ME 4440-5540 Lecture 1

Introduction

 Reference book for this class: Mechanical Vibration (S. Rao) 5th edition (Prentice Hall)

Mechanical Vibration (Structural Dynamics): A broad field of engineering or applied

mechanics
Engineering mechanics: It is one of the oldest disciplines in engineering and it’s the field that
deal with the action of forces or environmental effect on a body and how that body react to
forces.
Main courses in engineering mechanic (solid):
1) Statics
2) Mechanics of material
3) Dynamics
4) Kinematics
5) Mechanical vibrations

Any field in engineering can be represent by following diagram

A B C
Output
Input (Response)
System

This is all we do in study engineering mechanics. The only differences is related to different
assumptions or the nature of those components.

Statics:
A) Forces or load (time independent)
B) Assumed system be a rigid body (particles)
C) Forces

Example:
A

B
System (Rigid body)

1
 ME 4440-5540 Lecture 1

Because in statics, we have rigid body as the system, it cannot absorb energy and cannot be
deformed, it simply transfer the whole input force to supports.

Mechanics of material:
A) Forces or load (time independent)
B) Assumed system be a flexible body (deformable) – (Elasticity)
C) Forces/ Internal Forces (shear and moment inside of a beam)/ Deformation

Strain (𝜀)
Stress (𝜎)

Dynamics (rigid bodies):

A) Forces (time dependent)
B) Assumed system be a rigid body
C) Time dependent forces (Also, velocity, acceleration, etc.)

Kinematics (rigid bodies):

A) Input motion
B) Assumed system be a rigid body
C) Output motion (position, displacement, velocity, acceleration, etc.)

Mechanical vibrations:
A) Forces (time dependent) or any other time dependent phenomena can that causes a change in
the system (e.g. displacement). The forces can be desirable (like in engines) or undesirable
(like earth quick)!
B) Assumed system be a flexible body. That means not whole the force/energy that goes to the
system doesn’t get out of the system and some part of it absorb by system (damping). A
flexible body has inertia, elasticity, and energy absorption (dissipation).
C) Forces/ Displacements/ Stresses(𝜎) & Strains (𝜀)

2
 ME 4440-5540 Lecture 2

What we need for this course

A B C
Output
Input (Response)
System
Forces/ displacement Forces/ displacement/
(time dependent) Flexible body Stresses (𝜎) & Strains (𝜀)

 The ultimate goal in Mechanical vibration course: is it possible to set up a mathematical model
that has all of these elements in it and represents the whole system?

For formulating mechanical vibration equations, we will use the second order of ordinary
differential equation (ODE).

 There are two ways for solving an ODE:

1) Laplace transform (𝑙)
2) Direct formulation/ Direct integration

𝑙 𝑙 −1 (𝑠)
𝑆
Laplace Transform

𝑦 ′′ 𝑦

Direct Integral

∫𝟎

Rather than solving a very complicated integral, in the Laplace transform method, we are first
mapped the equation to another domain and then we doing inverse mapping to solve the problem.

In this course, we are dealing with some simple second order differential equations, so we are
using direct approach for solving the problems.

1
 ME 4440-5540 Lecture 2

 We are working with complex numbers in this course.

The real numbers have two dimensions: Real dimension & Imaginary dimension

𝑍 = 𝑥 + 𝑖𝑦

𝑥 R

Some number may have the imaginary part or may not (for instance, 5 = 5 + 𝑖(0))
Complex algebra has its own operation:
1) Summation/ subtraction
If: 𝑍1 = 𝑥1 + 𝑖𝑦1 , 𝑍2 = 𝑥2 + 𝑖𝑦2
𝑍𝑡 = 𝑍1 + 𝑍2 = (𝑥1 + 𝑥2 ) + 𝑖(𝑦1 + 𝑦2 )
2) Multiplication
Note: 𝑖 = √−1 so 𝑖 2 = 𝑖. 𝑖 = √−1. √−1 = −1
𝑍𝑡 = 𝑍1 . 𝑍2 =(𝑥1 + 𝑖𝑦1 ) . (𝑥2 + 𝑖𝑦2 ) = (𝑥1 𝑥2 − 𝑦1 𝑦2 ) + 𝑖(𝑥1 𝑦2 + 𝑥2 𝑦1 )

3) Division
𝑍1
𝑍𝑡 = 𝑍𝑡 . 𝑍2 = 𝑍1 From this equality we can find 𝑍𝑡
𝑍2

𝑥1 𝑥2 + 𝑦1 𝑦2 + 𝑖(𝑦1 𝑥2 + 𝑥1 𝑦2 )
𝑍𝑡 =
𝑥1 2 + 𝑦1 2
Note: If two complex numbers are equal that means real parts of them are equal and imaginary parts
of them are equal too.
𝑍1 = 𝑍2 𝑥1 = 𝑥2 & 𝑦1 = 𝑦2

2
 ME 4440-5540 Lecture 2

There is another way to look at an analogy of real numbers with using polar coordinates.

𝑍 = 𝑟 (cos(𝜃) + sin(𝜃))
𝑦 = 𝑟 sin(𝜃)
cos ⬚

𝜃
𝑥 = 𝑟 cos(𝜃)

cos
Also, “Z” can be written in form of 𝑍 = 𝑟𝑒 𝑖𝜃 .

 Complex conjugate: Two numbers which are symmetric with respect to the real axis (they have
same real values and opposite imaginary values)

𝑍1 = 𝑥1 + 𝑖𝑦1
𝑍1 ∗ = 𝑥1 − 𝑖𝑦1

𝑦1 𝑍1

𝑥1

−𝑦1 𝑍1 ∗

3
 ME 4440-5540 Lecture 3

Mechanical Vibration
 What is the objective?
The objective of mechanical vibration is we want to analyze and understand the behavior of a
system under the action of a desirable/undesirable motion (to use this vibration in an effective
way or eliminate it from system).
In vibration we have focus on the system and that is the most important part. Also, we have to
understand how model and present the force.
 System can be simple or very complex. For the complex system, as an engineer we have
to simplify the system to be more understandable.
 Force is time dependent.

Output
Input 𝑓(𝑡) (Response)
System
Disturbance (Forces,
displacement,
acceleration, etc. (time
dependent)

System is that element with all inherent characteristics of the real physical structure. Any physical
structure has in general no more than three major inherent properties or characteristics that define
basically what any system made of or is capable of doing in order to resist the action of a complain.
These properties are: 1) Mass (m) 2) Elasticity (k) 3) Energy Absorption (c).

Mass for the inertia of the system which as the result of the action of the force moves in the certain
direction. The system or that structure has the ability to resist this motion. This resistance can
comes in two distinct ways: Elasticity, which is the structure resists by going through some
deformation, bending, reacting to the action of the force or Energy Absorption, that system or
structure resist the imposed motion or the force by trying to dissipate the effect of that action or
dissipating energy that resists that motion.

The goal of vibrations analysis, is finding the system as the most appropriate mathematical model
of a real/physical structure or mechanism.

1
 ME 4440-5540 Lecture 3

Idealization

 Idealization: 1) Making logical assumptions based on knowing the physics of the structure 2)
Defining some things that can help us to develop the proper mathematical model.

1) Disturbance (𝑓(𝑡)): This force has certain forms or a behavior

2) Motion: We are dealing with two types of motion: Translational and rotational

 What is the main difference between translational and rotational motion?

In both of them we are talking about the mass go to some motion. But in translational motion we
are look at the motion in terms of what is happening to the center of mass. The distribution of the
mass or the way the mass is distributed or the inertia are not so important in translational motion
while in rotational motion mass distribution is super important.

Translational motion: Dealing with the mass or the center of the mass or how the mass is lumped
at one point.
Rotational motion: Dealing with mass moment of inertia

3) Degree of freedom: it describes when the structure start to move how every single mass
element in that structure moves. If there is one independent deformation that the rest of
structure can be defined according to that, we will have one degree of freedom.

Example: In the following light, if we assume whole wright of structure is just concentrate in the
light (chain is weightless), then we will have one degree of freedom and we can describe motion
of everything else respect to the blob. In other word, if you get the overall motion of a structure
represented by the motion of a single point we call this one degree of freedom.

If we can assume the entire mass of a structure or system is lumped in a single point which moving
in a certain direction (can be a translational or rotational) then we will have a system with a single
degree of freedom.

2
 ME 4440-5540 Lecture 3

4) Type and shapes of 𝑓(𝑡): This force can be in various shape and various types.

a) Harmonic force: Sinusoidal cyclic force (like the engine of the car)

b) Periodic (not harmonic) force:

c) General force: 1) Deterministic: like a blast load, an impact load etc. 2) Random force:
(completely unpredictable and predictability associated with some probability) like
earthquake, wind load, etc.

Mechanical vibration can be categorized on two distinct ways.

First, based on type of force that acts on the system. So, the system can vibrating without the
force acting on it and it is called free vibration. For instance, wind has been blowing and move
the trees then wind stops but trees still vibrating and moving. Or the system is vibrating due to
the action of the force that continuously is acting on it and exist and it is called forced vibration.

The following system has one degree of freedom. This system includes a mass (m) which is
moving because of a force (f(t)) in one direction (translational). This motion (displacement) is a
function of time (u(t)). Also, we have something that is trying to stop the motion of the mass (it
is named stiffness or elasticity element (k)). We will assume this system is linear, so that
deformation would be a linear function of the displacement. We can represent that by a linear
force which can be shown by a spring. We will have another kind of resistance against the
motion which is the energy absorption and we call it damping element (c).

3
 ME 4440-5540 Lecture 4

Mechanical Vibration
In this course, we use “u” rather than “x” to be able to analyze the multi directional motion and
not make any confusion with x-y frame. Also, we assume small values for “u” and that means
the deformation of structure and material will be remain in the elastic range.

Input: 𝒇(𝒕)
System: (m, k, c)
Output/response: u(t)
All structures in the World can be can be described by the motion of a single point where the
entire mass of the structure is lumped at that point. So, we can assumed and present most
structures in the World by a single degree of freedom (S.D.O.F) model.

1
 ME 4440-5540 Lecture 4

Mechanical vibration has two main Objectives:

1) How to model a physical system as S.D.O.F. system. If we have a structure, how we can
define the stiffness element, damping element, and mass of structure.
2) How to find the response (displacement u(t), velocity 𝑢̇ , acceleration 𝑢̈ , stress 𝛿 and
strain 𝜀, etc.)

A physical system in general can be presented in one of the following forms:

1) Undamped System: System with not ability to absorb energy (or negligible)

2) Damped System: System with ability to absorb energy

Undamped Damped

Type of Vibration Type of System Type of Force

Free Vibration (Force was in Undamped System

the past but No longer any
force now)
Damped System

Harmonic
Forced Vibration (Continues Undamped System Periodic (not harmonic)
force all the time)
General Deterministic
Damped System Random

For solving the system we will use the Free Body Diagram. There two ways to set up the
free body diagram in dynamics and solving dynamic vibration problems:
1) Newton’s low
2) D' Alembert’s principle : In this method in order to set up or solve a dynamic problem,
you treat the problem as a static case by drawing the free body diagram and place the
D'Alembert’s force which is basically is an inertia force on it in opposite direction of the
motion (mass times acceleration 𝑚 𝑢̈ ). Then with this free body diagram, we can write the
equations of equilibrium which are similar to the static equilibrium.

2
 ME 4440-5540 Lecture 4

In this diagram:

𝑓𝑘 : Force of elasticity. This force is a linear function to displacement and can be find from
multiplication of displacement to some constant value (stiffness coefficient).

𝑓𝑐 : Damping force. In reality this force is a very complex and hard to calculate. To simplify
that, we are making an assumption and define this force as multiplication of velocity to some
constant value (damping coefficient).

Equations of equilibrium:

∑𝐹 = 0

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑓(𝑡) Equation of motion

System Response Input

Equation of motion: The mathematical representation form of the entire physical system,
input, and response.

Whole mechanical vibration problems will be solved by using “equation of motion” for
different condition. Our focus in this course is on how to set up model of a physical system
and convert it to the “equation of motion” , then solve the equation to find response and
find out what is the meaning of that response and what kind of information that will give
us.

3
 ME 4440-5540 Lecture 5

Undamped Free Vibration S.D.O.F System

Summary of last time:

Any physical system that with a single coordinate can describe the motion, it can be modeled as a
single degree of freedom system.
Any system can be assumed which is moving in only one direction that would be a single degree
of freedom system. This system would be include the moving mass (translation/rotational) and
motion of system resisted by two elements: stiffness element and damping element.

Stiffness

Damping

Then, we just used D' Alembert’s principle which says, you can set up the free body diagram of a
dynamic system by including a fictitious force which we call force of inertia. This force for a
translational motion would be 𝑚𝑢̈ and for rotational motion would be mass moment of inertia
multiply to rotational acceleration.

Equations of equilibrium:

∑𝐹 = 0

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑓(𝑡) Equation of motion

System Response Input

1
 ME 4440-5540 Lecture 5

The mechanical vibration consists of two important areas:

1. We need to know how to set up the S.D.O.F model of a physical system

Basically till now, we just accept that any system in the world can be model as a
combination of mass, stiffness, and damping elements but if we have a physical system
how we can model it with those parameters?
2. Solving the equation of motion

Classification of Equation of Motion:

In this class we will study:
Case 1: Undamped system (𝑪 = 𝟎): We assume the physical system has no energy absorption
capability. So, whole the force only causes the deformation (like what you studied in mechanics
of material with only difference that here the force is function of time)
1a) Undamped System-Free Vibration (𝒇(𝒕) = 𝟎): Source of vibration is not exist anymore.
1b) Undamped System-Forced Vibration (𝒇(𝒕) ≠ 𝟎): Source of vibration is still exist.
1b1) Undamped System-Forced Vibration with Harmonic force
1b2) Undamped System-Forced Vibration with Periodic force
1b3) Undamped System-Forced Vibration with General force

Case 2: Damped system (𝑪 ≠ 𝟎): We assume the physical system has energy absorption
capability.
2a) Damped System-Free Vibration (𝒇(𝒕) = 𝟎)
2b) Damped System-Forced Vibration (𝒇(𝒕) ≠ 𝟎)
2b1) Damped System-Forced Vibration with Harmonic force
2b2) Damped System-Forced Vibration with Periodic force
2b3) Damped System-Forced Vibration with General force

2
 ME 4440-5540 Lecture 5

Case 1: Undamped system (𝑪 = 𝟎):

1a) Undamped System-Free Vibration (𝒇(𝒕) = 𝟎)
First of all we need to define some characteristic of parameters which are very important in
mechanical vibration and basically they are basis to set up the S.D.O.F model of a physical
system.
Study an undamped system with free vibration give us some specific information and essential
characteristics about our system.
This system include only stiffness and mass as you can see in following schematic model (or can
be a rotational system).

Equation of motion for

𝑚𝑢̈ + 𝑘𝑢 = 0
Undamped system-free vibration

How to solve this differential equation?

𝑚𝑥 2 + 𝑘 = 0

𝑘 It has two roots, so we

𝑥1,2 = ±𝑖√𝑚
will have two solutions

Two solutions:
𝑘
𝑖√ 𝑡
𝑢1 (𝑡) = 𝑐1 𝑒 𝑚

𝑘
−𝑖√ 𝑡
𝑢2 (𝑡) = 𝑐2 𝑒 𝑚

General solution:
𝑘 𝑘
𝑖√ 𝑡 −𝑖√ 𝑡
𝑢(𝑡) = 𝑐1 𝑒 𝑚 + 𝑐2 𝑒 𝑚

3
 ME 4440-5540 Lecture 5

We can write “𝑢 (𝑡)” in the form of sin and cos (Note: 𝑒 𝑖𝜃 = 𝑐𝑜𝑠 𝜃 + 𝑖 𝑠𝑖𝑛 𝜃):

𝑘 𝑘 𝑘 𝑘
𝑢(𝑡) = 𝑐1 (cos (√ 𝑡) + 𝑖 𝑠𝑖𝑛 (√ 𝑡)) + 𝑐2 (cos (√ 𝑡) − 𝑖 𝑠𝑖𝑛 (√ 𝑡))
𝑚 𝑚 𝑚 𝑚

𝑘 𝑘
𝑢(𝑡) = (𝑐1 + 𝑐2 ) (cos (√ 𝑡)) + (𝑐1 − 𝑐2 ) 𝑖 (𝑠𝑖𝑛 (√ 𝑡))
𝑚 𝑚

Note: (𝑐1 + 𝑐2 ) & 𝑖 (𝑐1 − 𝑐2 ) both are constant values. So, we can replace them with some other
constant values.

𝑘 𝑘
𝑢(𝑡) = 𝐴1 (cos (√ 𝑡)) + 𝐴2 (𝑠𝑖𝑛 (√ 𝑡))
𝑚 𝑚

So, the displacement response for an undamped system with free vibration would be a harmonic
sinusoidal function.

The frequency of that harmonic oscillation named “Natural Frequency” of the system. Actually,
this frequency represents the natural characteristics of the system and each system in the World
has its own unique natural frequency (because each system has different mass and stiffness).

Natural Frequency: 𝑘
𝜔𝑛 = √
𝑚

General Solution: 𝑢(𝑡) = 𝐴1 cos(𝜔𝑛 𝑡) + 𝐴2 𝑠𝑖𝑛 (𝜔𝑛 𝑡)

4
 ME 4440-5540 Lecture 6

Undamped Free Vibration S.D.O.F System

For finding the values of 𝐴1 & 𝐴2 , we need two initial conditions:

𝑡=0 𝑢(0) = 𝑢0 & 𝑢̇ (0) = 𝑢̇ 0

1 0
𝑢(0) = 𝐴1 cos(0) + 𝐴2 𝑠𝑖𝑛 (0) = 𝐴1 = 𝑢0

𝑢̇ (𝑡) = −𝜔𝑛 𝐴1 sin(𝜔𝑛 𝑡) + 𝜔𝑛 𝐴2 𝑐𝑜𝑠 (𝜔𝑛 𝑡)

0 1
𝑢̇ (0) = −𝜔𝑛 𝐴1 sin(0) + 𝜔𝑛 𝐴2 𝑐𝑜𝑠 (0) = 𝜔𝑛 𝐴2 = 𝑢̇ 0

So:
𝑢̇
𝐴1 = 𝑢0 & 𝐴2 = 𝜔0
𝑛

𝑢̇ 0
Solution in terms of some 𝑢(𝑡) = 𝑢0 cos(𝜔𝑛 𝑡) + 𝑠𝑖𝑛 (𝜔𝑛 𝑡)
initial conditions: 𝜔𝑛

Also, we can simplify this General solution with multiply and dividing this equation by the

constant value of √𝐴1 2 + 𝐴2 2 :

a b

𝐴1 𝐴2
𝑢(𝑡) = √𝐴1 2 + 𝐴2 2 cos(𝜔𝑛 𝑡) + 𝑠𝑖𝑛 (𝜔𝑛 𝑡)
2 2 2 2
√𝐴1 + 𝐴2 √𝐴1 + 𝐴2
( )

𝑎2 + 𝑏 2 = 1 𝑠𝑖𝑛 2 𝜃 + 𝑐𝑜𝑠 2 𝜃 = 1

(Always there is an angle 𝜃 which can satisfy this equation. Also, we can replace √𝐴1 2 + 𝐴2 2
with some other constant value A which can be find from the initial condition.

1
 ME 4440-5540 Lecture 6

𝑎 = 𝑠𝑖𝑛𝜃 , 𝑏 = 𝑐𝑜𝑠𝜃

From initial condition:

𝑢̇ 2
𝐴 = √𝐴1 2 + 𝐴2 2 = √𝑢0 2 + (𝜔0 ) = amplitude
𝑛

𝐴 𝑢0 𝜔𝑛
𝜃 = 𝑡𝑎𝑛−1 (𝐴1 ) = 𝑡𝑎𝑛−1 ( )= Phase angle
2 𝑢̇ 0

Phase angle shows how the signal lags behind the sinusoidal function.

𝑢(𝑡) = 𝐴(𝑠𝑖𝑛𝜃 cos(𝜔𝑛 𝑡) + 𝑐𝑜𝑠𝜃 𝑠𝑖𝑛 (𝜔𝑛 𝑡))

General Solution (in the term

of single sinusoidal form): 𝑢(𝑡) = 𝐴 sin(𝜔𝑛 𝑡 + 𝜃)

𝑢(𝑡)

𝜃
𝜃

𝑢(𝑡) 𝑢(𝑡) = 𝐴 sin(𝜔𝑛 𝑡 + 𝜃)

2
 ME 4440-5540 Lecture 6

𝑢̇ 0
𝑢(𝑡)
Velocity zero
𝑇

𝑢0

𝑢̇ 0
𝑢0 2
𝑢(𝑡) = 𝐴 sin(𝜔𝑛 𝑡 + 𝜃)

Velocity is the derivative of displacement function and would be in form of sin function. Therefore,
whenever displacement is maximum, velocity is zero and for zero displacement, we will have
𝜋
maximum velocity and the change from zero velocity to maximum velocity takes (𝑡 = ).
2𝜔𝑛

 As it discussed before, one of the important concepts for any physical system is natural
frequency.
 The motion of S.D.O.F. for an undamped system is harmonic (it repeats itself after each period
2𝜋
𝑇 = 𝜔 sec).
𝑛

𝑘
 𝜔𝑛 = √𝑚 (rad/sec)
𝜔𝑛
 Hertz: We can define the natural frequency in another form (unit: 𝑠 −1 ): 𝑓 = (Hz)
2𝜋
 If we write the mass in term of the weight natural frequency would be equal:
𝑘.𝑔 1 𝑘.𝑔
𝜔𝑛 = √ 𝑊 𝑓 = 2𝜋 √ 𝑊
𝑊
However, “g” is constant value & = 𝛿𝑠𝑡 (𝛿𝑠𝑡 : static displacement/deformation of the system)
𝑘
1 𝑘.𝑔 1 𝑔 1 𝛿
𝑓 = 2𝜋 √ 𝑊 = 2𝜋 √𝛿 𝑇 = 𝑓 = 2𝜋√ 𝑔𝑠𝑡
𝑠𝑡

That means the natural frequency of any system can be calculate by static equilibrium. So, if
you know the stiffness (k) of the system and weight (W) of the system you can find the natural
frequency of that system!

