stress state of the
pump nozzles.

From the analysis, we see that the sustained case has a code stress of 48064.6 kPa, an allowable of
137895.1 kPa and a highest percentage code stress of 34.86 at node 410 and the expansion case, has a
code stress of 36157.2 kPa, an allowable of 322722.9 kPa and a highest percentage code stress of 11.20
at node 60.

The piping is ok, but the API 610 code reports a problem with the discharge nozzles of the centrifugal
pump.

117 | P a g e
To check centrifugal pump nozzle loads, go to analysis and click on API 610. It says job does not exist,
create new? Respond yes. Give the job a description and click on the input data tab and fill out the
form as below.

To get CAESAR II fill the forces and moments cells. Click on the get loads from output file select the job
name and open. After that, select the OPE case and click ok.

118 | P a g e
This is also the responsibility of the stress engineer to ensure that the connected equipment nozzles are
not failing as a result of the transfer of piping loads to such nozzles.

The code will not be reviewed in this book, but it is a necessary code for the stress engineer to be
familiar with. We will only apply it to show pass or fail conditions of the nozzles.

It is now our responsibility to reduce the nozzle loads. The stress engineer will have to know whether
the nozzle failure has occurred with the primary sustained loads or thermal loads. It is evident from the
operating and sustained cases restraint summary that the system developed very small thermal loads
when in operation but the sustained load is very comparable with the operating load (5378 N is close to
6604 N). From visual interpretation, it is clear that the sustained loads are responsible for the pump
failure. The expansion and sustained loads are -1225 N and -5378 N respectively. The total operating
load at node 250 is -6604 N; all are in the Y direction. These loads have caused failure at the pump
nozzles. We will go ahead to reduce these hot and cold loads at that node (250), but when in design
mood, the sustained loads should be the first ones to be reduced because they are very high; the hot
loads developed during operation is small. The user should be aware that the operating case is a

119 | P a g e
combination case. The sustained and thermal loads simultaneously affect the pump nozzle that was
why we used the operating loads during the pump analysis.

At this stage, the user should be familiar with some piping techniques to reduce hot and cold loads.

This system also has a standard schedule which may not be the least schedule to contain pressure. If the
system’s schedule has been excessively specified, it could be reduced in order to increase flexibility.

Working with the given schedule

Here we have assumed that for some reasons, we must work with the given schedule.

By adding expansion loops in the horizontal and vertical straight runs leading up to the nozzle at
node 250.

120 | P a g e
 Creating the other expansion loop on the element leading directly to the pump nozzle:

We reduce the length of the last leg of the expansion loop leading to the valve assembly by 300 mm
and add that value (300 mm) to the vertical run of the loop (we have known that the closer the loop is
to the ANC (nozzle node), the more its effectiveness from the example on thermal stresses).

121 | P a g e
122 | P a g e
The model now looks:

We added the second expansion loop just to exaggerate our point that the expansion stresses developed
is not the cause of the pump failure.

This shows a reduction in the nozzle loads but the pump’s status still reports a FAILURE.

We add a spring hanger at node 190 (first bend element from the pump nozzle) with an
operating load of 6000 N to be carried as shown below to take some loads off that nozzle node
(250):

123 | P a g e
With the spring hanger added, the pump overall status remained FAIL, even though we have reduced
the loads due to thermal effects at the nozzle node (250). With the expansion loads well reduced, and
no success with the pump’s status, we investigate the sustained loads, which is what we should have
done at first.

A look into the sustained displacement report shows that there are high permissible (by ASME B31.3)
displacements in the nodes leading to the nozzle node under review. Looking further into the
sustained global force report shows that there are also high forces causing those displacements. These
loads should be reduced. These loads are due to weight and pressure; therefore, supports to carry the
dead weight of the piping system are called for.

Following a rather conservative approach, we attach two +Y supports each at new nodes 55 and
85.

To create these nodes, break element 50-60 for node 55 and element 80-90 for node 85.

124 | P a g e
With these supports at those nodes, the overall pump’s status still reports pump failure. Looking back
into the sustained displacement and sustained global forces reports, we see that these two +Y supports
have done little in reducing those high loads. The element (welding tee) linking the pump’s discharge
line with the run line still does not have enough support from these two attached +Y supports. These
high loads point to inadequate support spacing. Here we are only interested in providing support to the
pump’s discharge from the run line. We therefore bring the supports closer in order to reduce the
spacing between them so as to limit the stress transferred to the discharge line from the main line.

125 | P a g e
With the analysis re-initialized, the pump’s overall status reveals a PASS.

Assuming we can resize the thickness

Here, we assume that we can suggest a structurally satisfying thickness and a relatively cheap one.

126 | P a g e
We select schedule 10 because:

i. It contained the design pressure ii.

It will reduce cost iii. It reduced
weight iv. It increased the system’s
flexibility.
The pump analysis is re-run, with no expansion loops (checking if the flexibility added by the
reduction in schedule is sufficient to make the pump pass). The overall pump’s status is FAIL.
We add expansion loops as we did before (on elements 170-180 and 200-210), re-run the
analysis, the pump is still not ok. We add a spring hanger at node 190 as before, but without
specifying an operating load for the spring to carry (this will reduce spring size and
consequently cost).

The analysis is re-initialized, with the pump’s overall status now showing a PASS.

Just like before, expansion joints can be used instead of loops if it can effectively be managed.

127 | P a g e
Comment
Piping stress analysis is not complete until connected equipment nozzle loads become permissible as
specified by their applicable codes. From this example, we see that when the piping system (without
equipment nozzles) reports it is ok; the piping loads were transferred to the system’s boundaries
(nozzles). We also see that an attempt to comply with equipments codes can significantly increase the
number of pipe supports and stress relieve techniques which will consequently increase cost.

We have also shown an added advantage of using the minimum schedule that can withstand
hoop stress which is an increase in the system’s flexibility. With this added flexibility, we didn’t
use the +Y supports we used when the schedule was standard. The user can try removing the
loops to know what good it offers since the thermal load developed during operation was
small to start with. This could mean more load to be carried by the spring hanger.
The example also included friction coefficient. Friction between supports and pipes are important
especially when there is a relative movement between the two. The value of the friction
coefficient used in this example is between two steel surfaces. Friction coefficient values affect
stress results and are a very important input in stress analysis.

We have only considered the nozzle at Node 250, leaving the other two nozzle nodes (360 and 470) for
the user to investigate and resolve if there is an overloading of the nozzles.

1. .

128 | P a g e