3
 ME 4440-5540 Lecture 7

Model a System as S.D.O.F

Summary of last time:
 We talked about a single degree of freedom system.
 The D' Alembert’s principle was expressed.
 Based on damped/Undamped and free vibration/forced vibration we classified all possible
S.D.O.F. systems.
 We analyzed part of Undamped System-Free Vibration and defined an important
characteristic of system, natural frequency (𝜔𝑛 )

Look at this system:

Equation of motion: 𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑓(𝑡)

Where is the gravitation force (mg) in the
equation of motion? Why it’s not appear in this
equation?

The equation of motion is based on the assumption that the dynamic motion starts from the state
of static equilibrium.

𝑓(𝑡)
𝑚𝑔 𝑢 = 𝑢𝑠 + 𝑢′

𝑚𝑔

1 2 3

1
 ME 4440-5540 Lecture 7

Free Body Diagram 1 to 2:

𝑓𝑘 = 𝑘. 𝑢𝑠 = 𝑚𝑔

𝑚𝑔
𝑢𝑠 =
𝑘

Free Body Diagram 1 to 3:

𝑚𝑢̈′ + 𝑘𝑢′ = 𝑓(𝑡) − 𝑚𝑔

𝑚𝑔
𝑢 = 𝑢𝑠 + 𝑢′ 𝑢′ = 𝑢 − 𝑢𝑠 = 𝑢 − 𝑘
𝑚𝑔
= 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝑣𝑎𝑙𝑢𝑒 𝑢̈′ = 𝑢̈
𝑘

𝑚𝑔
𝑚𝑢̈ + 𝑘 (𝑢 − ) = 𝑚𝑢̈ + 𝑘𝑢 − 𝑚𝑔 = 𝑓(𝑡) − 𝑚𝑔
𝑘

𝑚𝑢̈ + 𝑘𝑢 = 𝑓(𝑡)

So, as you can see 𝑚𝑔 is not part of the equation of motion when the dynamic motion starts from
the state of static equilibrium.

2
 ME 4440-5540 Lecture 7

Before continue talking about undamped System-Free Vibration, we want to know how to set up
the S.D.O.F model of a physical system?

First of all, you need to know some basic concept of stiffness (review this part from mechanics
of material)

Based on the complexity of the system, there are three main approaches can be used to model a
system.
1) Simple cases: Direct Derivation
2) Medium cases: Using Structural analysis approach
3) Complex cases: Energy method:
 Conservation of energy
 Lagrange’s approach (for more complex systems)
 Rayleigh’s method (for the most complex systems)

1) Direct Derivation
Direct Derivation: Using Free Body Diagram and driving the equation of motion.
What is the stiffness? Ability of the structure to resist about deformation (Or force that generate a
unit displacement).
Consider we have a one directional system like a spring:
𝐹 = 𝑘. 𝑥
In this case for any input force, if displacement become equal 1 (𝑥 = 1), then stiffness of system
(𝑘) would be equal to the applied load (𝐹). 1
𝐹 = 𝑘. 𝑥 𝐹=𝑘
Example 1: We have a beam (system) which is fixed from one end and load is applying at the
other end of it.

One direction translational deformation case

If we have the properties of this beam (𝐸, 𝐼, 𝐿) we can write the displacement for the load (𝑃):
𝑃𝐿3
𝛿=
3𝐸𝐼

3
 ME 4440-5540 Lecture 7

Can you replace this beam with a single spring? Yes, because if you look at the figure, the beam
resisting against the displacement in the same direction of the applied load. So, this system is
equivalent to:

What is the stiffness of this beam equivalent?

If we put displacement of system equal to one, from previous discussion the force would be
equal to the stiffness (𝐹 = 𝑘) 1 k
𝑃𝐿3 3𝐸𝐼
𝛿= 𝑘=
3𝐸𝐼 𝐿3

So, this is how we can find an equivalent model of a physical system.

4
 ME 4440-5540 Lecture 8

Model a System as S.D.O.F

Example 2: We have a shaft with a disk (lumped mass) is subjected to a time dependent torque.
This shaft has the geometric properties of length (L), shear module (G), and moment of inertia of
the cross section (J). Find a single degree of freedom model that represent the physical system.
(We have to find mass and stiffness equivalent to this system)

One direction rotational deformation case

One of the objectives is finding the natural frequency of this system to be able to write the
equation of motion for this system.
In this system, we have a torsional force and a shaft which is resisting an angular deformation.
The rotational motion of the disk is resisted by the shaft (torsional spring).
𝑇𝐿
𝜃=
𝐽𝐺
So, if we put 𝜃 (rotational displacement/angular deformation) equal one then the load 𝑇 would be
equivalent to torsional stiffness of the spring (𝑘𝑡 ).

1 𝑘𝑡
𝑇𝐿 𝐽𝐺
𝜃 = 𝐽𝐺 𝑘𝑡 = 𝐿

𝜋𝐷4
(𝐽 = , D: Diameter of the shaft)
32

Free Body Diagram:

I: Moment of inertia

∅̈: Angular acceleration

1
 ME 4440-5540 Lecture 8

Equation of motion:

𝐼∅̈ + 𝑘𝑡 ∅ = 𝑇(𝑡)
𝐽𝐺
𝐼∅̈ + 𝐿 ∅ = 0

𝐽𝐺
𝑚𝑢̈ + 𝑘𝑢 = 𝑓(𝑡) 𝜔𝑛 = √ 𝐿𝐼

This system is equivalent to:

 What is the difference between ∅ & 𝜃? ∅ is time-dependent angular deformation when 𝜃 is

static deformation and stiffness is based on static deformation.

𝑘 𝑘
𝜔𝑛 = √ =√
𝑀𝑡𝑜𝑡𝑎𝑙 33
(𝑀 + 140 𝑚)

Example 3: A rigid beam with length of “L” is pinned to the wall at point A and a mass “m” is
connected to the other end of this beam. A spring with constant of “k” is attached somewhere
along the beam to the bottom of it. This beam is oscillating up and down with angle of 𝜃. Find
the equivalent system?

Free Body Diagram:

2
 ME 4440-5540 Lecture 8

Note: If angle 𝜃 is small enough, we can assume “x” in the Free Body diagram will be equal to:
(𝑥 = 𝑎. 𝜃).
Note: Because we have a rotational motion about point “A”, we will use ∑𝑀𝐴 = 0.

∑𝑀𝐴 = 0

𝐼𝜃̈ + 𝑓𝑘 (𝑎) = 0

Note: We assume the bar is mass less and the inertia term is just related to the lumped mass
(point mass), so 𝐼 = 𝑚𝐿2 .

𝑚𝐿2 𝜃̈ + 𝑘 (𝑎𝜃). 𝑎 = 𝑚𝐿2 𝜃̈ + 𝑘𝑎2 𝜃 = 0

𝑚𝑢̈ + 𝑘𝑢 = 𝑓(𝑡)

𝑎 𝑘
𝑚 = 𝑚𝐿2 , 𝑘 = 𝑘𝑎2 𝜔𝑛 = 𝐿 √𝑚

Example 4: A rigid bar with a mass of “m” and length of “L” is pinned at point A to the wall. This
bar is supported by a damper with a coefficient of “c” at point “C” and a spring with stiffness of
𝑥
“k” at point “B”. This bar is subjected to a triangular distributed load (𝑝(𝑥, 𝑡) = 𝑝0 𝐿 𝑓(𝑡)). Find
the equation of motion and natural frequency of the system?

3
 ME 4440-5540 Lecture 8

Free Body Diagram:

Note: If angle 𝜃 is small enough, we can assume displacement at point “B” and “C” in the Free
Body diagram will be equal to 𝑎. 𝜃 & 𝐿. 𝜃 respectively.

Note: Force of spring would be 𝑓𝑘 = 𝑘(𝑎. 𝜃) & force of damp would be 𝑓𝐷 = 𝑐(𝐿. 𝜃̇)
Note: The distributed force can be replaced by a single concentrated force at 2/3 of the length of
the beam.
∑ 𝑀𝐴 = 0
𝐿 2
𝐼𝜃̈ + 𝑓𝐷 (𝐿) + 𝑓𝑘 (𝑎) = [𝑝0 𝑓(𝑡) 2] (3 𝐿)

𝑚𝐿2
𝐼= 3

𝑚𝐿2 𝐿 2
( ) 𝜃̈ + 𝑐(𝐿. 𝜃̇)(𝐿) + 𝑘(𝑎. 𝜃)(𝑎) = [𝑝0 𝑓(𝑡) 2] (3 𝐿)
3

𝑚𝐿2 𝐿2
( ) 𝜃̈ + (𝑐𝐿2 )𝜃̇ + (𝑘𝑎2 )𝜃 = 𝑝0 3 𝑓(𝑡) Equation of Motion
3

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑓(𝑡)

𝑚𝐿2 𝑎 3𝑘
𝑚= , 𝑘 = 𝑘𝑎2 𝜔𝑛 = 𝐿 √ 𝑚
3

This system is equivalent to:

4
 ME 4440-5540 Lecture 9

Model a System as S.D.O.F

Example 5: Look at following simple structure. That includes a vertical rigid bar with mass of “m”
and length of “L” which is pinned at top and attached to a horizontal spring with a coefficient (k)
from bottom. This bar can rotate back and forth about the pinned point. Find the equivalent mass,
equivalent stiffness, and equivalent natural frequency.

Free Body Diagram:

We will have 5 forces on this system:
 Two forces in x & y directions to hold the bar
 If we assume 𝜃 be small then the reaction force from the spring would be 𝑓𝑘 = k. (𝐿𝜃)
 Force from the weight of bar (𝑚𝑔) which applied to the center of mass
 Force of inertia which is equal to 𝐼𝜃̈

1
 ME 4440-5540 Lecture 9

∑ 𝑀𝐴 = 0
𝐿𝜃
𝐼𝜃̈ + (𝑚𝑔 2 ) + 𝑓𝑘 (𝐿) = 0

𝑚𝐿2
I for the rigid bar about point “A” would be 𝐼 = & 𝑓𝑘 = 𝑘. 𝐿𝜃.
3

𝑚𝐿2 𝑚𝑔𝐿 𝑚𝐿 𝑚𝑔𝐿 2 𝑚𝐿 2

𝑚𝑔𝐿
( ) 𝜃̈ + 2 𝜃 + 𝑘. 𝐿𝜃(𝐿) = ( 3 ) 𝜃̈ + 2 𝜃 + 𝑘𝐿2 𝜃 = ( 3 ) 𝜃̈ + ( 2 + 𝑘𝐿2 ) 𝜃 = 0
3

𝑚𝑢̈ + 𝑘𝑢 = 𝑓(𝑡)
𝑚𝐿2
𝑚𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 =
3
𝑚𝑔𝐿
𝑘𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 = + 𝑘𝐿2
2

𝑚𝑔𝐿 𝑚𝑔
( +𝑘𝐿2 ) ( +𝑘𝐿)
2 2
𝜔𝑛 = √ 𝑚𝐿2
=√ 𝑚𝐿
( ) ( )
3 3

Example 6: In the following structure, we have a massless L shaped rigid bar. A lumped mass
attached to the point “C” of this bar. This bar is supported at point B and also supported at point
“A” with a spring with a coefficient of “k”. Other information for this structure is shown in the
figure. Find the equivalent mass, equivalent stiffness, and equivalent natural frequency.

2
 ME 4440-5540 Lecture 9

Free Body Diagram:

 Because bar assumed to be massless, so “𝐼𝜃̈” only related to the lumped mass.
 Remember “𝐼𝜃̈” is the force for D' Alembert’s principle and it is always in opposite
direction of motion.
 “m” is a point mass rotating about point “B”, so 𝐼 = 𝑚𝑏 2 .
 If we assume 𝜃 be small then the reaction force from the spring would be 𝑓𝑘 = k. (𝑎𝜃)

∑ 𝑀𝐵 = 0

𝐼𝜃̈ − 𝑚𝑔(𝑏𝜃) + 𝑓𝑘 (𝑎) = 0

(𝑚𝑏 2 )𝜃̈ − (𝑚𝑔𝑏)𝜃 + 𝑘. 𝑎𝜃(𝑎) = (𝑚𝑏 2 )𝜃̈ − (𝑚𝑔𝑏)𝜃 + 𝑘𝑎2 (𝜃) = (𝑚𝑏 2 )𝜃̈ + (𝑘𝑎2 − 𝑚𝑔𝑏)𝜃 = 0

𝑚𝑢̈ + 𝑘𝑢 = 𝑓(𝑡)
𝑚𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 = 𝑚𝑏2

𝑘𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 = 𝑘𝑎2 − 𝑚𝑔𝑏

(𝑘𝑎2 −𝑚𝑔𝑏)
𝜔𝑛 = √ (𝑚𝑏2 )

3
 ME 4440-5540 Lecture 10

2) Using Structural Analysis Approach

Structural analysis approach: Usually when focus is on finding equivalent stiffness of system
(hard to find) and equivalent of mass is easy to find (using technics like: superposition ,
definition of stiffness/flexibility, combination of stiffness elements with using the concept of
parallel and series springs)
In modeling a structure, the most important part is related to equivalent of stiffness element of
the structure. In reality most of the structures made of many pieces (e.g. truss, frame, building,
which are made of lots of loaded members) and each of these pieces have their own stiffness can
be represent by a spring. For such a complex structure like that, how do you add up the effect of
all of them and calculate the equivalent stiffness of the system?
In this part we will study two cases to see how we can replace all the stiffness elements in the
structure with a single stiffness element.
Case 1: Spring in Series: If all stiffness elements (springs) experience the same force.

Example 7: There are two bars with different geometric and material properties connected to
each other and subjected to an axial load. This can be replace by two springs with different
stiffness but they are subjected to the same force. So, these two elements (springs) will be
deformed differently. That is possible to replace these two elements with a single spring?

Displacement Displacement
of element-1 of element-2

In this case, both elements are subjected to the same force. For these two springs we can write:

𝑓
𝑓 = 𝑘1 (𝑢2 − 𝑢1 ) or = (𝑢2 − 𝑢1 ) 1
𝑘1
𝑓
𝑓 = 𝑘2 (𝑢3 − 𝑢2 ) or = (𝑢3 − 𝑢2 ) 2
𝑘2

For the equivalent spring we will have:

𝑓
𝑓 = 𝑘𝑒𝑞 (𝑢3 − 𝑢1 ) or 𝑘𝑒𝑞
= (𝑢3 − 𝑢1 ) 3

1
 ME 4440-5540 Lecture 10

But the summation of right side of equations 1 & 2 would be equal right side of equation 3. So,
the equivalent stiffness of two parallel springs can be find from following equation:

𝑓 𝑓 𝑓 1 1 1 𝑘 𝑘
(𝑢2 − 𝑢1 ) + (𝑢3 − 𝑢2 ) = (𝑢3 − 𝑢1 ) + =𝑘 =𝑘 + or 𝑘𝑒𝑞 = 𝑘 1+𝑘2
𝑘1 𝑘2 𝑒𝑞 𝑘𝑒𝑞 1 𝑘2 1 2

For “n” springs in series:

𝑛
1 1
=∑
𝑘𝑒𝑞 𝑘𝑖
𝑖=1

Example 8: We have two shafts with different diameters and properties, if they are subjected a
same torsional deformation, what is the stiffness of each element and equivalent stiffness?

𝐺1 𝐽1 𝐺2 𝐽2 𝐺1 𝐺2 𝐽1 𝐽2
𝐺1 𝐽1 𝐺2 𝐽2 × 𝐺1 𝐺2 𝐽1 𝐽2
𝐿1 𝐿2 𝐿1 𝐿2
𝑘𝑡1 = 𝑘𝑡2 = 𝑘𝑡_𝑒𝑞 = 𝐺1 𝐽1 𝐺2 𝐽2 = 𝐿2 𝐺1 𝐽1 +𝐿1 𝐺2 𝐽2 =𝐿
𝐿1 𝐿2 + 2 𝐺1 𝐽1 +𝐿1 𝐺2 𝐽2
𝐿1 𝐿2 𝐿1 𝐿2

Case 2: Spring in Parallel: If all stiffness elements (springs) experience the same deformation.

Example 9: There are two bars with different geometric and material properties and in a parallel
form connected to each other. This can be replace by two springs with different stiffness. These
two elements (springs) will have same deformation. That is possible to replace these two springs
with a single spring?

2
 ME 4440-5540 Lecture 10

From the free body diagram, we have:

𝑓 = 𝑓1 + 𝑓2

𝑓1 , 𝑓2 , and 𝑓 can be find from following equations:

𝑓1 = 𝑘1 (𝑢2 − 𝑢1 )
𝑓2 = 𝑘2 (𝑢2 − 𝑢1 )
𝑓 = 𝑘𝑒𝑞 (𝑢2 − 𝑢1 )

So,
𝑘𝑒𝑞 (𝑢2 − 𝑢1 ) = 𝑘1 (𝑢2 − 𝑢1 ) + 𝑘2 (𝑢2 − 𝑢1 ) 𝑘𝑒𝑞 = 𝑘1 + 𝑘2

For “n” parallel springs:

𝑘𝑒𝑞 = ∑ 𝑘𝑖
𝑖=1

Right now, the main challenges is related to find out if the springs are parallel or series with each
other.

Example 10: look at following Figure. There is a beam which is connected to two springs at the
end. This beam start to oscillate up and down. Is this structure a parallel or series case?

It seems the springs are align with each other and experience same force so it should be a series
case but it’s not! Actually beam has a deformation of 𝛿 and both springs also have 𝛿 deformation
too, so this structure is a parallel case!

3
 ME 4440-5540 Lecture 10

Note: That is so important to find if a structure is acting as a parallel case or series case. For
instance, in the above example, if we assumed that as a series case, if 𝑘1 and 𝑘2 are small values
(weak stiffness), based on series case equation, increasing the stiffness of bar would not effect on
equivalent stiffness (basically that part of equation goes to zero) while if it is assumed as a parallel
case, then increasing the stiffness of bar will effect a lot on the equivalent stiffness.
Zero

1 1 1 1
= + + Big number
𝑘𝑒𝑞 𝑘1 𝑘2 𝑘3

Big number Big number

𝑘𝑒𝑞 = 𝑘1 + 𝑘2 + 𝑘3

4
 ME 4440-5540 Lecture 11

2) Using Structural Analysis Approach

Example 11: Find the equivalent stiffness for the following structure.

Free Body Diagram:

This is Parallel case. So, we will have:

3𝐸1 𝐼1 𝐴2 𝐸2
𝑘𝑒𝑞𝑢𝑖𝑣𝑎𝑙𝑒𝑛𝑡 = 3 + + 𝑘3
𝐿1 𝐿2
Beam Spring
Bar

1
 ME 4440-5540 Lecture 11

Example 12: Find the equivalent stiffness for the following structure.

Free Body Diagram:

In this case, same force (W) is applied on both beam and rope but the rope and beam have
different displacements. So, this structure can be equivalent to series springs.

1 1 1 1 1 𝐿1 3 𝐿2
= + = + = +
𝑘𝑒𝑞 𝑘1 𝑘2 3𝐸1 𝐼1 𝐴2 𝐸2 3𝐸1 𝐼1 𝐴2 𝐸2
3 𝐿2
𝐿1

2
 ME 4440-5540 Lecture 11

Example 13: Find the equivalent stiffness for the following structure.

3
4

Step 1: Springs with stiffness 𝑘1 are parallel to each other.

𝑘𝑒𝑞_1 = 𝑘1 + 𝑘1 = 2𝑘1

3
 ME 4440-5540 Lecture 11

Step 2: Springs with stiffness 𝑘3 are parallel to each other.

𝑘𝑒𝑞_2 = 𝑘3 + 𝑘3 = 2𝑘3

Step 3: The equivalent of springs (𝑘𝑒𝑞_1 ) and spring (𝑘2 ) and equivalent of springs (𝑘𝑒𝑞_3 ) are
series with each other.

1 1 1 1 1 1 1 𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2
= + + = + + =
𝑘𝑒𝑞_3 𝑘𝑒𝑞_1 𝑘2 𝑘𝑒𝑞_2 2𝑘1 𝑘2 2𝑘3 2𝑘1 𝑘2 𝑘3

2𝑘1 𝑘2 𝑘3
𝑘𝑒𝑞_3 =
𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2

Step 4: The equivalent of springs (𝑘𝑒𝑞_3 ) and spring (𝑘4 ) are parallel to each other.

2𝑘1 𝑘2 𝑘3 𝑘4 (𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2 ) + 2𝑘1 𝑘2 𝑘3

𝑘𝑒𝑞_4 = 𝑘𝑒𝑞_3 + 𝑘4 = + 𝑘4 =
𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2 𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2

Step 5: The equivalent of springs (𝑘𝑒𝑞_4 ) and spring (𝑘5 ) are series with each other.

1 1 1 1 1
= + = +
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 𝑘𝑒𝑞_4 𝑘5 𝑘4 (𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2 ) + 2𝑘1 𝑘2 𝑘3 𝑘5
𝑘2 𝑘3 + 2𝑘1 𝑘3 + 𝑘1 𝑘2

𝑘2 𝑘3 𝑘4 𝑘5 + 2𝑘1 𝑘3 𝑘4 𝑘5 + 𝑘1 𝑘2 𝑘4 𝑘5 + 2𝑘1 𝑘2 𝑘3 𝑘5
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 =
𝑘2 𝑘3 𝑘4 + 𝑘2 𝑘3 𝑘5 + 𝑘1 𝑘3 𝑘4 + 2𝑘1 𝑘3 𝑘5 + 𝑘1 𝑘2 𝑘4 + 𝑘1 𝑘2 𝑘5 + 2𝑘1 𝑘2 𝑘3

4
 ME 4440-5540 Lecture 12

2) Using Structural Analysis Approach

Example 14: Find the equivalent stiffness for the following structure.

Free Body Diagram:

For the beams and bar in this structure we have same force but different displacement so they are
series with each other. The equivalent stiffness would be equal to:

3𝐸1 𝐼1
𝑘1 = 𝐿1 3
3𝐸2 𝐼2
𝑘2 = 𝐿2 3
𝐴3 𝐸3
𝑘3 = 𝐿3

1 1 1 1 𝐿1 3 𝐿2 3 𝐿3
= + + = + +
𝑘𝑒𝑞 3𝐸1 𝐼1 3𝐸2 𝐼2 𝐴3 𝐸3 3𝐸1 𝐼1 3𝐸2 𝐼2 𝐴3 𝐸3
𝐿1 3 𝐿2 3 𝐿3

1
 ME 4440-5540 Lecture 12

Example 15: Three springs and a mass are attached to a rigid, weightless bar PQ as shown in
following figure. Find the natural frequency of vibration of the system.

Free Body Diagram:

There are different methods to solve this problem. In this case because we don’t know about
relation between third spring and two others, we are going to use equivalent force system. Based
on this method, we can combine springs 1 & 2 and put an equivalent spring for them at the end
of the beam (Q). Then it possible to find the equivalent for whole structure.

For finding the equivalent of spring 1 &2, the moment respect to point P have to remain same.

𝑓𝑘1 = 𝑘1 (𝑙1 𝜃)
𝑓𝑘2 = 𝑘2 (𝑙2 𝜃)
𝑓𝑘 𝑒𝑞−12 = 𝑘𝑒𝑞−12 (𝑙3 𝜃)

∑ 𝑀𝑃 = 𝑓𝑘 𝑒𝑞−12 . 𝑙3 = 𝑓𝑘1 . 𝑙1 + 𝑓𝑘2 . 𝑙2

𝑘𝑒𝑞−12 (𝑙3 𝜃). 𝑙3 = 𝑘1 (𝑙1 𝜃). 𝑙1 + 𝑘2 (𝑙2 𝜃). 𝑙2

𝑘𝑒𝑞−12 𝑙3 2 𝜃 = 𝑘1 𝑙1 2 𝜃 + 𝑘2 𝑙2 2 𝜃

𝑙 2 𝑙 2
𝑘𝑒𝑞−12 = 𝑘1 (𝑙1 ) + 𝑘2 (𝑙2 )
3 3

2
 ME 4440-5540 Lecture 12

The third spring and this equivalent spring are series with each other. So, the total equivalent of
these three springs would be:

1 1 1 1 1
= + = 2 2 +
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 𝑘𝑒𝑞−12 𝑘3 𝑙 𝑙 𝑘3
𝑘1 ( 1 ) + 𝑘2 ( 2 )
𝑙3 𝑙3

𝑙 2 𝑙 2
[𝑘1 ( 1 ) + 𝑘2 ( 2 ) ] 𝑘3
𝑙3 𝑙3
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 =
𝑙 2 𝑙 2
𝑘1 ( 1 ) + 𝑘2 ( 2 ) + 𝑘3
𝑙3 𝑙3

𝑘1 𝑘3 𝑙1 2 + 𝑘2 𝑘3 𝑙2 2
𝜔𝑛 = √
𝑚(𝑘1 𝑙1 2 + 𝑘2 𝑙2 2 + 𝑘3 𝑙3 2 )

Example 16: Find the equivalent stiffness for the following structure. Assume that 𝑘1 , 𝑘2 , 𝑘3 ,
and 𝑘4 are torsional and 𝑘5 and 𝑘6 are linear spring constants.

3
 ME 4440-5540 Lecture 12

First of all, the rotation is happened on the disk between shaft 3 and 4. So, the shaft 1, 2, and 3
will experience same force and they are series with each other. On the other hands, shaft 4 and
equivalent of shafts 1-3 are in different sides of the moving disk so they will have same
deformation and parallel to each other. Springs 5 and 6 both have same deformation and parallel
to each other and rest of the system. However, for springs, we have to change the linear motion
to rotational motion to be able to adding them to the shafts. In this case, we can use the moment
relation between the spring and shafts.

Step 1:

1 1 1 1 𝑘1 𝑘2 𝑘3
=𝑘 +𝑘 +𝑘 𝑘𝑒𝑞_1 = 𝑘
𝑘𝑒𝑞_1 1 2 3 1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1

Step 2:

𝑘1 𝑘2 𝑘3 𝑘1 𝑘2 𝑘3 +𝑘4 (𝑘1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1 )

𝑘𝑒𝑞_2 = 𝑘𝑒𝑞_1 + 𝑘4 = 𝑘 + 𝑘4 =
1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1 𝑘1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1

Step 3:

𝑘𝑒𝑞_3 = 𝑘5 + 𝑘6

Step 4:

∑ 𝑀 = 𝑅. 𝐹𝑘
Moment
from the
For 𝜃 rotation: 𝐹𝑘 shafts

𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 . (𝜃) = 𝑅. 𝑘𝑒𝑞_3 . (𝑅𝜃)+𝑘𝑒𝑞_2 . (𝜃)

𝑘1 𝑘2 𝑘3 +𝑘4 (𝑘1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1 )

𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 = 𝑘𝑒𝑞_3 . 𝑅 2 +𝑘𝑒𝑞_2 = (𝑘5 + 𝑘6 ). 𝑅 2 + 𝑘1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1

4
 ME 4440-5540 Lecture 13

2) Using Structural Analysis Approach

Example 17: In the following figure, we have a rigid mass that rigidly connected to the ground.
Find the equivalent stiffness for this structure.

If there is a rigid connection between the mass and supports, there is not any rotation happened
between them (𝜃 = 0) and we only have 𝛿 displacement. For solving this problem, we will use
the superposition method.

Internal moment
because of rigid
connection
We know 𝜃 = 0, so 𝜃1 + 𝜃2 = 0 and 𝛿 = 𝛿1 + 𝛿2 .

From mechanical of material:

𝐹𝐿3 𝐹𝐿2
𝛿1 = , 𝜃1 =
3𝐸𝐼 2𝐸𝐼
𝑀𝐿2 𝑀𝐿
𝛿2 = − 2𝐸𝐼 , 𝜃2 = − 𝐸𝐼

𝐹𝐿2 𝑀𝐿 𝐹𝐿2 𝑀𝐿 𝐹𝐿
𝜃1 + 𝜃2 = − =0 = 𝑀=
2𝐸𝐼 𝐸𝐼 2𝐸𝐼 𝐸𝐼 2

1
 ME 4440-5540 Lecture 13

𝐹𝐿 2
𝐹𝐿3 𝑀𝐿2 𝐹𝐿3 ( )𝐿 𝐹𝐿3 𝐹𝐿3 𝐹𝐿3 𝐹𝐿3
2
𝛿 = 𝛿1 + 𝛿2 = 3𝐸𝐼 − = 3𝐸𝐼 − = − 4𝐸𝐼 = 12𝐸𝐼 𝛿 = 12𝐸𝐼
2𝐸𝐼 2𝐸𝐼 3𝐸𝐼
1 k
𝐹𝐿3 12𝐸𝐼
So, stiffness for each support element would be equal: 𝛿 = 12𝐸𝐼 𝑘= 𝐿3

24𝐸𝐼
Total stiffness of structure have to multiply by two. 𝑘𝑒𝑞 = 𝐿3

Example 18: In the following figure, we have a weightless disk with a spring with coefficient of
“k” attached to the wall. This disk can rotate around point “O”. A mass with a non-deformable
cable is connected to this disk. Find the equivalent stiffness and natural frequency for this structure.

If we want to solve this problem from the main definition of stiffness which is force based on unit
displacement, it would be hard to identify the unit displacement. So, in this case we will use a
method is named Flexibility.
1
Flexibility (the displacement for unit force) is the inverse of equivalent stiffness (𝑘 ) or in other
𝑒𝑞
words, it is the displacement based on the unit force.

2
 ME 4440-5540 Lecture 13

Free Body Diagram:

∑ 𝑀𝑂 = 0

𝑓𝑘 . 𝑎 = 𝑟. 1 𝑘(𝑎𝜃) = 𝑟
𝛿𝐵 𝑟2
𝑘 (𝑎 )𝑎 = 𝑟 𝛿𝐵 = 𝑘.𝑎2
𝛿𝐵 𝑟
𝑟𝜃 = 𝛿𝐵 𝜃= 𝑟

As it mentioned before in the flexibility method, the displacement for unit force (flexibility)
1
equal (𝑘 ). So, the equivalent stiffness and natural frequency for this example would be:
𝑒𝑞

1 𝑟2 𝑘.𝑎2 𝑎 𝑘
= 𝑘.𝑎2 𝑘𝑒𝑞 = 𝜔𝑛 = 𝑟 √𝑚
𝑘𝑒𝑞 𝑟2

3
 ME 4440-5540 Lecture 14

3) Energy method

Energy method is based on:

A) Conservation of energy
Summation of kinetic and potential energy is constant: 𝐾𝑒 + 𝑃𝑒 = 𝐸
B) Rayleigh’s method
C) Lagrange’s approach

A) Conservation of energy
In conservation of energy, we say the total energy of system is made of two major source kinetic
energy (Ke) and potential energy (Pe).
𝑃𝑒 + 𝐾𝑒 = 𝐸𝑡𝑜𝑡𝑎𝑙
Also, 𝐸𝑡𝑜𝑡𝑎𝑙 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 so:
𝑑 𝑑
𝐸𝑡𝑜𝑡𝑎𝑙 = 0 (𝑃𝑒 + 𝐾𝑒) = 0
𝑑𝑡 𝑑𝑡

If you remember, the kinetic and potential energy of a spring respectively equal to:
1 1
𝐾𝑒 = 2 𝑚𝑢̇ 2 𝑃𝑒 = 2 𝑘𝑢2

 Let’s check this method for the general solution of displacement (based on initial conditions)
of an undamped system. From above equations if velocity become zero, the kinetic energy
become zero and potential energy become maximum. On the other hands, if displacement
become zero the potential energy become zero and kinetic energy become maximum.
First, we try the energy method for the case that system has initial displacement but its initial
velocity is zero. (𝑢0 ≠ 0 , 𝑢̇ 0 = 0)
𝑢̇ 0
𝑢(𝑡) = 𝑢0 cos(𝜔𝑛 𝑡) + 𝑠𝑖𝑛 (𝜔𝑛 𝑡) 𝑢(𝑡) = 𝑢0 cos(𝜔𝑛 𝑡) & 𝑢̇ (𝑡) = −𝜔𝑛 𝑢0 sin(𝜔𝑛 𝑡)
𝜔𝑛

𝑘
𝜔𝑛 = √
𝑚

1 1 1 1 1
𝐸 = 2 𝑘𝑢2 + 2 𝑚𝑢̇ 2 = 2 𝑘[𝑢0 cos(𝜔𝑛 𝑡)]2 + 2 𝑚[−𝜔𝑛 𝑢0 sin(𝜔𝑛 𝑡)]2 = 2 𝑘𝑢0 2 cos 2 (𝜔𝑛 𝑡) +
1 𝑘 1 1
𝑚 (𝑚) 𝑢0 2 sin2 (𝜔𝑛 𝑡) = 2 𝑘𝑢0 2 [cos2 (𝜔𝑛 𝑡) + sin2 (𝜔𝑛 𝑡)] = 2 𝑘𝑢0 2 Max potential energy
2

Now, we try the energy method for the case that system has initial velocity but its initial
displacement is zero. (𝑢̇ 0 ≠ 0 , 𝑢0 = 0)

Max kinetic energy

 ME 4440-5540 Lecture 14

𝑢̇ 0 𝑢̇
𝑢(𝑡) = 𝑢0 cos(𝜔𝑛 𝑡) + 𝑠𝑖𝑛 (𝜔𝑛 𝑡) 𝑢(𝑡) = 𝜔0 𝑠𝑖𝑛 (𝜔𝑛 𝑡) & 𝑢̇ (𝑡) = 𝑢̇ 0 𝑐𝑜𝑠 (𝜔𝑛 𝑡)
𝜔𝑛 𝑛

1 1 1 𝑢̇ 2 1 1 𝑚
𝐸 = 2 𝑘𝑢2 + 2 𝑚𝑢̇ 2 = 2 𝑘 [𝜔0 𝑠𝑖𝑛 (𝜔𝑛 𝑡)] + 2 𝑚[𝑢̇ 0 𝑐𝑜𝑠 (𝜔𝑛 𝑡)]2 = 2 𝑘𝑢̇ 0 2 ( 𝑘 ) sin2 (𝜔𝑛 𝑡) +
𝑛
1 2 1 1
𝑚 𝑢̇ 0 cos 2 (𝜔𝑛 𝑡) = 2 𝑚𝑢̇ 0 2 [sin2 (𝜔𝑛 𝑡) + cos 2 (𝜔𝑛 𝑡)] = 2 𝑚𝑢̇ 0 2
2

1 1
As it mentioned before, 𝐸 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡, so: 𝑘𝑢0 2 = 2 𝑚𝑢̇ 0 2 or 𝑃𝑒𝑚𝑎𝑥 = 𝐾𝑒𝑚𝑎𝑥
2

 We can write the conservation of energy equation for the general solution of displacement (in
the term of single sinusoidal) for an undamped system. In this equation “A” is amplitude
which is equal maximum displacement. If we replace the conservation energy with the
values of Pe and Ke for the spring:
Max 1

𝑢(𝑡) = 𝐴 sin(𝜔𝑛 𝑡 + 𝜃) 𝑢𝑚𝑎𝑥 = 𝐴

Max 1

𝑢̇ (𝑡) = 𝐴𝜔𝑛 cos(𝜔𝑛 𝑡 + 𝜃) 𝑢̇ 𝑚𝑎𝑥 = 𝐴𝜔𝑛 𝑢̇ 𝑚𝑎𝑥 = 𝜔𝑛 𝑢𝑚𝑎𝑥

We have 𝑢̇ 𝑚𝑎𝑥 for maximum kinetic energy and 𝑢𝑚𝑎𝑥 for maximum potential energy. From
previous calculation we have:

1 1 𝑘
𝑚𝑢̇ 𝑚𝑎𝑥 2 = 2 𝑘𝑢𝑚𝑎𝑥 2 𝑢̇ 𝑚𝑎𝑥 = √𝑚 𝑢𝑚𝑎𝑥
2

If we replace it in previous equation:

𝑘 𝑘
√ 𝑢𝑚𝑎𝑥 = 𝜔𝑛 𝑢𝑚𝑎𝑥 𝜔𝑛 = √𝑚
𝑚

 We can write the conservation of energy equation for an undamped system. If we replace the
conservation energy with the values of Pe and Ke for the spring:
𝑑 𝑑 1 1
(𝑃𝑒 + 𝐾𝑒) = ( 𝑘𝑢2 + 2 𝑚𝑢̇ 2 ) = 0
𝑑𝑡 𝑑𝑡 2

𝑘𝑢𝑢̇ + 𝑚𝑢̇ 𝑢̈ = 0 𝑘𝑢 + 𝑚𝑢̈ = 0 Equation of motion for a system

with spring (no damped)

2
 ME 4440-5540 Lecture 14

Example 19: A vehicle with speed of 90 km/h hitting the guardrails in high way. This vehicle has
the weight of 1000kg and assume the guardrails have a displacement of 𝛿 = 0.25 𝑚. Find the
stiffness of this guardrails (𝑁/𝑚). The friction would be negligible in this problem.

Free Body Diagram:

First we have to change everything in a uniform unit.

𝑘𝑚 1ℎ 1000 𝑚 𝑚
𝑢̇ = 90 × 3600 𝑠 × = 25
ℎ 1 𝑘𝑚 𝑠

We know 𝐾𝑒𝑚𝑎𝑥 = 𝑃𝑒𝑚𝑎𝑥 and from the velocity of vehicle we can find the maximum kinetic
energy.
1 1 𝑚 2
𝐾𝑒𝑚𝑎𝑥 = 2 𝑚𝑢̇ 0 2 = 2 × 1000 kg × (25 𝑠 ) = 312500 𝑁. 𝑚 = 𝑃𝑒𝑚𝑎𝑥
1 1
𝑃𝑒𝑚𝑎𝑥 = 𝑘𝑢0 2 = 2 𝑘(0.25𝑚)2 = 312500 𝑁. 𝑚
2
312500 𝑁
𝑘= = 10,000,000 𝑚
0.0625

3
 ME 4440-5540 Lecture 15

3) Energy method

Example 20: In the following figure, a spring with coefficient of “k” from one side attached to a
disk and from other side is attached to the wall. This disk can rotate around point “O”. A mass
with a non-deformable cable is connected to this disk. Find the equivalent stiffness and natural
frequency for this structure.

1
𝛿𝐵 = 𝑟𝜃 & 𝐼 = 2 𝑀𝑟 2
1 1 1 1 1 1
𝐾𝑒 = 2 𝑚(𝛿𝐵̇ )2 + 2 𝐼𝜃̇ 2 = 2 𝑚(𝑟𝜃̇)2 + 2 𝐼𝜃̇ 2 = 2 𝑚𝑟 2 𝜃̇ 2 + 2 𝐼𝜃̇ 2
1 1
𝑃𝑒 = 2 𝑘(𝑎𝜃)2 = 2 𝑘𝑎2 𝜃 2
𝑑
(𝑃𝑒 + 𝐾𝑒) = 0
𝑑𝑡
𝑑
(𝑃𝑒) = 𝑘𝑎2 𝜃𝜃̇
𝑑𝑡

1
 ME 4440-5540 Lecture 15

𝑑
(𝐾𝑒) = 𝑚𝑟 2 𝜃̇𝜃̈ + 𝐼𝜃̇𝜃̈
𝑑𝑡

𝑘𝑎2 𝜃𝜃̇ + 𝑚𝑟 2 𝜃̇𝜃̈ + 𝐼𝜃̇𝜃̈ = 0 𝑘𝑎2 𝜃 + 𝑚𝑟 2 𝜃̈ + 𝐼𝜃̈ = (𝑘𝑎2 )𝜃 + (𝑚𝑟 2 + 𝐼)𝜃̈ = 0

𝑘𝑒𝑞 = 𝑘𝑎2
1
𝑚𝑒𝑞 = (𝑚𝑟 2 + 𝐼) = (𝑚𝑟 2 + 2 𝑀𝑟 2 )

𝑘𝑎2
𝜔𝑛 = √ 1
(𝑚𝑟 2 + 𝑀𝑟 2 )
2

Example 21: Find the equivalent stiffness for the following structure. Assume that 𝑘1 , 𝑘2 , 𝑘3 ,
and 𝑘4 are torsional and 𝑘5 and 𝑘6 are linear spring constants.

Step 1:

1 1 1 1 𝑘1 𝑘2 𝑘3
=𝑘 +𝑘 +𝑘 𝑘𝑒𝑞_1 = 𝑘
𝑘𝑒𝑞_1 1 2 3 1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1

Step 2:

1 1 1 1 1 1
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 . 𝜃 2 = 𝑘𝑒𝑞_1 . 𝜃 2 + 𝑘4 . 𝜃 2 + 𝑘5 . (𝑅𝜃)2 + 𝑘6 . (𝑅𝜃)2 = [𝑘𝑒𝑞_1 . 𝜃 2 + 𝑘4 . 𝜃 2 + 𝑘5 . 𝑅2 𝜃 2 +
2 2 2 2 2 2
𝑘6 . 𝑅 2 𝜃 2 ]

𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 = 𝑘𝑒𝑞_1 + 𝑘4 + 𝑘5 . 𝑅2 + 𝑘6 . 𝑅2

𝑘1𝑘2 𝑘3
𝑘𝑒𝑞_𝑡𝑜𝑡𝑎𝑙 =
𝑘1 𝑘2 +𝑘2𝑘3 +𝑘3 𝑘1
+ 𝑘4 + 𝑘5 . 𝑅2 + 𝑘6 . 𝑅2 = (𝑘5 + 𝑘6 ). 𝑅2 + 𝑘1𝑘2𝑘3𝑘+𝑘𝑘4(𝑘1 𝑘2 +𝑘2 𝑘3 +𝑘3 𝑘1 )
+𝑘 𝑘 +𝑘 𝑘
1 2 2 3 3 1

2
 ME 4440-5540 Lecture 15

B) Rayleigh’s method
Let’s imagine a simple system with only a beam with a mass at one end but this time beam has
weight (distributed through the beam). In other words, we can’t simply replace the weight by a
lumped force and basically it includes infinite lumped forces.

The static deflection of the beam:

𝑃𝑥 2
𝑢(𝑥) = (3𝐿 − 𝑥)
6𝐸𝐼
𝑃𝐿2 𝑃𝐿3
𝑢𝑚𝑎𝑥 = 𝑢(𝐿) = (3𝐿 − 𝐿) =
6𝐸𝐼 3𝐸𝐼
𝑢𝑚𝑎𝑥 𝑥 2 𝑢𝑚𝑎𝑥
𝑢(𝑥) = 3
(3𝐿 − 𝑥) = (3 𝑥 2 𝐿 − 𝑥 3 )
2𝐿 2𝐿3

The total kinetic energy of the beam:

𝑢̇ 𝑚𝑎𝑥
𝑢̇ (𝑥) = (3 𝑥 2 𝐿 − 𝑥 3 )
2𝐿3

𝑚
The maximum kinetic energy of the beam itself (𝐾𝑒𝑚𝑎𝑥 ) where 𝑚 is the total mass and ( 𝐿 ) is
the mass per unit length of the beam.
𝐿 2
1 𝑚 2
𝑚 𝐿 𝑢̇ 𝑚𝑎𝑥
𝐾𝑒𝑚𝑎𝑥 = ∫ . . (𝑢̇ (𝑥)) 𝑑𝑥 = ∫ ( 3 (3 𝑥 2 𝐿 − 𝑥 3 )) 𝑑𝑥
0 2 𝐿 2𝐿 0 2𝐿
2 𝐿
𝑚 𝑢̇ 𝑚𝑎𝑥 4 2 6 5
𝑚 𝑢̇ 2 𝑚𝑎𝑥 33𝐿7 33
= 6
∫ 9 𝑥 𝐿 + 𝑥 − 6 𝑥 𝐿 𝑑𝑥 = 7
( )=( 𝑚) 𝑢̇ 2 𝑚𝑎𝑥
2𝐿 4𝐿 0 8𝐿 35 280
1
We know 𝐾𝑒𝑚𝑎𝑥 = 2 𝑚𝑢̇ 2 𝑚𝑎𝑥 , so:
1 33 33
𝐾𝑒𝑚𝑎𝑥 = 2 𝑚𝑒𝑞 𝑢̇ 2 𝑚𝑎𝑥 = (280 𝑚) 𝑢̇ 2 𝑚𝑎𝑥 𝑚𝑒𝑞 = 140 𝑚

3
 ME 4440-5540 Lecture 15

Thus the total effective mass acting at the end of the beam would be:
𝑀𝑡𝑜𝑡𝑎𝑙 = 𝑀 + 𝑚𝑒𝑞

Natural frequency of the system would be:

𝑘 𝑘
𝜔𝑛 = √ =√
𝑀𝑡𝑜𝑡𝑎𝑙 33
(𝑀 + 140 𝑚)

4
 ME 4440-5540 Lecture 16

Damped Free Vibration S.D.O.F System

Look at this system:

Free Body Diagram:

𝑓𝑘 = 𝑘𝑢 , 𝑓𝑐 = 𝑐𝑢̇
Equation of motion: 𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0
Let’s solve this differential equation:
𝑚𝑥 2 + 𝑐𝑥 + 𝑘 = 0
−𝑐±√𝑐 2 −4𝑚𝑘 It has two roots, so we
𝑥1,2 = 2𝑚
will have two solutions

Two solutions:
𝑢1 (𝑡) = 𝑐1 𝑒 𝑥1 𝑡
𝑢2 (𝑡) = 𝑐2 𝑒 𝑥2 𝑡
General solution:
𝑢(𝑡) = 𝑐1 𝑒 𝑥1 𝑡 + 𝑐2 𝑒 𝑥2 𝑡

1
 ME 4440-5540 Lecture 16

Critical damping coefficient (𝑐𝑐 ): The critical damping is defined as value of damping which
makes 𝑐 2 − 4𝑚𝑘 = 0.

√𝑚
×
√𝑚
𝑐𝑐 = √4𝑚𝑘 = 2√𝑚𝑘 𝑐𝑐 = 2𝑚𝜔𝑛

In this case:
−𝑐𝑐 −2𝑚𝜔𝑛
𝑥1,2 = = = −𝜔𝑛
2𝑚 2𝑚
That is why it’s named critical damping!

Damping ratio (𝜁): Damping ratio is defined as the ratio of current damping over critical
𝑐
damping (𝑐 ).
𝑐

𝑐
𝜁=
𝑐𝑐

Based on values of 𝜁 and 𝜔𝑛 , the equation of motion can be rewrite:

1
×
𝑚 𝑐 𝑘
𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0 𝑢̈ + 𝑚 𝑢̇ + 𝑚 𝑢 = 0 𝑢̈ + 2𝜁𝜔𝑛 𝑢̇ + 𝜔𝑛 2 𝑢 = 0

𝜁
𝑐 𝑐 𝑐 𝑐
= 𝑐 × 𝑚𝑐 = 2𝜁𝜔𝑛
𝑚 𝑚 𝑐

𝑐𝑐
𝑐𝑐 = 2𝑚𝜔𝑛 = 2𝜔𝑛
𝑚

Also, we can rewrite the equation of roots:

A B
−𝑐 ± √𝑐 2 − 4𝑚𝑘
𝑥1,2 =
2𝑚

A: 𝑐
− = −𝜁𝜔𝑛
2𝑚

B: √𝑐 2 − 4𝑚𝑘 𝑐 2 − 4𝑚𝑘 2
=√ = √𝜁 2 𝜔𝑛 − 𝜔𝑛 2 = 𝑖𝜔𝑛 √1 − 𝜁 2
2𝑚 4𝑚2

𝑥1,2 = −𝜁𝜔𝑛 ± 𝑖𝜔𝑛 √1 − 𝜁 2 𝑥1,2 = 𝜔𝑛 (−𝜁 ± 𝑖√1 − 𝜁 2 )

2
 ME 4440-5540 Lecture 16

Different solution cases based on damping values (𝜁):

𝑐 𝑘
Case1: Underdamped system (𝜁 < 1 or 𝑐 < 𝑐𝑐 or 2𝑚 < √𝑚): We will have two complex
conjugate roots.

𝑥1,2 = 𝜔𝑛 (−𝜁 ± 𝑖√1 − 𝜁 2 )

2) 𝜔 𝑡 2) 𝜔 𝑡
𝑢(𝑡) = 𝑐1 𝑒 (−𝜁+𝑖√1−𝜁 + 𝑐2 𝑒 (−𝜁−𝑖√1−𝜁
2 2
𝑛 𝑛
= 𝑒 −𝜁 𝜔𝑛𝑡 [𝑐1 𝑒 𝑖√1−𝜁 𝜔𝑛 𝑡
+ 𝑐2 𝑒 −𝑖√1−𝜁 𝜔𝑛 𝑡
]

𝑢(𝑡) = 𝑒 −𝜁 𝜔𝑛𝑡 [(𝑐1 + 𝑐2 ) cos (√1 − 𝜁 2 𝜔𝑛 𝑡) + 𝑖(𝑐1 − 𝑐2 ) sin (√1 − 𝜁 2 𝜔𝑛 𝑡)]

Note: 𝑒 𝑖𝜃 = 𝑐𝑜𝑠𝜃 + 𝑖𝑠𝑖𝑛 𝜃

Damped natural frequency (𝜔𝑑 ): The frequency of underdamped case. 𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛

The rest of simplifications and steps can be find in the book but final solution will be:
𝑢(𝑡) = 𝐴1 𝑒 −𝜁 𝜔𝑛𝑡 cos(𝜔𝑑 𝑡 − ∅) or 𝑢(𝑡) = 𝐴2 𝑒 −𝜁 𝜔𝑛𝑡 sin(𝜔𝑑 𝑡 + ∅0 )
Note: 𝐴1 , 𝐴2 , ∅, ∅0 will be find from initial conditions.
In terms of initial displacement and velocity:
𝑡=0 𝑢(0) = 𝑢0 & 𝑢̇ (0) = 𝑢̇ 0
𝑢̇ 0 +𝜁 𝜔𝑛 𝑢0
𝑢(𝑡) = 𝑒 −𝜁 𝜔𝑛𝑡 [𝑢0 𝑐𝑜𝑠𝜔𝑑 𝑡 + 𝑠𝑖𝑛𝜔𝑑 𝑡]
𝜔𝑑

Note: The above equation show the 𝑢(𝑡) is a sinusoidal function or in other words, it is a
harmonic function. However, the amplitude of this function because of 𝑒 −𝜁 𝜔𝑛 𝑡 will
exponentially decrease with time.

u(t)

t
𝑡1 𝑡2 𝑡3 𝑡4

3
 ME 4440-5540 Lecture 16

Note: As it is shown in above graph, for underdamped case, always going from zero to maximum
displacement is much faster than return (𝑡1 < 𝑡2 & 𝑡3 < 𝑡4 , ....) like any shock absorber.
We can plot the function of 𝑢(𝑡) for a system from experimental tests (like following figure). For an
underdamped system, we will be able to calculate the damping ratio (ζ) of the system from the amplitude
of graph at time 𝑡1 & 𝑡2 when 𝑡2 =𝑡1 +T𝑑 (T𝑑 : Period of damped vibration).

𝑢(𝑡)

𝑢1
𝑢2

2𝜋
𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛 & 𝜔𝑑 = 𝑇
𝑑
2𝜋 2𝜋
= √1 − 𝜁2 𝜔𝑛 𝑇𝑑 =
𝑇𝑑 √1−𝜁 2 𝜔𝑛

𝑢 𝑒 −𝜁𝜔𝑛𝑡1 𝑒 −𝜁𝜔𝑛𝑡1
∆= 𝑢1 = 𝑒 −𝜁𝜔𝑛𝑡2 = = 𝑒 𝜁𝜔𝑛𝑇𝑑
2 𝑒 −𝜁𝜔𝑛(𝑡1 +𝑇𝑑 )
If we use natural logarithm from both side of above equation we can make it simpler:
𝑢1 2𝜋 2𝜋𝜁
𝛿 = 𝑙𝑛 = 𝜁𝜔𝑛 𝑇𝑑 = 𝜁𝜔𝑛 =
𝑢2 √1 − 𝜁 2 𝜔𝑛 √1 − 𝜁 2
In the most of the case in real structures and systems the 𝜁 is very small. So, if that is the case the
√1 − 𝜁 2 ≅ 1 and we can rewrite above equation:

𝛿 = 2𝜋𝜁 (𝜁 ≪ 1)

4
 ME 4440-5540 Lecture 17

Damped Free Vibration S.D.O.F System

𝑐 𝑘
Case2: Critically damped system (𝜁 = 1 or 𝑐 = 𝑐𝑐 or 2𝑚 = √𝑚): We will have two equal real
roots. -1 0
𝑥1,2 = 𝜔𝑛 (−𝜁 ± 𝑖√1 − 𝜁 2 ) 𝑥1,2 = −𝜔𝑛

General solution:
𝑢(𝑡) = (𝑐1 + 𝑐2 𝑡)𝑒 −𝜔𝑛 𝑡 Exponential
Linear part part

The above equation show the 𝑢(𝑡) is combination of linear and exponential function and it
means the system will be damped but not very quickly! Also, it is not a harmonic function, so it
doesn’t have repetition (non-periodic).
For finding the constant values 𝑐1 & 𝑐2 we need to use two initial conditions:
𝑡=0 𝑢(0) = 𝑢0 & 𝑢̇ (0) = 𝑢̇ 0
𝑐1 = 𝑢0
𝑐2 = 𝑢̇ 0 + 𝜔𝑛 𝑢0

𝑐 𝑘
Case3: Overdamped system (𝜁 > 1 or 𝑐 > 𝑐𝑐 or 2𝑚 > √𝑚): This We will have two real roots.

𝑥1,2 = 𝜔𝑛 (−𝜁 ± 𝑖√1 − 𝜁 2 ) 𝑥1,2 = 𝜔𝑛 (−𝜁 ± √𝜁 2 − 1)

General solution:
𝑢(𝑡) = 𝑐1 𝑒 𝑥1 𝑡 + 𝑐2 𝑒 𝑥2 𝑡
2 −1) 𝜔 𝑡 2 −1) 𝜔 𝑡
𝑢(𝑡) = 𝑐1 𝑒 (−𝜁+√𝜁 𝑛
+ 𝑐2 𝑒 (−𝜁−√𝜁 𝑛

The above equation show the 𝑢(𝑡) is an exponential function and it means the system take
infinite time to be completely damped! Also, it is not a harmonic function, so it doesn’t have
repetition (non-periodic).
For finding the constant values 𝑐1 & 𝑐2 we need to use two initial conditions:
𝑡=0 𝑢(0) = 𝑢0 & 𝑢̇ (0) = 𝑢̇ 0

1
 ME 4440-5540 Lecture 17

𝑢0 𝜔𝑛 (𝜁 + √𝜁 2 − 1) + 𝑢̇ 0
𝑐1 =
2 𝜔𝑛 √𝜁 2 − 1

−𝑢0 𝜔𝑛 (𝜁 − √𝜁 2 − 1) − 𝑢̇ 0
𝑐2 =
2 𝜔𝑛 √𝜁 2 − 1

Watch this video for comparison between overdamped, critical damped, and underdamped
systems: https://www.youtube.com/watch?v=99ZE2RGwqSM&list=LL&index=2

Example 22: The following system is given. The mass of the system m=2.5kg, coefficient of
spring k=10N/m. If the initial conditions of the system are 𝑢0 = 0.05𝑚 & 𝑢̇ 0 = 0 .Find the
response of system for following cases.
a) If C=5 𝑁 − 𝑠⁄𝑚
b) If C=10 𝑁 − 𝑠⁄𝑚
c) If C=12 𝑁 − 𝑠⁄𝑚

2
 ME 4440-5540 Lecture 17

This is a free vibration S.D.O.F damped system. For this problem or similar cases we have to do
the following steps:

Step 1: Free Body Diagram:

Step 2: Finding the natural frequency of system (𝜔𝑛 ):

𝑘 10 𝑟𝑎𝑑
𝜔𝑛 = √𝑚 = √2.5 = 2 𝑠
Step 2: Finding the critical damping coefficient of system (𝑐𝑐 ):
𝑐𝑐 = 2𝑚𝜔𝑛 = 2 × 2.5 × 2 = 10 𝑁 − 𝑠⁄𝑚

Step 3: Finding the damping ratio of system (𝜁) and find that is related to which damping case:
𝑐1 5
a) 𝜁1 = = 10 = 0.5 Underdamped
𝑐𝑐
𝑐2 10
b) 𝜁2 = 𝑐 = 10 = 1 Critically damped
𝑐
𝑐3 12
c) 𝜁3 = 𝑐 = 10 = 1.2 Overdamped
𝑐

Step 4: Write the equation of motion (only for under-damped case, first you have to calculate
damped natural frequency (𝜔𝑑 )):

a) Underdamped
𝑟𝑎𝑑
𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛 = 2 × √1 − 0.52 = 1.73
𝑠
−𝜁 𝜔𝑛 𝑡 𝑢̇ 0 +𝜁 𝜔𝑛 𝑢0
𝑢(𝑡) = 𝑒 [𝑢0 𝑐𝑜𝑠𝜔𝑑 𝑡 + 𝑠𝑖𝑛𝜔𝑑 𝑡] General solution for underdamped case
𝜔 𝑑
0.5 × 2 × 0.05
𝑢(𝑡) = 𝑒 −0.5×2𝑡 [0.05cos(1.73𝑡) + sin(1.73𝑡)]
1.73

𝑢(𝑡) = 𝑒 −𝑡 [0.05cos(1.73𝑡) + 0.029sin(1.73𝑡)]

b) Critically damped
𝑢(𝑡) = (𝑐1 + 𝑐2 𝑡)𝑒 −𝜔𝑛 𝑡 General solution for critically damped case
𝑐1 = 𝑢0 = 0.05
𝑐2 = 𝑢̇ 0 + 𝜔𝑛 𝑢0 = 2 × 0.05 = 0.1
𝑢(𝑡) = (0.05 + 0.1𝑡)𝑒 −2 𝑡

3
 ME 4440-5540 Lecture 17

c) Overdamped
2 −1) 𝜔 𝑡 2 −1) 𝜔
𝑢(𝑡) = 𝑐1 𝑒 (−𝜁+√𝜁 𝑛
+ 𝑐2 𝑒 (−𝜁−√𝜁 𝑛𝑡
General solution for overdamped case
𝑢0 𝜔𝑛 (𝜁 + √𝜁 2 − 1) + 𝑢̇ 0 0.05 × 2(1.2 + √(1.2)2 − 1)
𝑐1 = = = 0.07
2 𝜔𝑛 √𝜁 2 − 1 2 × 2√(1.2)2 − 1

−𝑢0 𝜔𝑛 (𝜁 − √𝜁 2 − 1) − 𝑢̇ 0 −0.05 × 2(1.2 − √(1.2)2 − 1)

𝑐2 = = = −0.02
2 𝜔𝑛 √𝜁 2 − 1 2 × 2√(1.2)2 − 1

2 −1) 2𝑡 2 −1) 2𝑡
𝑢(𝑡) = 0.07𝑒 (−1.2+√(1.2) − 0.02𝑒 (−1.2−√(1.2)
𝑢(𝑡) = 0.07𝑒 −1.073𝑡 − 0.02𝑒 −3.727𝑡

4
 ME 4440-5540 Lecture 18

Damped Free Vibration S.D.O.F System

Example 23: The following system is given. This system includes a massless rigid beam attached
to a lumped mass at the end of the beam, a spring with coefficient of “k” and damping system with
coefficient of “c”. The length of the beam L=2m, m=5kg, coefficient of spring k=20N/m, damping
coefficient c=8 𝑁 − 𝑠⁄𝑚. Find the equation of motion and response of system for initial condition
of 𝑢(0) = 0.03𝑚 & 𝑢̇ (0) = 0.2 𝑚/𝑠.

Step 1: Free Body Diagram:

This system is on static equilibrium position, so we can ignore the force of mg in free body
diagram.
We can write the moment for point “A”.

∑ 𝑀𝐴 = 0
3𝐿 𝐿
𝐼𝜃̈ + 𝑓𝑐 ( ) + 𝑓𝑘 ( ) = 0
4 4
3𝐿 3×2
𝑓𝑐 = 𝑐𝑢̇ = 𝑐 ( 4 ) 𝜃̇ = 8 ( 4 ) 𝜃̇ = 12 𝜃̇
𝐿 2
𝑓𝑘 = 𝑘𝑢 = 𝑘 ( ) 𝜃 = 20 ( ) 𝜃 = 10 𝜃
4 4
𝐼 = 𝑚𝐿2 = 5 × 22 = 5 × 4 = 20

3×2 2
20 𝜃̈ + 12 𝜃̇ ( ) + 10 𝜃 (4) = 0 20 𝜃̈ + 18 𝜃̇ + 5 𝜃 = 0 Equation of motion
4

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0

1
 ME 4440-5540 Lecture 18

𝑚𝑒𝑞 = 20 , 𝑐𝑒𝑞 = 18, 𝑘𝑒𝑞 = 5

Step 2: Finding the natural frequency of system (𝜔𝑛 ):

𝑘𝑒𝑞 5 1 𝑟𝑎𝑑
𝜔𝑛 = √𝑚 = √20 = 2
𝑒𝑞 𝑠

Step 2: Finding the critical damping coefficient of system (𝑐𝑐 ):

1
𝑐𝑐 = 2𝑚𝑒𝑞 𝜔𝑛 = 2 × 20 × 2 = 20 𝑁 − 𝑠⁄𝑚

Step 3: Finding the damping ratio of system (𝜁) and find that is related to which damping case:
𝑐𝑒𝑞 18
𝜁= = 20 = 0.9 underdamped
𝑐𝑐

Step 4: Because it is under-damped case, first you have to calculate damped natural frequency
(𝜔𝑑 )):
1 𝑟𝑎𝑑
𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛 = × √1 − 0.92 = 0.218
2 𝑠
−𝜁 𝜔𝑛 𝑡 𝑢̇ 0 +𝜁 𝜔𝑛 𝑢0
𝑢(𝑡) = 𝑒 [𝑢0 𝑐𝑜𝑠𝜔𝑑 𝑡 + 𝑠𝑖𝑛𝜔𝑑 𝑡] General solution for underdamped case
𝜔 𝑑
1
1 0.2 + 0.9 × × 0.03
𝑢(𝑡) = 𝑒 −0.9×2𝑡 [0.03cos(0.218𝑡) + 2 sin(0.218𝑡)]
0.218

𝑢(𝑡) = 𝑒 −0.45𝑡 [0.03cos(0.218𝑡) + 0.98sin(0.218𝑡)] Response of system

Example 24: The following system is given. This system includes a massless flexible beam
attached to a lumped mass at the end of the beam and damping system with coefficient of “c”. The
length of the beam L=2m, m=5kg, damping coefficient c=8 𝑁 − 𝑠⁄𝑚 and the beam has E=207×
109 𝑃𝑎, I=5× 10−11 𝑚4 . Find the equation of motion and response of system for initial condition
of 𝑢(0) = 0.03𝑚 & 𝑢̇ (0) = 0.2 𝑚/𝑠.

2
 ME 4440-5540 Lecture 18

In this case, we have a flexible beam and that means there is not any more relation between angle
θ, distance from the origin and beam displacement (beam has curve displacement).

Also, we don’t have an obvious spring but the beam act as a vertical spring at point “B”. So, if
we can move the damping system to the point “B”, then we will have a simple damped system
which both spring and damper are connected to the mass.

In this case, we can use the static deflection of the beam same as what we did in Rayleigh’s
method (The dynamic displacement would be similar to static displacement).
From before we had:
𝑢𝑚𝑎𝑥 𝑢𝐿
𝑢(𝑥) = 3
(3 𝑥 2 𝐿 − 𝑥 3 ) = 3 (3 𝑥 2 𝐿 − 𝑥 3 )
2𝐿 2𝐿
𝑢̇ 𝑚𝑎𝑥 𝑢̇ 𝐿
𝑢̇ (𝑥) = 3
(3 𝑥 2 𝐿 − 𝑥 3 ) = 3 (3 𝑥 2 𝐿 − 𝑥 3 )
2𝐿 2𝐿

And now:
3𝐿
𝑓𝑐 = 𝑐𝑢̇ ( 4 ) & 𝑓𝑐 ′ = 𝑐′𝑢̇ (𝐿)

3
 ME 4440-5540 Lecture 18

For ∑ 𝑀𝐴 we can write:

3𝐿 3𝐿
𝑐𝑢̇ ( 4 ) × = 𝑐′𝑢̇ (𝐿) × 𝐿
4

3𝐿 𝑢̇ 3𝐿 2 3𝐿 3 81
𝑢̇ ( 4 ) = 2𝐿𝐿3 (3 ( 4 ) 𝐿 − ( 4 ) ) = 128 𝑢̇ 𝐿

𝑢̇
𝑢̇ (𝐿) = 2𝐿𝐿3 (3 𝐿3 − 𝐿3 ) = 𝑢̇ 𝐿
3𝐿 81
×𝑐× 𝑢̇ 𝐿 243
𝑐′ = 4 128
𝑐 ′ = 512 𝑐
𝐿×𝑢̇ 𝐿

3𝐸𝐼 243 3𝐸𝐼 243 3×207×109 ×5×10−11

𝑚𝑢̈ + 𝑐 ′ 𝑢̇ + 𝑢 =0 𝑚𝑢̈ + (512) 𝑐𝑢̇ + 𝑢 = 5𝑢̈ + (512) × 8𝑢̇ + 𝑢=0
𝐿3 𝐿3 (2)3

5𝑢̈ + 3.8 𝑢̇ + (3.9)𝑢 = 0 Equation of motion

𝑚𝑒𝑞 = 5 , 𝑐𝑒𝑞 = 3.8, 𝑘𝑒𝑞 = 3.9

Step 2: Finding the natural frequency of system (𝜔𝑛 ):

𝑘𝑒𝑞 3.9 𝑟𝑎𝑑
𝜔𝑛 = √𝑚 = √ 5 = 0.88
𝑒𝑞 𝑠

Step 2: Finding the critical damping coefficient of system (𝑐𝑐 ):

𝑐𝑐 = 2𝑚𝑒𝑞 𝜔𝑛 = 2 × 5 × 0.88 = 8.8 𝑁 − 𝑠⁄𝑚

Step 3: Finding the damping ratio of system (𝜁) and find that is related to which damping case:
𝑐𝑒𝑞 3.8
𝜁= = 8.8 = 0.43 underdamped
𝑐𝑐

Step 4: Because it is under-damped case, first you have to calculate damped natural frequency
(𝜔𝑑 )):
𝑟𝑎𝑑
𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛 = 0.88 × √1 − 0.432 = 0.79
𝑠
−𝜁 𝜔𝑛 𝑡 𝑢̇ 0 +𝜁 𝜔𝑛 𝑢0
𝑢(𝑡) = 𝑒 [𝑢0 𝑐𝑜𝑠𝜔𝑑 𝑡 + 𝑠𝑖𝑛𝜔𝑑 𝑡] General solution for underdamped case
𝜔 𝑑
0.2 + 0.43 × 0.88 × 0.03
𝑢(𝑡) = 𝑒 −0.43×0.88𝑡 [0.03cos(0.79𝑡) + sin(0.79𝑡)]
0.79

𝑢(𝑡) = 𝑒 −0.38𝑡 [0.03cos(0.79𝑡) + 0.267sin(0.79𝑡)] Response of system

4
 ME 4440-5540 Lecture 19

Forced Vibration S.D.O.F System

As it mentioned before, we have different type of forces: A) Harmonic force B) Periodic force C)
General force (which can be divided to Deterministic and Random forces). In this part we will
take a look how each of these types of forces act on an Undamped & Damped forced systems.

A) Harmonic Force

A.1) Analyzing an Undamped System That Excited by a Harmonic Force

Look at the following system.

If a force 𝐹(𝑡) = 𝐹0 𝑐𝑜𝑠𝜔𝑡 act on the mass m of an Undamped system, the equation of motion
would be:

𝑚𝑢̈ + 𝑘𝑢 = 𝐹0 𝑐𝑜𝑠𝜔𝑡 Equation of motion

The total solution for this equation of motion includes two parts: 1) homogeneous solution
(called transient) which already find for free vibration, and 2) particular solution (called steady
state).
𝑢(𝑡)𝑡𝑜𝑡𝑎𝑙 = 𝑢ℎ (𝑡) + 𝑢𝑝 (𝑡)

𝑢ℎ (𝑡) = 𝐴1 cos(𝜔𝑛 𝑡) + 𝐴2 𝑠𝑖𝑛 (𝜔𝑛 𝑡)

Because in this case we have a harmonic force, from differential equation we know the particular
solution also would be harmonic with same frequency of harmonic force (𝝎).
𝑢𝑝 (𝑡) = 𝐴3 cos 𝜔𝑡

We have to find the value of amplitude for steady state response (𝐴3 ). By substituting 𝑢𝑝 in the
equation of motion:

1
 ME 4440-5540 Lecture 19

𝑚𝑢̈ 𝑝 + 𝑘𝑢𝑝 = 𝐹0 𝑐𝑜𝑠𝜔𝑡

} −𝑚𝐴3 𝜔2 cos 𝜔𝑡 + 𝑘𝐴3 cos 𝜔𝑡 = 𝐹0 𝑐𝑜𝑠𝜔𝑡
𝑢̈ 𝑝 = −𝐴3 𝜔2 cos 𝜔𝑡 1
× 𝐹0
2 𝐹0 𝑘 𝑘
𝐴3 cos 𝜔𝑡 (−𝑚𝜔 + 𝑘) = 𝐹0 𝑐𝑜𝑠𝜔𝑡 𝐴3 = (𝑘−𝑚𝜔2) 𝐴3 = 𝜔2
(1− 2 )
𝜔𝑛

𝐹0
In above equation, let’s name the 𝛿𝑠𝑡 = static deflection (displacement) which is the
𝑘
𝜔
deflection of the mass under the force 𝐹0 . Also, let’s name the 𝑟 = 𝜔 frequency ratio.
𝑛

Static Displacement Dynamic Displacement

𝛿𝑠𝑡
𝐴3 =
(1 − 𝑟 2 )

So, the particular solution would be:

𝛿𝑠𝑡
𝑢𝑝 = cos 𝜔𝑡
(1 − 𝑟 2 )

Note: The initial condition for particular solution: 𝑢𝑝 (0) ≠ 0 & 𝑢̇ 𝑝 (0) = 0

And, the total solution would be:

𝛿𝑠𝑡
𝑢(𝑡) = 𝐴1 cos(𝜔𝑛 𝑡) + 𝐴2 𝑠𝑖𝑛 (𝜔𝑛 𝑡) + cos 𝜔𝑡
(1 − 𝑟 2 )

Depends on natural Depends on natural

frequency 𝜔𝑛 frequency (𝜔𝑛 ) and
force frequency (𝜔)

For the initial condition 𝑢(0) = 𝑢0 & 𝑢̇ (0) = 𝑢̇ 0 , 𝐴1 & 𝐴2 would be equal to:

𝛿
𝑠𝑡 𝑢̇
𝐴1 = 𝑢0 − (1−𝑟 2) & 𝐴2 = 𝜔0
𝑛

2
 ME 4440-5540 Lecture 19

𝛿𝑠𝑡 𝑢̇ 0 𝛿𝑠𝑡
𝑢(𝑡) = (𝑢0 − 2
) cos(𝜔𝑛 𝑡) + 𝑠𝑖𝑛 (𝜔𝑛 𝑡) + cos 𝜔𝑡
(1 − 𝑟 ) 𝜔𝑛 (1 − 𝑟 2 )

For the initial condition 𝑢(0) = 0 & 𝑢̇ (0) = 0, the total response:

𝛿𝑠𝑡
𝑢(𝑡) = [cos 𝜔𝑡 − cos 𝜔𝑛 𝑡]
(1 − 𝑟 2 )

The relation between maximum dynamic deformation (𝐴3 ) and maximum static deformation
(𝛿𝑠𝑡 ) would be:

𝐴3 1
=
𝛿𝑠𝑡 (1 − 𝑟 2 )

1
Note: 𝑀 = (1−𝑟 2 ) is named magnification factor and the absolute value of that is named
1
frequency response function: H(r) = (1−𝑟 2 )

𝐴3
𝛿𝑠𝑡

𝜔
𝑟=
𝜔𝑛
√2

This graph shows, the value of frequency is growing till frequency become close to natural
frequency of system and 𝑟 = 1, then dynamic response goes to infinity (resonance). Then if the
frequency still continue growing at 𝑟 = √2 (𝜔 = √2𝜔𝑛 ) the dynamic response and static
response become equal and static and dynamic displacement would be equal (This would be
happened in 𝑟 = 0 too but that means 𝜔 = 0 so basically we don’t have any dynamic forces!).
With growing 𝑟 more than √2, frequency response function H(r) become closer and closer to
zero. This means for a large value of frequency of harmonic force (𝜔) the dynamic force will
make less displacement than static force! So, that means: “Not always the response from a
dynamic force is worse than the response from a static force.”

3
 ME 4440-5540 Lecture 20

Forced Vibration S.D.O.F System

Based on value of the frequency ratio of “𝑟” there are three cases:

Case 1: 0 < 𝑟 < 1

𝛿
𝑠𝑡
That means 𝐴3 = (1−𝑟 2)
become positive and the particular (steady state) response of system (𝑢𝑝 )
and external force are in the same phase and they just have different amplitudes.
𝑢𝑝 (𝑡) = 𝐴3 𝑐𝑜𝑠𝜔𝑡
𝐹(𝑡) = 𝐹0 𝑐𝑜𝑠𝜔𝑡

𝑢𝑝 (𝑡) = 𝐴3𝑐𝑜𝑠𝜔𝑡

𝐴3

Case 2: 𝑟 > 1
𝛿
𝑠𝑡
That means 𝐴3 = (1−𝑟 2 ) become negative and the particular (steady state) response of system

(𝑢𝑝 ) and external force are in the opposite phase (180° out of phase).

𝑢𝑝 (𝑡) = 𝐴3𝑐𝑜𝑠𝜔𝑡

𝐴3

1
 ME 4440-5540 Lecture 20

Case 3: 𝑟 = 1 and that means 𝜔 = 𝜔𝑛

0
In this case, the response of system 𝑢(𝑡) indeterminate form of (0), so we have to use
L’Hopital’s rule to solve it.

𝛿𝑠𝑡 𝜔𝑛 𝑡
𝑢(𝑡) = sin 𝜔𝑛 𝑡
2

Here is the plot of this response function. As you can see, the amplitude of this response grows
linearly and goes to infinity this phenomenon is named resonance.

𝑢𝑝 (𝑡)

Case 3′: 𝑟 ≅ 1 or 𝜔 ≅ 𝜔𝑛 , if the harmonic forcing frequency is very close to natural frequency
of system but they are exactly same. This phenomenon is named beating.

The total response for initial conditions equal zero we had:

𝛿𝑠𝑡 𝐹0
𝑢(𝑡) = [cos 𝜔𝑡 − cos 𝜔𝑛 𝑡] = [cos 𝜔𝑡 − cos 𝜔𝑛 𝑡]
(1 − 𝑟 2 ) 𝑚(𝜔𝑛 2 − 𝜔 2 )

So, for the case of 𝜔 ≅ 𝜔𝑛 , we can write the following equation (𝜀 is very small value):

𝜔𝑛 − 𝜔 = 2𝜀 & 𝜔𝑛 + 𝜔 ≅ 2𝜔 𝜔𝑛 2 − 𝜔2 = (𝜔𝑛 − 𝜔)(𝜔𝑛 + 𝜔) = 2𝜀 × 2𝜔 = 4𝜀𝜔

So,
𝐹0 𝜔+𝜔𝑛 𝜔 − 𝜔𝑛 𝐹0 𝜔+𝜔𝑛 𝜔𝑛 − 𝜔
𝑢(𝑡) = [−2 sin 𝑡 . sin 𝑡] = [2 sin 𝑡 . sin 𝑡]
𝑚(𝜔𝑛 2 − 𝜔 2 ) 2 2 𝑚(𝜔𝑛 2 − 𝜔 2 ) 2 2

𝐹0 2𝜔 2𝜀 𝐹0
𝑢(𝑡) = [2 sin 𝑡 . sin 𝑡] = [sin 𝜔 𝑡 . sin 𝜀𝑡]
𝑚(4𝜀𝜔) 2 2 2𝑚𝜀𝜔
2

2
 ME 4440-5540 Lecture 20

𝐹0 sin 𝜀𝑡
𝑢(𝑡) = (sin 𝜔 𝑡)
2𝑚𝜀𝜔

Variable Phase
Amplitude
2𝜋 2𝜋
The period for sin 𝜀𝑡 would be 𝑇1 = and the period for sin 𝜔𝑡 would be 𝑇2 = . Because 𝜀 is
𝜀 𝜔
an very small number, 𝑇1 ≫ 𝑇2 . Basically, in this case, we have two signals and when they are in
phase, their amplitudes add to each other and when they are out of phase they cancel each other
𝜋
and this process repeat for each ( 𝜀 ).

𝑢(𝑡)

Example 25: In the following system, the springs has coefficient of 𝑘 = 10𝑁/𝑚 and mass of the
system is equal to 𝑚 = 5𝑘𝑔. A dynamic force, 𝑓(𝑡) = 10 𝑐𝑜𝑠 5𝑡 is applying to this system. For
initial conditions 𝑢(0) = 0 & 𝑢̇ (0) = 0, find the equation of motion and plot the response.

𝑠𝑡 𝛿
𝑢(𝑡) = 𝐴1 cos(𝜔𝑛 𝑡) + 𝐴2 𝑠𝑖𝑛 (𝜔𝑛 𝑡) + (1−𝑟 2 ) cos 𝜔𝑡

3
 ME 4440-5540 Lecture 20

Step 1: Let’s find the amplitude of the steady state part of response:

Springs are parallel, so:

𝑘𝑒𝑞 = 𝑘 + 𝑘 = 2𝑘 = 2 × 10 = 20𝑁/𝑚 , 𝐹0 = 10𝑁 , 𝜔 = 5 𝑟𝑎𝑑/𝑠

20
𝜔𝑛 = √ 5 = 2 𝑟𝑎𝑑/𝑠

𝐹 10
𝛿𝑠𝑡 = 𝑘 0 = 20 = 0.5 𝑚
𝑒𝑞
𝜔 5
𝑟 = 𝜔 = 2 = 2.5 > 1 Case 2
𝑛

𝛿𝑠𝑡 0.5
= (1−(2.5)2 ) = −0.095
(1−𝑟 2 )

Step 2: Finding 𝐴1 & 𝐴2 from initial condition.

𝑠𝑡 𝛿
𝑢̇ (𝑡) = −𝐴1 𝜔𝑛 sin(𝜔𝑛 𝑡) + 𝐴2 𝜔𝑛 𝑐𝑜𝑠 (𝜔𝑛 𝑡) − 𝜔 (1−𝑟 2 ) sin 𝜔𝑡

𝑢(𝑡) = 𝐴1 cos 2𝑡 + 𝐴2 𝑠𝑖𝑛 2𝑡 − 0.095 cos 5𝑡

𝑢(0) = 𝐴1 − 0.095 = 0 𝐴1 = 0.095

𝑢̇ (0) = 𝐴2 𝜔𝑛 = 0 𝐴2 = 0

Step 3: Total response would be equal to:

𝑢(𝑡) = 0.095 cos 2𝑡 − 0.095 cos 5𝑡

Step 4: Two cos function the plot would be look like this:

But for more exact sketch, you can plot it in MATLAB.

4
 ME 4440-5540 Lecture 21

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Look at the following system.

If a force 𝐹(𝑡) = 𝐹0 𝑐𝑜𝑠𝜔𝑡 act on the mass m of a Damped system, the equation of motion
would be:

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝐹0 𝑐𝑜𝑠𝜔𝑡

Or

𝐹0
𝑢̈ + 2𝜁 𝜔𝑛 𝑢̇ + 𝜔𝑛 2 𝑢 = 𝑐𝑜𝑠𝜔𝑡 Equation of motion
𝑚

The total solution for this equation of motion includes two parts: 1) homogenous solution (called
transient) which already find for free vibration, and 2) particular solution (called steady state).
𝑢(𝑡)𝑡𝑜𝑡𝑎𝑙 = 𝑢ℎ (𝑡) + 𝑢𝑝 (𝑡)

𝑢ℎ (𝑡) = 𝐴1 𝑒 −𝜁 𝜔𝑛𝑡 cos(𝜔𝑑 𝑡 − ∅0 )

Because in this case we have a harmonic force, from differential equation we know the particular
solution also would be harmonic with but not in phase with frequency of the harmonic force
(𝝎).
𝑢𝑝 (𝑡) = 𝐴3 cos(𝜔𝑡 − ∅)

So in this case, we have to find the value of amplitude for steady state response (𝐴3 ) and phase
angle of response (∅). By substituting 𝑢𝑝 in the equation of motion:

𝑚𝑢̈ 𝑝 + 𝑐𝑢̇ 𝑝 + 𝑘𝑢𝑝 = 𝐹0 𝑐𝑜𝑠𝜔𝑡

𝑢̈ 𝑝 = −𝐴3 𝜔2 cos(𝜔𝑡 − ∅) } −𝑚𝐴3 𝜔2 cos(𝜔𝑡 − ∅) − 𝑐𝐴3 𝜔 sin(𝜔𝑡 − ∅) + 𝑘𝐴3 cos(𝜔𝑡 − ∅) = 𝐹0 𝑐𝑜𝑠𝜔𝑡
𝑢̇ 𝑝 = −𝐴3 𝜔 sin(𝜔𝑡 − ∅)

𝐴3 [(𝑘 − 𝑚𝜔2 ) cos(𝜔𝑡 − ∅) − 𝑐𝜔 sin(𝜔𝑡 − ∅)] = 𝐹0 𝑐𝑜𝑠𝜔𝑡

1
 ME 4440-5540 Lecture 21

cos(𝜔𝑡 − ∅) = cos 𝜔𝑡 cos ∅ + sin 𝜔𝑡 sin ∅

sin(𝜔𝑡 − ∅) = sin 𝜔𝑡 cos ∅ − cos 𝜔𝑡 sin ∅

𝐴3 [(𝑘 − 𝑚𝜔2 )[cos 𝜔𝑡 cos ∅ + sin 𝜔𝑡 sin ∅ ] − 𝑐𝜔[sin 𝜔𝑡 cos ∅ − cos 𝜔𝑡 sin ∅ ]] = 𝐹0 𝑐𝑜𝑠𝜔𝑡

Anything in the left side of above equation which is coefficient of cos 𝜔𝑡 would be equal to 𝐹0
and rest of them (parts Not include𝑠 cos 𝜔𝑡) would be equal to zero. So, we will have:

1) 𝐴3 [(𝑘 − 𝑚𝜔2 ) cos ∅ + 𝑐𝜔 sin ∅ ] = 𝐹0

2) 𝐴3 [(𝑘 − 𝑚𝜔2 ) sin ∅ − 𝑐𝜔 cos ∅] = 0

For finding 𝐴3 , square both equations and add them with each other:

𝐴3 2 [(𝑘 − 𝑚𝜔2 )2 cos2 ∅ + 𝑐 2 𝜔2 sin2 ∅ + 2(𝑘 − 𝑚𝜔2 )𝑐𝜔 cos ∅ sin ∅] = 𝐹0 2

+ 𝐴3 2 [(𝑘 − 𝑚𝜔2 )2 sin2 ∅ + 𝑐 2 𝜔2 cos 2 ∅ − 2(𝑘 − 𝑚𝜔2 )𝑐𝜔 cos ∅ sin ∅] = 0

𝐴3 2 [(𝑘 − 𝑚𝜔2 )2 + 𝑐 2 𝜔2 ] = 𝐹0 2

𝐹0
𝐹0 𝑘
𝐴3 = =
√[(𝑘 − 𝑚𝜔 ) + 𝑐 𝜔 ] √[(1 − 𝑟 )2 + (2𝜁𝑟)2 ]
2 2 2 2 2

𝑐𝜔 2𝜁𝑟
∅ = 𝑡𝑎𝑛−1 ( 2
) = ∅ = 𝑡𝑎𝑛−1 ( )
𝑘 − 𝑚𝜔 1 − 𝑟2

𝑐
Note: (𝑟 = 𝜔/𝜔𝑛 , 𝜁 = 𝑐 , 𝑐𝑐 = 2𝑚𝜔𝑛 )
𝑐

Let’s check this values for an Undamped case (𝜁 = 0):

Amplitude of the steady state response:

𝐹0 Static displacement
𝐴3 = 𝑘
(1 − 𝑟 2 )

1 1
𝑀 = (1−𝑟 2 ) (𝑀𝑎𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟) 𝐻(𝑟) = |(1−𝑟 2 )| Frequency response function

However this values for a Damped case (𝜁 ≠ 0):

2
 ME 4440-5540 Lecture 21

𝐹0 Static displacement

𝐴3 = 𝑘
√[(1 − 𝑟 )2 + (2𝜁𝑟)2 ]
2

1 1 Frequency
𝑀= (𝑀𝑎𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟) |𝐻(𝑟)| =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] √[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] response function

In this case we have “complex frequency response function:

1
𝐶𝑜𝑚𝑝𝑙𝑒𝑥 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑟𝑒𝑠𝑝𝑜𝑛𝑠𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛
(1 − 𝑟 2) + 𝑖2𝜁𝑟

The magnitude of this function would be equal to the Magnification factor:

A B
Proof:
1 (1 − 𝑟 2 ) − 𝑖2𝜁𝑟 (1 − 𝑟 2 ) 2𝜁𝑟
2
× 2
=( 2 2 2
)−𝑖( )
(1 − 𝑟 ) + 𝑖2𝜁𝑟 (1 − 𝑟 ) − 𝑖2𝜁𝑟 (1 − 𝑟 ) + (2𝜁𝑟) (1 − 𝑟 )2 + (2𝜁𝑟)2
2

The magnitude of it can be find from: √𝐴2 + 𝐵 2

2 2
(1 − 𝑟 2 ) 2𝜁𝑟 1
√( ) + ( ) =
(1 − 𝑟 2 )2 + (2𝜁𝑟)2 (1 − 𝑟 2 )2 + (2𝜁𝑟)2 √[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ]

The Frequency response function 𝐻(𝑟) can give us lots of information. In this part, we will talk
about them one by one.

1
|𝐻(𝑟)| =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]

1) Undamped system (𝜁 = 0)
𝑟=0 𝐻(𝑟) = 1
𝑟=1 𝐻(𝑟) = ∞

𝑟 = √2 𝐻(𝑟) = 1
𝑟= ∞ 𝐻(𝑟) = 0

3
 ME 4440-5540 Lecture 21

2) For Damped system (𝜁 ≠ 0), the Frequency response function 𝐻(𝑟) is dependent on two
parameters 𝜁 & 𝑟.

2a) For (0 < 𝜁 < 0.707)

The maximum value of the curves of Frequency response function 𝐻(𝑟) will be change between
0-1 and as much as the value of 𝜁 increases, the maximum of the curve will be shift more to the
left of the line 𝑟 = 1 on the graph.
Frequency response function 𝑯(𝒓)

𝟏
𝑯(𝒓) =
𝟐𝜻

2b) For (0.707 < 𝜁)

The maximum value of the curves of Frequency response function 𝐻(𝑟) for all value of 𝜁 will be
on 𝑟 = 0!
Important Note: Only for 𝟎 < 𝜻 < 𝟎. 𝟕𝟎𝟕 for any system, the maximum dynamic response can
be larger than corresponding static response! For 𝟎. 𝟕𝟎𝟕 < 𝜻, always dynamic response would
be less than corresponding static response (except at 𝑟 = 0 which means we have a static case).
So, in order to decrease the effect of dynamic motion, you need to increase the damping which is
not always a good idea because the consequence of it would be waste of energies, generating
heat, etc.

4
 ME 4440-5540 Lecture 22

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Mathematical analysis of the graphs:

For finding the maximum value of the curves, we have to find the roots of “𝑟” for derivative of
𝐻(𝑟) equal to zero. For constant value of damping we will have:
𝑑|𝐻(𝑟)| 𝑟(1−𝑟 2 −2𝜁 2 )
=0 3 =0 𝑟(1 − 𝑟 2 − 2𝜁 2 ) = 0
𝑑𝑟 [(1−𝑟 2 )2 +4𝜁 2 𝑟 2 ] ⁄2

So,
𝑟=0 or (1 − 𝑟 2 − 2𝜁 2 ) = 0
1) For 𝑟 = 0 can be maximum or minimum of the graph. For finding if it is the maximum or
minimum, we need to take the second derivative of 𝐻(𝑟) to see if it is negative (maximum) or
positive (minimum)!
𝑑2 𝐻(𝑟)
<0 Maximum
𝑑𝑟 2

𝑑2 𝐻(𝑟)
>0 Minimum
𝑑𝑟 2

If you take the second derivative and put 𝑟 = 0, you will find:
𝑑2 𝐻(𝑟=0)
= 1 − 2𝜁 2
𝑑𝑟 2

√2 𝑑2 𝐻(𝑟=0)
For 𝜁 > ≅ 0.707 we will have <0 Maximum
2 𝑑𝑟 2

√2 𝑑2 𝐻(𝑟=0)
For 𝜁 < ≅ 0.707 we will have >0 Minimum
2 𝑑𝑟 2

Note: Therefore, for all systems with 𝜁 > 0.707, the dynamic response will never get to
𝐹
corresponding static displacement ( 𝑘0 ).

2) For (1 − 𝑟 2 − 2𝜁 2 ) = 0 (𝑟 ≥ 0 always):
1 1 1
𝑟 = √1 − 2𝜁 2 𝐻(𝑟) = = =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] √[(1−1+2𝜁 2 )2 +4𝜁 2 (1−2𝜁 2 )] 2𝜁√1−𝜁 2

For 𝜻 ≤ 𝟎. 𝟕𝟎𝟕 1
𝜁 2 become very small number √1 − 𝜁 2 ≈ 1 |𝐻(𝑟)|𝑚𝑎𝑥 ≈
2𝜁

1
So, the maximum value of 𝐻(𝑟) for 𝜻 ≤ 𝟎. 𝟕𝟎𝟕 would be almost equal to (2𝜁 )! Therefore, if you
find and plot 𝐻(𝑟) for a system from experiment and measure the height of the maximum
1
of the function, then you can calculate and find 𝜁 from (2𝜁 ). So, one of the benefit of
frequency response function 𝐻(𝑟) is finding damping ratio. 𝐻(𝑟)𝑚𝑎𝑥 is named “Quality

1
 ME 4440-5540 Lecture 22

Factor” and is showing with “Q”. High Q system means systems with low damping and low
Q systems means systems that have high damping.

Example 26: The following figure showing a structure which is hinged to the floor and supported
rotating machinery that exerts a force F= 900 cos 5t N. This system has a damping ratio of 𝜁 =
0.05 and mass of 𝑚 = 6800𝑘𝑔. For the beams in this structure, = 207 × 109 𝑃𝑎 , 𝐼 = 2.88 ×
𝐼
10−5 𝑚4 , 𝑙 = 5𝑚 and section modulus (𝑐) = 35 × 10−5 𝑚3.
a) Find the maximum steady-state displacement.
b) Find the maximum dynamic bending stress.

a)
First let find the equivalent of stiffness for the beams. The beams are parallel, so:
3𝐸𝐼 3𝐸𝐼 207×109 ×2.88×10−5 𝑁
𝑘= 𝑘𝑡 = 2 × = 2×3× = 286.157 × 103 𝑚
𝑙3 𝑙3 125

2
 ME 4440-5540 Lecture 22

Then find the equivalent static displacement (𝐴𝑠𝑡 )

𝐹0 900
𝐴𝑠𝑡 = = = 0.003𝑚
𝑘𝑡 286.157 × 103
Now, we can calculate the amplitude or maximum steady-state displacement from following
equation:
𝐹0
𝑈𝑚𝑎𝑥 = |𝐻(𝑟)|. 𝐴𝑠𝑡 = 𝑘
√[(1 − 𝑟 )2 + (2𝜁𝑟)2 ]
2

286.157×103 𝑟 𝑟 𝜔 5
𝜔𝑛 = √ = 6.5 & 𝜔 = 5𝑠 𝑟 = 𝜔 = 6.5 = 0.77
6800 𝑠 𝑛

0.003𝑚
𝑈𝑚𝑎𝑥 = = 0.007𝑚
√[(1 − (0.77)2 )2 + (2 × 0.05 × 0.77)2 ]

As you can see the maximum steady-state dynamic displacement is larger than static
displacement in this case.

𝑈𝑚𝑎𝑥 > 𝐴𝑠𝑡

3
 ME 4440-5540 Lecture 22

b)
What is this force on the beam? Because the force is perpendicular to the beam, so it is a shear
force (V)!

Maximum shear force can be find from following equation:

3𝐸𝐼
𝐹𝑚𝑎𝑥 = 𝑉𝑚𝑎𝑥 = 𝑘 × 𝑈𝑚𝑎𝑥 = × 𝑈𝑚𝑎𝑥
𝑙3

𝑉𝑚𝑎𝑥 = 143.078 × 103 × 0.007 = 1001.5 𝑁

Note: From mechanic of material: 𝑀𝑚𝑎𝑥 = 𝑉𝑚𝑎𝑥 × 𝑙

𝑀𝑚𝑎𝑥 = 𝑉𝑚𝑎𝑥 × 𝑙 = 1001.5 × 5 = 5007.5 𝑁. 𝑚

4
 ME 4440-5540 Lecture 22

𝑀×𝑦 𝑀×𝑐
From mechanics of material bending stress can be find from 𝜎 = =
𝐼 𝐼

Cross section of beam

Maximum bending stress would be equal to:

𝑀𝑚𝑎𝑥 5007.5
𝜎𝑚𝑎𝑥 = = = 1.43 ∗ 10^7 𝑃𝑎
𝐼/𝑐 35 × 10−5

5
 ME 4440-5540 Lecture 21

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Rotating Unbalance

Rotating unbalance damped system is one of the special cases of the harmonically excited
mechanical systems. In general, most of the systems are subjected to some kind of harmonic
excitation (e.g. engine of the cars, jet engines, rotating machineries, etc.). If the center of mass
and center of rotation do not coincide, that will be the rotating unbalance case. To know about
importance of this topic, for instance the breakage of only one turbine blade of a jet engine
causes a huge rotating unbalance and causes the total failure of the system!

Example 27: The following figure is showing a supported beam that holding a rotating
machinery (has a rotating component). The rotating part has a frequency of 𝜔. There is a small
eccentric mass (𝑚𝑒 ) on the rotating part when the center of rotation and center of this mass are
not coincide (they have a distance of “𝑒”). The total mass of the rotating part is (𝑚 − 𝑚𝑒 ). (The
beam is assumed massless)

Without (𝑚𝑒 ) this example was a harmonic excited damped system which we studied before.
This beam can be shown with spring-damp system, where the 𝑘 is the stiffness of the beam and 𝑐
is given.

Beam

So, whole structure can be model like this:

1
 ME 4440-5540 Lecture 21

In this case additional to the rotation part which experience a displacement of 𝑢(𝑡), (𝑚𝑒 ) is
experiences a displacement of 𝑑 = 𝑢(𝑡) + 𝑒 𝑠𝑖𝑛 𝜔𝑡

Free body diagram would be:

Write the equation of motion for the system:

𝑑2
(𝑚 − 𝑚𝑒 )𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 + 𝑚𝑒 (𝑢 + 𝑒 𝑠𝑖𝑛 𝜔𝑡) = 0
𝑑𝑡 2
𝑚𝑢̈ − 𝑚𝑒 𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 + 𝑚𝑒 𝑢̈ − 𝑚𝑒 𝑒𝜔2 𝑠𝑖𝑛 𝜔𝑡 = 0

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑚𝑒 𝑒𝜔2 𝑠𝑖𝑛 𝜔𝑡 Equation of motion

There are some observation from the equation of motion for rotating unbalanced system.
𝑚𝑒 𝑒𝜔2 = constant, so the right side of this equation is very similar to 𝐹0 sin 𝜔𝑡. We can say that
is a harmonically excited system with amplitude of forcing function equal to (𝑚𝑒 𝑒𝜔2 ) or the
amplitude of the forcing function is dependent on the frequency (𝜔).

If we write 𝐹0 = 𝑚𝑒 𝑒𝜔2 , then 𝑈𝑚𝑎𝑥 would be:

2
 ME 4440-5540 Lecture 21

𝐹0 𝑚𝑒 𝑒𝜔2
𝑈𝑚𝑎𝑥 = 𝑘 = 𝑘
√[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ] √[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ]

Or
𝑚𝑒 𝑒𝜔2 𝜔𝑛 2 𝑚 𝑒
× 2 ( 𝑚𝑒 ) 𝑟 2
𝑘 𝜔𝑛
𝑈𝑚𝑎𝑥 = =
√[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ] √[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ]

If we compare the results with a damped harmonic excited system:

For a general harmonically excited damped system:

Forcing function: 𝑓(𝑡) = 𝐹0 𝑠𝑖𝑛 𝜔𝑡

1
Frequency response function: |𝐻(𝑟)| =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]
𝐹0
Maximum steady-state response: 𝑈𝑚𝑎𝑥 = |𝐻(𝑟)| × 𝑘

For a rotating unbalance damped system:

Forcing function: 𝑓(𝑡) = 𝑚𝑒 𝑒𝜔2 𝑠𝑖𝑛 𝜔𝑡 (The amplitude of forcing function is frequency
dependent in this case)
𝑟2
Frequency response function: |𝐻(𝑟)| =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]
𝑚𝑒 𝑒𝜔2 𝑚 𝑒
( 𝑒 )𝑟 2
𝑘 𝑚
Maximum steady-state response: 𝑈𝑚𝑎𝑥 = =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] √[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]

General harmonically Rotating unbalance

excited damped system damped system

𝜁 < 0.707 𝜁 < 0.707

3
 ME 4440-5540 Lecture 21

As you can see from the graphs, the general harmonically excited damped system & the rotating
unbalance damped system are completely different.
In the rotating unbalance damped system:
1) In this case, the peaks of the functions instead of being on the left of the r=1 are on the right
of the r=1. So, if the peak is on the right of r=1, that means there is an unbalance in the
system! That is how they find the unbalance in the tires/wheels of vehicle.
2) As damping ratio (𝜁) increases, the peaks are shift further to the right of the 𝑟 = 1.
𝑟2
3) If 𝑟 → ∞ |𝐻(𝑟)| → = 1 , This means (for 𝜁 < 0.707 & 𝑟 > 1) there is no way under
𝑟2
rotation unbalance system that you can get below the maximum static displacement! At the
time that the forcing frequency become so much larger then natural frequency of system
(𝜔 ≫ 𝜔𝑛 ), at best you would approach the static displacement!
4) For 𝑟 = 0 |𝐻(𝑟)| = 0 (no matter what is the value of 𝜁)

Exercise: Find the maximum values for the peaks on the rotating unbalance damped system and
show why they are shifting to the right of 𝑟 = 1 with increasing value of 𝜁 (for 𝜁 < 0.707).
Example 28: The following figure is showing a supported beam that holding a rotating machinery
(laundry machine). The rotating part has a frequency of 𝜔 and mass of 𝑚 =7250kg. For the beam
in this structure, = 207 × 109 𝑃𝑎 , 𝐼 = 5 × 10−5 𝑚4 , 𝑙 = 3.5𝑚. The motor has a speed of 300
rpm. There is a small eccentric mass (𝑚𝑒 = 20𝑘𝑔) on the rotating part that the center of rotation
and center of this mass are not coincide (they have a distance of 𝑒 = 0.25𝑚). The damping ratio
of the system is 𝜁 = 10%. The beam is assumed massless. Find the maximum displacement of the
system 𝑈𝑚𝑎𝑥 ?

𝑈𝑚𝑎𝑥 =?
𝑚𝑒 𝑒𝜔2
𝑘
𝑈𝑚𝑎𝑥 =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]
𝑟 =?

4
 ME 4440-5540 Lecture 21

2𝜋 𝑟𝑎𝑑
𝜔 = 300 𝑟𝑝𝑚 × 60 = 31.4 𝑠
48 𝐸𝐼 48×207×109 × 5×10−5 49680×104 𝑁
𝑘= = = = 11.6 × 106 𝑚
𝑙3 (3.5)3 42.875
𝑘 11.6×106 𝑟𝑎𝑑
𝜔𝑛 = √𝑚 = √ = 40
7250 𝑠

𝜔 31.4
𝑟=𝜔 = = 0.785
𝑛 40

𝑚𝑒 𝑒𝜔2 20×0.25×31.42
𝑘 11.6×106 4.25×10−4
𝑈𝑚𝑎𝑥 = = = = 0.001 𝑚
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] √[(1−(0.785 )2 )2 +(2×0.1×0.785 )2 ] 0.414

So, for just a very small unbalance mass compare to the total mass causes 1mm displacement!
20
(7250 × 100 = 0.27% of total weight)

5
 ME 4440-5540 Lecture 24

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Time Dependent Input Displacement

This is a special case of damped (or Undamped) systems and instead of system be subjected to
an external force, the base of system is subjected to a displacement (e.g. earthquake). Input
displacement doesn’t have to be necessarily harmonic but here we are discussing the harmonic
and later we will discuss a general case.

𝑧 =𝑢+𝑦

We can write the equation of motion in term of “u”:

𝑚𝑧̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0 𝑚𝑢̈ + 𝑚𝑦̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0 𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = −𝑚𝑦̈

This way would be good for civil engineers because usually the earthquake input measure
respect to the ground moving acceleration.

Also, we can write the equation of motion in term of “z”:

𝑢 =𝑧−𝑦

𝑚𝑧̈ + 𝑐𝑢̇ + 𝑘𝑢 = 0 𝑚𝑧̈ + 𝑐𝑧̇ − 𝑐𝑦̇ + 𝑘𝑧 − 𝑘𝑦 = 0 𝑚𝑧̈ + 𝑐𝑧̇ + 𝑘𝑧 = 𝑐𝑦̇ + 𝑘𝑦

This way would be good for mechanical cases where usually we are dealing with an input
displacement.
If we want to write these equations as the form of the forced vibration, we will have:

1
 ME 4440-5540 Lecture 24

−𝑚𝑦̈
𝑓(𝑡) = { or
𝑐𝑦̇ + 𝑘𝑦
Example 29: Let’s consider that a car is travelling on a road with roughness in shape of sinusoidal
𝑥
function. The amplitude of the roughness for the surface of the road is 𝑦 = 𝑌 sin 2𝜋 𝑙 (“𝑙” is one
period for the wave of the road). The car is modeled as an Undamped SDOF system (𝑐 = 0). This
car travelling in 𝑥 direction with constant speed “𝑣”. Find the most undesirable speed.

When the most undesirable speed will be happened? In case of resonance (When the input
frequency (𝜔) and natural frequency of the system (𝜔𝑛 ) become equal and 𝑟 = 1)

The equation of motion would be:

𝑥
𝑚𝑧̈ + 𝑘𝑧 = 𝑘𝑦 = 𝑘𝑌 sin 2𝜋
𝑙
𝑥 2𝜋𝑣𝑡 2𝜋𝑣
Velocity is constant 𝑥 = 𝑣. 𝑡 2𝜋 𝑙 = = 𝜔𝑡 𝜔=
𝑙 𝑙

𝑘 𝑘𝑔
𝜔𝑛 = √ = √
𝑚 𝑊

𝑘𝑔 2𝜋𝑣 𝑙 𝑘𝑔
𝜔𝑛 = 𝜔 √ = 𝑣 = 2𝜋 √ 𝑊 The most undesirable speed
𝑊 𝑙

2
 ME 4440-5540 Lecture 24

Example 30: The following system is supported at both sides with walls. The wall at the right side
𝑁
subjected to a displacement 𝑦(𝑡) = 0.1 𝑠𝑖𝑛 𝑡. In this system, 𝑚 = 1 𝑘𝑔, 𝑘 = 1 𝑚,
𝑐2 = 2 𝑁 − 𝑠⁄𝑚, and 𝜁 = 0.2. Find the maximum displacement that the mass experience (𝑈𝑚𝑎𝑥 ).

Free body diagram:

Equation of motion:
−𝑚𝑢̈ + 𝑐2 (𝑦̇ − 𝑢̇ ) − 𝑘𝑢 − 𝑐1 𝑢̇ = 0

𝑚𝑢̈ + (𝑐1 + 𝑐2 )𝑢̇ + 𝑘𝑢 = 𝑐2 𝑦̇

We don’t have value of 𝑐1 but damping ratio is given (𝜁 = 0.2). So, we can divide both side of
equation by “𝑚” and we will have:

𝑐 𝑐
(Note: 𝑐𝑡𝑜𝑡𝑎𝑙 = 𝑐1 + 𝑐2 𝜁 = 2𝑚𝜔 = 2𝜁𝜔𝑛 )
𝑛 𝑚

𝑐2
𝑢̈ + 2𝜁𝜔𝑛 𝑢̇ + 𝜔𝑛 2 𝑢 = 𝑦̇
𝑚
𝑘 1
𝜔𝑛 = √𝑚 = √1 = 1 , 𝜁 = 0.2 , 𝑐2 = 2 , 𝑚 = 1 , 𝑦 = 0.1 𝑠𝑖𝑛 𝑡 𝑦̇ = 0.1 𝑐𝑜𝑠 𝑡

3
 ME 4440-5540 Lecture 24

𝑢̈ + 0.4𝑢̇ + 1𝑢 = 0.2 𝑐𝑜𝑠 𝑡 Equation of motion

𝑓(𝑡) = 0.2 𝑐𝑜𝑠 𝑡

𝐹0

𝐹0 𝐹0 1 0.2 1
𝑈𝑚𝑎𝑥 = |𝐻(𝑟)|𝑚𝑎𝑥 = × = ×
𝑘 𝑘 √[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ] 1 √[(1 − 1)2 + (2 × 0.2 × 1)2 ]

0.2
𝑈𝑚𝑎𝑥 = = 0.5 𝑚
0.4

4
 ME 4440-5540 Lecture 25

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Transmissibility Ratio (TR)

Till now we discussed about systems which are subjected to a force or a displacement. But we
want to know, what is the effect of this force or displacement on our system to protect the system
from undesirable vibrations.

Transmissibility Ratio (TR): It is a quantity that measures the impact of the Force or Displacement
acting on a system.
Case 1: Force Transmitted
If we have a force acting on the system, it causes the system displacement, so the structure supports
which are “k” and “c” as a result of that displacement would be subjected the transmitted force
(𝑓𝑇 ).

𝑓𝑇 = 𝑘𝑢 + 𝑐𝑢̇

Structure supports

In the case of force acting on the system, we define Transmissibility Ratio (TR) as:

𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝑇𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 𝑘𝑢 + 𝑐𝑢̇

𝑇𝑅 = =
𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝐹0
For a harmonic force:

𝑓(𝑡) = 𝐹0 𝑠𝑖𝑛 𝜔𝑡
𝐹0
𝑢(𝑡) = 𝐻(𝑟) 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃)
𝑘
𝐹0
𝑢̇ (𝑡) = 𝐻(𝑟) 𝜔 𝑐𝑜𝑠 (𝜔𝑡 − 𝜃)
𝑘
𝐹0 𝐹0
𝑓𝑇 = 𝑘 × 𝑢𝑚𝑎𝑥 + 𝑐 × 𝑢̇ 𝑚𝑎𝑥 = 𝑘 × 𝐻(𝑟) 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃) + 𝑐 × 𝐻(𝑟) 𝜔 𝑐𝑜𝑠 (𝜔𝑡 − 𝜃)
𝑘 𝑘
𝑐𝜔
𝑓𝑇 = 𝐹0 𝐻(𝑟) 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃) + 𝐹0 𝐻(𝑟) 𝑐𝑜𝑠 (𝜔𝑡 − 𝜃)
𝑘
𝐴1 𝐴2

1
 ME 4440-5540 Lecture 25

𝐴1 𝐴2
𝑓𝑇 = √𝐴1 2 + 𝐴2 2 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃) + 𝑐𝑜𝑠 (𝜔𝑡 − 𝜃)
2 2 2 2
√ √𝐴1 + 𝐴2
( 𝐴1 + 𝐴2 )

(𝑓𝑇 )𝑚𝑎𝑥

𝑐2𝜔2
(𝑓𝑇 )𝑚𝑎𝑥 = 𝐹0 𝐻(𝑟)√1 +
𝑘2

𝑐 2𝜔2
(𝑓𝑇 )𝑚𝑎𝑥 𝐹0 𝐻(𝑟)√1 + 𝑘 2 𝑐2𝜔2
𝑇𝑅 = = = 𝐻(𝑟)√1 + 2
𝐹0 𝐹0 𝑘

1
𝐻(𝑟) =
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]

𝑐 2 𝜔2 𝜔 2 4 𝑐2
× 𝜔𝑛2 × 4 = 4 4𝑘𝑚 × 𝑟 2 = 4𝜁 2 𝑟 2
𝑘2 𝑛

𝑐 𝜁2
Note: 𝜁 = 2√𝑘𝑚

1+4𝜁 2 𝑟 2
𝑇𝑅 = √(1−𝑟 2 )2 +(2𝜁𝑟)2

This is showing the ratio of maximum force that the supports of a system experience to the
maximum harmonic force that acts on the system.

The following plot showing the Force Transmissibility Ratio function. We care about the case
𝜁 < 0.707 because for 𝜁 > 0.707 we know the dynamic displacement is always less than static
displacement.

2
 ME 4440-5540 Lecture 25

Transmissibility Ratio (TR)

Following observation from this plot:

- All curves go through 𝑟 = √2 and intersect on that point

- 𝑇𝑅 < 1 Only if 𝑟 > √2, That means in order to reduce the unpleasant effect of
a forcing function on your system, you have to make sure always 𝜔 > √2 𝜔𝑛
- For 𝑟 > √2 as damping ratio increases, TR increases too. That means, we want to have a
system less damped in order to less force transmitted to the system.
- For having 𝑟 > √2 , the best way is having system with smaller natural frequency
𝑘
(𝜔𝑛 = √𝑚). Because in most of the cases, we don’t have much control over the mass so
we need system with smaller stiffness (𝑘). So, softer system better absorb the force but
𝑊
then you will have larger static displacement (𝛿 = 𝑘 )! Therefore, vibration isolation is a
balancing act that you need to find out what is your optimum stiffness (𝑘) value.

Example 31: The following figure is showing a supported beam that holding a rotating machinery
with a weight of 𝑀 = 500 𝑘𝑔 and frequency of 𝜔 = 7200 𝑟𝑝𝑚. This system has an unbalanced
mass with 𝑚𝑒 = 1 𝑔 and 𝑒 = 20 𝑐𝑚.The damping ratio of the system is negligible (𝜁 = 0). Design
an isolation system that assures the force transmitted (𝑓𝑇 ) is less than 250N?

3
 ME 4440-5540 Lecture 25

𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝑇𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 1

𝜁=0 𝑇𝑅 = =| 2 |
𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝑟 −1

𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝐴𝑝𝑝𝑙𝑖𝑒𝑑: This an unbalanced system, so, maximum applied force would be:

𝐹0 = 𝑚𝑒 𝑒𝜔2

𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝑇𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑: That is given by problem equal to 250 N.

So, we will have:

250 1
2
=| 2 |
𝑚𝑒 𝑒𝜔 𝑟 −1

2𝜋 𝑟𝑎𝑑
𝜔 = 7200 𝑟𝑝𝑚 × 60 = 754 , 𝑚𝑒 = 0.001 kg , 𝑒 = 0.2 𝑚
𝑠

250 250 1
= 113.7 = 2.2 = |𝑟 2 −1| 𝑟 = 1.2
0.001×0.2×(754)2

𝜔 754 𝑟𝑎𝑑
𝑟=𝜔 1.2 = 𝜔𝑛 = 628.3
𝑛 𝜔𝑛 𝑠

So, for having force transmitted (𝑓𝑇 ) less than 250N, the support system (beam) need to have a
𝑟𝑎𝑑
natural frequency of 𝜔𝑛 = 628.3 . Based on this natural frequency, we can find the stiffness
𝑠
of the system.

𝑘 𝑘 𝑁
𝜔𝑛 = √𝑚 628.3 = √500 𝑘 = 1.9 × 108 𝑚

However, we have to make sure, system with this stiffness not make a large static displacement.

𝑊 500 × 9.81
𝛿= = = 2.6 × 10−5 𝑚
𝑘 1.9 × 108

4
 ME 4440-5540 Lecture 25

The displacement is very small and almost negligible. That means this stiffness (𝑘) is acceptable
for the system.

5
 ME 4440-5540 Lecture 26

A.2) Analyzing a Damped System That Excited by a Harmonic Force

Example 32: The following figure is showing a supported beam that holding a cylinder-piston
𝑟𝑎𝑑
machinery with a weight of 𝑀 = 1750 𝑘𝑔 and frequency of 𝜔 = 60 . The properties of the
𝑠
9 −5 4
beam are, = 207 × 10 𝑃𝑎 , 𝐼 = 5 × 10 𝑚 , 𝑙 = 3 𝑚. This machinery generate a harmonic
force with amplitude 𝐹0 = 32000 𝑁. The damping ratio of the system is 𝜁 = 10%.
a) Find the amplitude (maximum displacement) of the motion
b) Find the force transmitted to the beam

a)
This is simply supported beam, so:

48 𝐸𝐼 48 × 207 × 109 × 5 × 10−5 𝑁

𝑘= 3
= = 18.4 × 106
𝑙 27 𝑚

𝑘 18.4 × 106 𝑟𝑎𝑑

𝜔𝑛 = √ = √ = 102.5
𝑚 1750 𝑠

𝜔 60
𝜁 = 0.1 , 𝑟 = 𝜔 = 102.5 = 0.585
𝑛

𝐹 𝐹 1 32000 1
𝑈𝑚𝑎𝑥 = ( 𝑘0 ) 𝐻(𝑟) = ( 𝑘0 ) = (18.4×106 ) = 2.6 × 10−3 𝑚
√[(1−𝑟 2 )2 +(2𝜁𝑟)2 ] √[(1−0.5852 )2 +(2×0.1×0.585)2 ]

b)

1+4𝜁 2 𝑟 2 1+4×0.12 ×0.62

𝑇𝑅 = √(1−𝑟 2 )2 +(2𝜁𝑟)2 = √(1−0.62)2 +(2×0.1×0.6)2 = 1.547

𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝑇𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑 (𝑓𝑇 )𝑚𝑎𝑥

𝑇𝑅 = =
𝑀𝑎𝑥 𝐹𝑜𝑟𝑐𝑒 𝐴𝑝𝑝𝑙𝑖𝑒𝑑 𝐹0

(𝑓𝑇 )𝑚𝑎𝑥 = 𝑇𝑅 × 𝐹0 = 1.547 × 32000 = 49504 𝑁

1
 ME 4440-5540 Lecture 26

As you can see, this machinery with weight of 1750 𝑘𝑔 (≅ 17500𝑁), if it operates at the
𝑟𝑎𝑑
frequency of 𝜔 = 60 , the system experiences almost three times higher force than the weight
𝑠
of machine!

Case 2: Input Displacement Transmitted

𝑀𝑎𝑥 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑇𝑟𝑎𝑛𝑠𝑚𝑖𝑡𝑡𝑒𝑑

𝑇𝑅 =
𝑀𝑎𝑥 𝐼𝑛𝑝𝑢𝑡 𝐷𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = −𝑚𝑦̈ Equation of motion

Let’s assume the system experiences a harmonic displacement:

𝑦 = 𝑌 sin 𝜔𝑡

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑚𝜔2 𝑌 sin 𝜔𝑡

Maximum Input
Displacement (𝐹0 )

If we solve for 𝑢(𝑡), we will have:

𝐹0 𝑚𝜔2 𝑌
𝑢(𝑡) = ( ) 𝐻 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃) = ( ) 𝐻 𝑠𝑖𝑛 (𝜔𝑡 − 𝜃)
𝑘 𝑘

The absolute displacement would be:

Absolute displacement = 𝑢(𝑡) + 𝑦(𝑡)

If you are going to the same process as we did for the force transmitted case, you can see the
transmissibility ratio will be exactly same as previous case and it has an identical expression!

1+4𝜁 2 𝑟 2
𝑇𝑅 = √(1−𝑟 2 )2 +(2𝜁𝑟)2

Note: The goal is keep the “𝑇𝑅” less than “1” for the systems!

Half Power Method

2
 ME 4440-5540 Lecture 26

1
If you remember, we used 𝐻(𝑟) plot and quality (Q) (which was equal to 𝐻(𝑟)𝑚𝑎𝑥 = 2𝜁 ) to find
damping ratio (𝜁). The “Half Power Method” is another approach for finding the damping ratio
(𝜁).

The frequency response function is not smooth like that is showing in the graphs and usually it is
full of noises. So, finding the exact maximum value of the graph would be hard and using
estimation is not very accurate.

𝐻𝑚𝑎𝑥
Let’s assume we can find and draw a line to cross the graph (where the graph is smoother).
√2
Find the frequency ratio (r) for two intersection points. In this method, we are using these (r) values
to calculate and find the peak of the graph and damping ratio.

For finding 𝑟1 & 𝑟2 , we can solve the following equation and roots of that equation will be 𝑟1 &
𝑟2 values:

1 1 1 1
𝐻𝑚𝑎𝑥 = 2𝜁 × 2𝜁 =
√2 √[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]

𝑟1 2 = 1 − 2𝜁 2 − 2𝜁√1 + 𝜁 2
𝑟2 2 = 1 − 2𝜁 2 + 2𝜁√1 + 𝜁 2

We know the damping is very small (𝜁 ≪ 1), so, 𝑟1 & 𝑟2 can write as:

𝑟1 2 ≅ 1 − 2𝜁 2 − 2𝜁
𝑟2 2 ≅ 1 − 2𝜁 2 + 2𝜁

With a good approximation 𝑟1 & 𝑟2 will be equal to:

3
 ME 4440-5540 Lecture 26

𝑟1 ≅ 1 − 𝜁 2 − 𝜁
𝑟2 ≅ 1 − 𝜁 2 + 𝜁

The value of 𝜁 can be find from subtracting these two equations from each other.

𝑟2 − 𝑟1 = 2𝜁

1
𝜁 = (𝑟2 − 𝑟1 )
2

4
 ME 4440-5540 Lecture 27

B) Periodic Force

In the real life there are many cases that system is subjected to a periodic force. In reality a true
harmonic excitation is very rare and mostly exist and study on the laboratories.

In reality you can see many situations that the forcing function is periodic and repeated by period
of “T” but it is not harmonic!

Harmonic Function

Periodic Functions

T
T

For any periodic functions, we can find a proper Fourier series (include summation of large number
of harmonic terms) that reproduce the periodic function. In other words, a periodic function can
be find by summation of lots of harmonic functions as a Fourier series.

2𝜋
In general, if we have a periodic function with period of (T) and frequency of 𝜔 = , Fourier
𝑇
series represent this function can be write as follow:

𝑓(𝑡) = 𝑎0 + 𝑎1 𝑐𝑜𝑠 𝜔𝑡 + 𝑎2 𝑐𝑜𝑠 2𝜔𝑡 + 𝑎3 𝑐𝑜𝑠 3𝜔𝑡 + ⋯ + 𝑎𝑛 𝑐𝑜𝑠 𝑛𝜔𝑡 + 𝑏1 𝑠𝑖𝑛 𝜔𝑡 + 𝑏2 𝑠𝑖𝑛 2𝜔𝑡
+ 𝑏3 𝑠𝑖𝑛 3𝜔𝑡 + ⋯ + 𝑏𝑛 𝑠𝑖𝑛 𝑛𝜔𝑡

For any periodic function, we only need to find all proper coefficients for above Fourier series (𝑎0 ,
𝑎1 , 𝑎2 , …, 𝑎𝑛 , 𝑏1 , 𝑏2 , …, 𝑏𝑛 ). Then, that periodic function would be converted to a summation of
series of harmonic functions.

1
 ME 4440-5540 Lecture 27

This Fourier series can be write as:

𝑓(𝑡) = 𝑎0 + ∑[𝑎𝑛 𝑐𝑜𝑠 𝑛𝜔𝑡 + 𝑏𝑛 𝑠𝑖𝑛 𝑛𝜔𝑡 ]

𝑛=1

For the Fourier series, 𝑎0 , 𝑎𝑛 , 𝑏𝑛 would be:

1 𝑇
𝑎0 = 𝑇 ∫0 𝑓(𝑡) 𝑑𝑡

2 𝑇
𝑎𝑛 = 𝑇 ∫0 𝑓(𝑡)𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡

2 𝑇
𝑏𝑛 = 𝑇 ∫0 𝑓(𝑡)𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡

So, for any problem, only we need to insert any given periodic function and period of (𝑇) in these
equations, find these coefficients and apply them in Fourier series.

Example 33: The following graph showing a periodic function. Find the proper Fourier series
which is representing this function.

Step1: The first step in solving a problem like this is finding the 𝑓(𝑡) from the given graph.

1 0 < 𝑡 < 𝑇/2

𝑓(𝑡) = {
−1 𝑇/2 < 𝑡 < 𝑇
Step2: find the coefficients

As this function is discontinues, we have to break the integral.

2
 ME 4440-5540 Lecture 27

1 𝑇 1 𝑇⁄ 𝑇 1 𝑇 𝑇
𝑎0 = 𝑇 ∫0 𝑓(𝑡) 𝑑𝑡 = 𝑇 [∫0 2 1 𝑑𝑡 + ∫𝑇⁄ −1 𝑑𝑡] = 𝑇 [2 − (𝑇 − 2)] = 0
2

2 𝑇 2 𝑇⁄ 𝑇
𝑎𝑛 = 𝑇 ∫0 𝑓(𝑡)𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡 = 𝑇 [∫0 2 1. 𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡 + ∫𝑇⁄ −1 . 𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡]
2

2 1 𝑇 1 𝑇
𝑎𝑛 = [ 𝑠𝑖𝑛 𝑛𝜔𝑡| ⁄2 − 𝑠𝑖𝑛 𝑛𝜔𝑡|𝑇⁄ ]
𝑇 𝑛𝜔 0 𝑛𝜔 2

2 1 𝑇 1 𝑇 2 𝑇
𝑎𝑛 = 𝑇 [𝑛𝜔 𝑠𝑖𝑛 𝑛𝜔 2 − 𝑛𝜔 (𝑠𝑖𝑛 𝑛𝜔𝑇 − 𝑠𝑖𝑛 𝑛𝜔 2)] = 𝑛𝜔𝑇 [2 𝑠𝑖𝑛 𝑛𝜔 2 − 𝑠𝑖𝑛 𝑛𝜔𝑇]

2 2𝜋 𝑇 2𝜋
𝑎𝑛 = 𝑛𝜔𝑇 [2 𝑠𝑖𝑛 𝑛 − 𝑠𝑖𝑛 𝑛 𝑇]
𝑇 2 𝑇 0
4 0 2
𝑎𝑛 = 𝑠𝑖𝑛 𝑛𝜋 − 𝑠𝑖𝑛 2𝑛𝜋 = 0
𝑛𝜔𝑇 𝑛𝜔𝑇

2 𝑇 2 𝑇⁄ 𝑇
𝑏𝑛 = 𝑇 ∫0 𝑓(𝑡)𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡 = 𝑇 [∫0 2 1. 𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡 + ∫𝑇⁄ −1 . 𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡]
2

𝑇 𝑇
𝑏𝑛 = 𝑇 [− 𝑛𝜔 𝑐𝑜𝑠 𝑛𝜔𝑡| ⁄2 + 𝑛𝜔 𝑐𝑜𝑠 𝑛𝜔𝑡|𝑇⁄ ]
2 1 1

0 2

2 1 𝑇 1 1 1 𝑇
𝑏𝑛 = 𝑇 [− 𝑛𝜔 𝑐𝑜𝑠 𝑛𝜔 2 + 𝑛𝜔 + 𝑛𝜔 𝑐𝑜𝑠 𝑛𝜔𝑇 − 𝑛𝜔 𝑐𝑜𝑠 𝑛𝜔 2]

2 𝑇 2 2𝜋 𝑇 2𝜋
𝑏𝑛 = 𝑛𝜔𝑇 [−2 𝑐𝑜𝑠 𝑛𝜔 2 + 1 + 𝑐𝑜𝑠 𝑛𝜔𝑇] = 2𝜋 [−2 𝑐𝑜𝑠 𝑛 + 1 + 𝑐𝑜𝑠 𝑛 𝑇]
𝑛 𝑇 𝑇 2 𝑇
𝑇

1
𝑏𝑛 = 𝑛𝜋 [−2 𝑐𝑜𝑠 𝑛𝜋 + 1 + 𝑐𝑜𝑠 2𝑛𝜋]

𝑏𝑛 = 0 𝑛 = 𝑒𝑣𝑒𝑛 𝑛𝑢𝑚𝑏𝑒𝑟𝑠 (2,4,6, … )

𝑏𝑛 = { 4
𝑏𝑛 = 𝑛 = 𝑜𝑑𝑑 𝑛𝑢𝑚𝑏𝑒𝑟𝑠 (1,3,5, … )
𝑛𝜋

Step3: write the periodic function in form of Fourier series

∞
4 1 2𝑛𝜋
𝑓(𝑡) = ∑ 𝑠𝑖𝑛 𝑡
𝜋 𝑛 𝑇
𝑛=1

3
 ME 4440-5540 Lecture 27

If we plot this function, for 𝑛 = 1, 𝑛 = 5, 𝑛 = 12 terms of this Fourier series and compare it with
the original function, you will see, with increasing the number of terms in Fourier series, the
results will be closer to the original function. Usually for 𝑛 > 100 this function become very close
(almost same) to the original function with the advantage of working with series of harmonic terms.

𝑛=1
𝑛=5

𝑛 = 12

Application for using Fourier series in vibration:

If we having forcing function in the form of periodic function, we can write it in form of Fourier
series, find the solution for each term in series and add them with each other (use superposition).

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑎0 + ∑∞
𝑛=1[𝑎𝑛 𝑐𝑜𝑠 𝑛𝜔𝑡 + 𝑏𝑛 𝑠𝑖𝑛 𝑛𝜔𝑡 ]

𝑎0 → 𝑢𝑠𝑡𝑎𝑡𝑖𝑐 Find the solution for the static term

∑∞
𝑛=1 𝑎𝑛 𝑐𝑜𝑠 𝑛𝜔𝑡 → 𝑢𝐶𝑛 (cos 𝑡𝑒𝑟𝑚𝑠) Find the solution for all of the cos terms

∑∞
𝑛=1 𝑏𝑛 𝑠𝑖𝑛 𝑛𝜔𝑡 → 𝑢𝑠𝑛 (sin 𝑡𝑒𝑟𝑚𝑠) Find the solution for all of the sin terms

Add all solutions to get final solution. It would be hard with hand calculation but you can use some
computational software (e.g. Matlab) to do this.

4
 ME 4440-5540 Lecture 27

In general form we have from before:

𝐹0 1
𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝐹0 𝑐𝑜𝑠𝜔𝑡 = ( )
𝑘 √[(1 − 𝑟 2 )2 + (2𝜁𝑟)2 ]
∞

𝑢(𝑡) = 𝑢𝑠𝑡𝑎𝑡𝑖𝑐 + ∑(𝑢𝐶𝑛 + 𝑢𝑠𝑛 )

𝑛=1

𝐹0 𝑎0 1
𝑢𝑠𝑡𝑎𝑡𝑖𝑐 = = ( = 0)
𝑘 𝑘 √[(1−𝑟 2 )2 +(2𝜁𝑟)2 ]

∞
𝑎0 𝑎𝑛 1
𝑢(𝑡) = +∑ cos(𝑛𝜔𝑡 − 𝜃𝑛 )
𝑘 𝑘 √[(1 − 𝑛2 𝑟 2 )2 + (2𝜁𝑛𝑟)2 ] General
𝑛=1
∞ Response
𝑏𝑛 1
+∑ s 𝑖𝑛(𝑛𝜔𝑡 − 𝜃𝑛 )
𝑘 √[(1 − 𝑛2 𝑟 2 )2 + (2𝜁𝑛𝑟)2 ]
𝑛=1

When for harmonically excited system 𝜃𝑛 would be:

2𝜁𝑛𝑟
𝜃𝑛 = 𝑡𝑎𝑛−1
1 − 𝑛2 𝑟 2

5
 ME 4440-5540 Lecture 28

B) Periodic Force

Example 34: A SODF system is shown in following figure. This system includes two springs with
same stiffness of 𝑘 = 3.5 𝑘𝑁/𝑚 and a damper with damping coefficient of 𝑐 = 0.2 𝑘𝑁. 𝑠𝑒𝑐/𝑚
and a mass 𝑚 = 20 𝑘𝑔. A cam drive mechanism is attached to the spring and move the system up.
The cam operating at a frequency of 𝜔 = 60 𝑟𝑝𝑚. The input displacement function from the cam
to the system is showing on the following graph. Find the response of the system u(t) based on this
input displacement function.

The input displacement is just related to spring (1), so the equation of motion would be:

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑘1 𝑦(𝑡)

2𝜋 𝑟𝑎𝑑
𝑘1 = 3500 𝑁⁄𝑚 , 𝜔 = 60 × 60 = 2𝜋 , 𝑦𝑚𝑎𝑥 = 25 𝑚𝑚 = 0.025𝑚 , 𝑇 = 1 𝑠𝑒𝑐
𝑠

𝑓(𝑡) = 𝑘1 𝑦(𝑡) = (3500)(0.025𝑡) = 87.5𝑡 (0 < 𝑡 < 1)

1 𝑇 1
87.5 2 1 87.5
𝑎0 = ∫ 𝑓(𝑡) 𝑑𝑡 = ∫ 87.5𝑡 𝑑𝑡 = 𝑡 | =
𝑇 0 0 2 0 2

2 𝑇 1
𝑎𝑛 = ∫ 𝑓(𝑡)𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡 = 2 ∫ 87.5𝑡 𝑐𝑜𝑠 𝑛𝜔𝑡 𝑑𝑡
𝑇 0 0
1 1 1
𝑎𝑛 = 2 × 87.5 [ 2
𝑐𝑜𝑠 𝑛𝜔𝑡 + 𝑡 𝑠𝑖𝑛 𝑛𝜔𝑡 ] |
(𝑛𝜔) 𝑛𝜔 0
1 1
𝑎𝑛 = 2 × 87.5 [ 2
+0− −0]=0
(2𝜋𝑛) (2𝜋𝑛)2

2 𝑇 1
𝑏𝑛 = ∫ 𝑓(𝑡)𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡 = 2 ∫ 87.5𝑡 𝑠𝑖𝑛 𝑛𝜔𝑡 𝑑𝑡
𝑇 0 0

1
 ME 4440-5540 Lecture 28

1 1 1
𝑏𝑛 = 2 × 87.5 [ 2
𝑠𝑖𝑛 𝑛𝜔𝑡 − 𝑡 𝑐𝑜𝑠 𝑛𝜔𝑡 ] |
(𝑛𝜔) 𝑛𝜔 0
1 87.5
𝑏𝑛 = 2 × 87.5 [0 − −0+0] =−
2𝜋𝑛 𝜋𝑛

Note: ∫ 𝑢𝑑𝑣 = 𝑢𝑣 − ∫ 𝑣𝑑𝑢 , For example:

∫ 𝑡 𝑐𝑜𝑠 (𝑡)𝑑𝑡 , 𝑢 = 𝑡, 𝑑𝑣 = 𝑐𝑜𝑠 (𝑡) 𝑣 = 𝑠𝑖𝑛 (𝑡) , 𝑑𝑢 = 1

∫ 𝑢𝑑𝑣 = 𝑢𝑣 − ∫ 𝑣𝑑𝑢 = 𝑡 𝑠𝑖𝑛 (𝑡) − ∫ 𝑠𝑖𝑛 (𝑡)𝑑𝑡 = 𝑡 𝑠𝑖𝑛 (𝑡) + 𝑐𝑜𝑠 (𝑡)

The springs in this system are going through same deformation, so they are parallel with each
other. So total stiffness of the system would be equal to:

𝑘𝑒𝑞 = 𝑘1 + 𝑘2 = 3500 + 3500 = 7000 𝑁/𝑚

𝑐 𝑐 0.2×103
𝜁 = 𝑐 = 2√𝑘.𝑚 = 2√7000×20 = 0.267
𝑐

𝑘 7000 𝑟𝑎𝑑
𝜔𝑛 = √ =√ = 18.71
𝑚 20 𝑠

𝜔 2𝜋
𝑟= = = 0.11𝜋
𝜔𝑛 18.71

∞
𝑎0 𝑎𝑛 1
𝑢(𝑡) = +∑ cos(𝑛𝜔𝑡 − 𝜃𝑛 )
𝑘 𝑘 √[(1 − (𝑟𝑛)2 )2 + (2𝜁𝑛𝑟)2 ]
𝑛=1
∞
𝑏𝑛 1
+∑ s 𝑖𝑛(𝑛𝜔𝑡 − 𝜃𝑛 )
𝑘 √[(1 − (𝑟𝑛)2 )2 + (2𝜁𝑛𝑟)2 ]
𝑛=1

2𝜁𝑛𝑟
𝜃𝑛 = 𝑡𝑎𝑛−1
1 − 𝑛2 𝑟 2

∞
87.5 87.5 1 s 𝑖𝑛(2𝜋𝑛𝑡 − 𝜃𝑛 )
𝑢(𝑡) = − ∑
2(7000) 𝜋(7000) 𝑛 √[(1 − (0.11𝑛𝜋)2 )2 + (2 × 0.267 × 𝑛 × 0.11𝜋)2 ]
𝑛=1

0.0587𝑛𝜋
𝜃𝑛 = 𝑡𝑎𝑛−1
1 − (0.11𝑛𝜋)2

2
 ME 4440-5540 Lecture 29

C) General Forcing Function

As the last type of forcing function, we study the general forcing function without any limitation.
The force function can be impulse force, impact load, blast load, earthquake load, etc.

Look at the response of the mass under the action of a general function 𝑓(𝑡). In this case the
approach to drive the response is done in two different ways: 1) Convolution Theorem 2)
Duhamel's Integral. With using both approaches, we reach the same results. However, in this class,
we are only working on Duhamel's Integral method.

First of all, we assume system is linear and that means if you have several loads/forces acting on
the system, if you find the response to the system under the action of each individual load and
add them up, that will give you total response.

For the previous function, we can break down that function to infinite number of loads.

If we can find the response of the system under the load of 𝑓(𝜏) for any arbitrary point at the
time of (𝜏) with duration of (𝑑𝜏), then we can apply it throughout the entire period and find total
solution.

1
 ME 4440-5540 Lecture 29

Let’s introduce some terminologies:

1) Impulse loading: Load with a short duration.

𝐼𝑚𝑝𝑢𝑙𝑠𝑒 = 𝑓(𝜏). 𝑑𝜏
2) Impulse apply to a system causes a change in the momentum

For
simplicity 𝑑
𝑓𝑖 (𝜏). 𝑑𝜏 𝑓. 𝑑𝜏 = 𝑑(𝑚𝑢̇ ) 𝑓 = 𝑑𝜏 (𝑚𝑢̇ )

If the mass is constant, we can say the impulse causes a change in velocity.

𝑑𝑢 𝑓.𝑑𝜏 𝑑
𝑓. 𝑑𝜏 = 𝑑(𝑚𝑢̇ ) = 𝑚𝑑( 𝑑𝑡 ) = 𝑑𝑡 (𝑑𝑢)
𝑚

In the other words, from initial condition for system would be:
𝑢0 = 0
{ 𝑓. 𝑑𝜏
𝑢̇ 0 =
𝑚
This concept is the general form of the Newton's second law!
Newton's second law: Rate of change of momentum of the mass is equal to the force acting on it.

2
 ME 4440-5540 Lecture 29

For finding the response for general force, we assume we don’t have a force vibration but we have
a system which due to an impulse, it is subjected to an initial velocity. So, that means we have a
free vibration, due to initial conditions (just initial velocity).
From before we have the steady state response for free vibration of a damped system would be:

𝜁𝜔𝑛 𝑠𝑖𝑛 𝜔𝑑 𝑡
𝑢(𝑡) = 𝑒 −𝜁𝜔𝑛𝑡 [( 𝑠𝑖𝑛 𝜔𝑑 𝑡 + 𝑐𝑜𝑠𝜔𝑑 𝑡) 𝑢0 + 𝑢̇ 0 ]
𝜔𝑑 𝜔𝑑
𝑓.𝑑𝜏
But in this case and from initial conditions 𝑢0 = 0 & 𝑢̇ 0 = , so we have:
𝑚
𝑓. 𝑑𝜏
0 𝑚
𝜁𝜔𝑛 𝑠𝑖𝑛 𝜔𝑑 𝑡
𝑢(𝑡) = 𝑒 −𝜁𝜔𝑛𝑡 [( 𝑠𝑖𝑛 𝜔𝑑 𝑡 + 𝑐𝑜𝑠𝜔𝑑 𝑡) 𝑢0 + 𝑢̇ 0 ]
𝜔𝑑 𝜔𝑑

𝑓. 𝑑𝜏 1 −𝜁𝜔 𝑡
𝑢(𝑡) = × 𝑒 𝑛 𝑠𝑖𝑛 𝜔 𝑡
𝑑
𝑚 𝜔𝑑
However, we are looking to find the response of system at any point (moment) after this impulse
was applied (a general point).

The response of the system on that point would be:

𝑓. 𝑑𝜏 1 −𝜁𝜔 (𝑡−𝜏)
𝑢(𝑡 − 𝜏) = × 𝑒 𝑛 𝑠𝑖𝑛 𝜔𝑑 (𝑡 − 𝜏)
𝑚 𝜔𝑑

3
 ME 4440-5540 Lecture 29

However, this is the response of the system to one of these impulses! Now, we need to find the
total response of system by adding the responses for all small impulses. The total response would
be:

𝑡
1 Duhamel's
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑒 −𝜁𝜔𝑛(𝑡−𝜏) 𝑠𝑖𝑛 𝜔𝑑 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑑 0 Integral

𝜏: A dummy variable.
For an Undamped system(𝜁 = 0), we will have:

𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0

4
 ME 4440-5540 Lecture 30

C) General Forcing Function

Let’s see some application of this concept.

Example 35: We have an undamped system which is subjected to an impact load (𝐹0 ). Find the
maximum displacement due to this load.

Impact load: Whenever a machine members are subjected to load with a sudden impact due to
falling or hitting one object on another (zero force at time zero but suddenly change to some
constant force).
0 𝑡≤0
𝑓(𝑡) = {
𝐹0 𝑡>0

Note: There is big difference between the impact load and static load. In static load, the load
adding to the system gradually during process but for impact load, we dropping an object on system
and from zero load suddenly adding a constant load to system (that causes initial conditions on
system).

𝑡 𝑡
1 𝐹0
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏 = ∫ 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑚𝜔𝑛 0

𝐹0 1 𝑡 𝐹0 𝐹0
𝑢(𝑡) = [ 𝑐𝑜𝑠 𝜔𝑛 (𝑡 − 𝜏)] | = (1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡) = (1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡)
𝑚𝜔𝑛 𝜔𝑛 0 𝑚𝜔𝑛 2 𝑘
𝑚𝑚

𝐹0
𝑢(𝑡) = (1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡)
𝑘
𝐹0
𝛿𝑠𝑡𝑎𝑡𝑖𝑐 = (Static displacement)
𝑘

1
 ME 4440-5540 Lecture 30

The maximum of 𝑢(𝑡) would be happened on 𝑐𝑜𝑠 𝜔𝑛 𝑡 = −1

𝐹0
𝑢𝑚𝑎𝑥 = 2 = 2 𝛿𝑠𝑡𝑎𝑡𝑖𝑐
𝑘

Dynamic Factor (DF)

Dynamic Load Factor (DLF)

That is why for an impact load in machine design we always consider factor of safety of 2 for a
dynamic load!
Now let’s look at the damped case but with very small damping ratio (𝜁 ≠ 0 & 𝜁 ≪ 1)
So, we have to assume τ as a dummy variable and solve the Duhamel's integral.
𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑒 −𝜁𝜔𝑛(𝑡−𝜏) 𝑠𝑖𝑛 𝜔𝑑 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑑 0
𝐹0 𝜁
𝑢(𝑡) = [1 − 𝑒 −𝜁𝜔𝑛𝑡 (𝑐𝑜𝑠 𝜔𝑑 𝑡 + 𝑠𝑖𝑛 𝜔𝑑 𝑡)]
𝑘 √1 − 𝜁 2

Static DLF (We have to find the

Displacement 𝑑𝑢 maximum of this part)
To find the 𝑢𝑚𝑎𝑥 , we have to solve ( 𝑑𝑡 = 0) and find 𝑡, and substitute it in above equation.
(constant)

≠0 ≠0
𝑑𝑢 𝜁 𝜁𝜔𝑑
= 𝜁𝜔𝑛 𝑒 −𝜁𝜔𝑛 𝑡 (𝑐𝑜𝑠 𝜔𝑑 𝑡 + 𝑠𝑖𝑛 𝜔𝑑 𝑡) − 𝑒 −𝜁𝜔𝑛 𝑡 (−𝜔𝑑 𝑠𝑖𝑛 𝜔𝑑 𝑡 + 𝑐𝑜𝑠 𝜔𝑑 𝑡) = 0
𝑑𝑡 √1 − 𝜁 2 √1 − 𝜁 2
For 𝜁 ≪ 1, ≈ 1 For 𝜁 ≪ 1, ≈ 1

𝑑𝑢
= 𝜁𝜔𝑛 𝑐𝑜𝑠 𝜔𝑑 𝑡 + 𝜁 2 𝜔𝑛 𝑠𝑖𝑛 𝜔𝑑 𝑡 + 𝜔𝑑 𝑠𝑖𝑛 𝜔𝑑 𝑡 − 𝜁𝜔𝑑 𝑐𝑜𝑠 𝜔𝑑 𝑡 = 0
𝑑𝑡

𝑑𝑢
= (𝜁 2 𝜔𝑛 + 𝜔𝑑 ) 𝑠𝑖𝑛 𝜔𝑑 𝑡 = 0
𝑑𝑡

Constant and it
can’t be zero

2
 ME 4440-5540 Lecture 30

𝜋
𝑠𝑖𝑛 𝜔𝑑 𝑡 = 0 𝜔𝑑 𝑡 = 𝜋 𝑡=𝜔
𝑑

For 𝜁 ≪ 1, ≈ 1
−𝜁𝜋
𝜋
−𝜁𝜔𝑛 𝜔 √1−𝜁2
𝐷𝐿𝐹 = 1 + 𝑒 𝑑 =1+𝑒 = 1 + 𝑒−𝜁𝜋

Note: (𝜔𝑑 = √1 − 𝜁 2 𝜔𝑛 )

𝐷𝐿𝐹 = 1 + 𝑒−𝜁𝜋

Even in case of damped system, as long as we have a small damping ratio (𝜁 less than 10%), 𝐷𝐿𝐹
would be very close to 𝐷𝐿𝐹 for undamped system (𝐷𝐿𝐹 ≅ 2). So, for the case of impact loading, it
would be a fair assumption to consider 𝐷𝐿𝐹 = 2 in the design.

𝜁 = 0.01 𝐷𝐿𝐹 = 1 + 𝑒−𝜁𝜋 = 1.97 ≅ 2

𝜁 = 0.02 𝐷𝐿𝐹 = 1 + 𝑒−𝜁𝜋 = 1.94 ≅ 2

Here is comparison between displacement caused by a static load and an impact load:

Static Load Impact Load

3
 ME 4440-5540 Lecture 30

Applications for Duhamel's Integral:

𝐹
First Application: How to evaluate the maximum displacement 𝑢𝑚𝑎𝑥 vs static displacement ( 0 ), if we
𝑘
have general forcing function.
Second Application: Shock Spectrum

First Application:
We already study the simple differential equations and saw how to solve, but in the case of Duhamel's
Integral, the integration is the most important part. In the forced vibration, we studied the harmonic forces
(and periodic forces) to drive displacement function 𝑢 (𝑡). But the ultimate goal is finding the maximum
displacement 𝑢𝑚𝑎𝑥 to find the maximum stresses 𝜎𝑚𝑎𝑥 for the design purposes. In the case of harmonic
forces, independent to details of force function (amplitude, frequency, etc.), the maximum displacement
𝐹
was equal to multiplication of static displacement and frequency response function 𝑢𝑚𝑎𝑥 = 0 × |𝐻|. 𝑘
However, in general force function, we have to divide the function to different regions and check each of
them for maximum displacement and compare with each other to find where the maximum displacement
will be happened.

4
 ME 4440-5540 Lecture 31

C) General Forcing Function

Example 36: There is a ramp load apply to an undamped system as it is shown in following graph. Time
𝜏 represent any arbitrary time from zero to time 𝑡𝑑 . Find the maximum displacement 𝑢𝑚𝑎𝑥 vs static
𝐹
displacement ( 𝑘0 ) of this system.

Wrong assumption: There is no load after time 𝑡𝑑 on this system so we can assume there is no load
applied on this system at all and system has not any vibration!
This is wrong assumption because before time 𝑡𝑑 , the force causes some initial condition (initial velocity)
on system and after time 𝑡𝑑 , we can assume system as a free vibration case!
According to the graph, forcing function would be:

𝜏
(1 − )𝐹 0 ≤ 𝑡 < 𝑡𝑑
𝑓(𝑡) = { 𝑡𝑑 0
0 𝑡 > 𝑡𝑑
In this case, we can divide the function to two regions: 1) 0 ≤ 𝑡 < 𝑡𝑑 2) 𝑡 > 𝑡𝑑
1) 0 ≤ 𝑡 < 𝑡𝑑

Displacement function for an Undamped system(𝜁 = 0), would be:

Note: we are taking integral from 0 to 𝑡!

𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0
𝑡
𝐹0 𝜏
𝑢(𝑡) = ∫ (1 − ) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡𝑑

𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) = 𝑠𝑖𝑛 𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 𝜏 − 𝑐𝑜𝑠 𝜔𝑛 𝑡 𝑠𝑖𝑛 𝜔𝑛 𝜏

Constant

1
 ME 4440-5540 Lecture 31

𝑡 𝑡
𝐹0 𝜏 𝐹0 𝜏
𝑢(𝑡) = ∫ (1 − ) 𝑠𝑖𝑛 𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 − ∫ (1 − ) 𝑐𝑜𝑠 𝜔𝑛 𝑡 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏
𝑚𝜔𝑛 0 𝑡𝑑 𝑚𝜔𝑛 0 𝑡𝑑

1 2 3 4
𝑡
𝐹0 1 𝑡 𝑡
1 𝑡
𝑢(𝑡) = {𝑠𝑖𝑛 𝜔𝑛 𝑡 [∫ 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 − ∫ 𝜏 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 ] − 𝑐𝑜𝑠 𝜔𝑛 𝑡 [∫ 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 − ∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 ]}
𝑚𝜔𝑛 0 𝑡𝑑 0 0 𝑡𝑑 0

We have 4 integrals that need to be solved:

𝑡 1 𝑡 1
1) ∫0 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 = 𝜔 𝑠𝑖𝑛 𝜔𝑛 𝜏| = 𝜔 𝑠𝑖𝑛 𝜔𝑛 𝑡
0
𝑛 𝑛

𝑡 1 𝑡 1 1
3) ∫0 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 = − 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝜏| = 𝜔 − 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝑡
0 𝑛 𝑛 𝑛

For integration number 2 & 4, we need to use the integration by parts:

𝑏 𝑏
𝑏
∫ 𝑢 𝑑𝑣 = 𝑢𝑣| − ∫ 𝑣 𝑑𝑢
𝑎 𝑎 𝑎
𝑡
2) ∫0 𝜏 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏
1
𝑢=𝜏 , 𝑑𝑢 = 𝑑𝜏 , 𝑑𝑣 = 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 , 𝑣=
𝜔𝑛
𝑠𝑖𝑛 𝜔𝑛 𝜏

𝑡 𝜏 𝑡 𝑡 1 𝑡 1 1 𝑡
∫0 𝜏 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 = 𝜔 𝑠𝑖𝑛 𝜔𝑛 𝜏| − ∫0 𝜔 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 = 𝜔 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝜔 [− 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝜏] |
𝑛 0 𝑛 𝑛 𝑛 𝑛 0

𝑡
𝑡 1
∫ 𝜏 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 = 𝑠𝑖𝑛 𝜔𝑛 𝑡 + 2 (𝑐𝑜𝑠 𝜔𝑛 𝑡 − 1)
0 𝜔𝑛 𝜔𝑛
𝑡
4) ∫0 𝜏 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏
1
𝑢=𝜏 , 𝑑𝑢 = 𝑑𝜏 , 𝑑𝑣 = 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 , 𝑣=−
𝜔𝑛
𝑐𝑜𝑠 𝜔𝑛 𝜏

𝑡 𝜏 𝑡 𝑡 1 𝑡 1 1 𝑡
∫0 𝜏 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 = − 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝜏| + ∫0 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 = − 𝜔 𝑐𝑜𝑠 𝜔𝑛 𝑡 + 𝜔 [𝜔 𝑠𝑖𝑛 𝜔𝑛 𝜏] |
𝑛 0 𝑛 𝑛 𝑛 𝑛 0
𝑡
𝑡 1
∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏 = − 𝑐𝑜𝑠 𝜔𝑛 𝑡 + 2 𝑠𝑖𝑛 𝜔𝑛 𝑡
0 𝜔𝑛 𝜔𝑛

Total response of system, 𝑢(𝑡) for (𝟎 < 𝒕 ≤ 𝒕𝒅 ) :

𝐹0 1 𝑡 1 1
𝑢(𝑡) = {𝑠𝑖𝑛 𝜔𝑛 𝑡 [ 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 2 𝑐𝑜𝑠 𝜔𝑛 𝑡 + 2 ]
𝑚𝜔𝑛 𝜔𝑛 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑
1 1 𝑡 1
− 𝑐𝑜𝑠 𝜔𝑛 𝑡 [ − 𝑐𝑜𝑠 𝜔𝑛 𝑡 + 𝑐𝑜𝑠 𝜔𝑛 𝑡 − 2 𝑠𝑖𝑛 𝜔𝑛 𝑡]}
𝜔𝑛 𝜔𝑛 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑

2
 ME 4440-5540 Lecture 31

Simplify version of 𝑢(𝑡):

𝐹0 𝑡 1
𝑢(𝑡) = {1 − + 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝑐𝑜𝑠 𝜔𝑛 𝑡 }
𝑘 𝑡𝑑 𝜔𝑛 𝑡𝑑
Now for finding the maximum displacement (𝑢𝑚𝑎𝑥 ), we have to take derivative of 𝑢(𝑡) and set it
equal zero and solve it for 𝑡, put this 𝑡 value back to the 𝑢(𝑡) to find the 𝑢𝑚𝑎𝑥 .
2) 𝑡 > 𝑡𝑑
For this part, we can use two different methods:
A) As it can be seen after time 𝑡𝑑 , there is no force acting on the system and the system is undamped.
So, we have a free vibration of an undamped system that having some initial conditions which can
be find from previous part (𝑢(𝑡𝑑 ) & 𝑢̇ (𝑡𝑑 ))

B) We can write the Duhamel's integral for part 1 (integral from 0 𝑡𝑜 𝑡𝑑 ) plus part 2 (integral
from 𝑡𝑑 𝑡𝑜 𝑡) for any 𝑡 > 𝑡𝑑

𝒕𝒅 𝑡 0
𝐹0 𝜏 1
𝑢(𝑡) = ∫ (1 − ) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏 + ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡𝑑 𝑚𝜔𝑛 𝒕𝒅

𝑓(𝜏) = 0 for 𝑡 > 𝑡𝑑

If we solve it, we will get the same result as previous section, but this time, integral from 0 to 𝑡𝑑 . So, the
result for second part of the graph would be:
𝐹0 1 1 1
𝑢(𝑡) = {𝑠𝑖𝑛 𝜔𝑛 𝑡 [− 𝑐𝑜𝑠 𝜔𝑛 𝑡𝑑 + ] − 𝑐𝑜𝑠 𝜔𝑛 𝑡 [1 − 𝑠𝑖𝑛 𝜔𝑛 𝑡𝑑 ]}
𝑚𝜔𝑛 2 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑
𝐹0 1 1 1
𝑢(𝑡) = {𝑠𝑖𝑛 𝜔𝑛 𝑡 [− 𝑐𝑜𝑠 𝜔𝑛 𝑡𝑑 + ] − 𝑐𝑜𝑠 𝜔𝑛 𝑡 [1 − 𝑠𝑖𝑛 𝜔𝑛 𝑡𝑑 ]}
𝑘 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑 𝜔𝑛 𝑡𝑑
Note: The results from both methods would be same!
Then, we can find the maximum displacement (𝑢𝑚𝑎𝑥 ), for this part same as the first part of the
graph and compare them to find which of them is the true 𝑢𝑚𝑎𝑥 for the system.

In general, a “General Force Function”, based on the shape of the function, can be divide to several
sections/parts (any changes in the forcing function will be consider as a new section). So, we have
to evaluate a response function section by section, find the 𝑢𝑚𝑎𝑥 for each of them, and compare to
find the true 𝑢𝑚𝑎𝑥 for the system.

 Do you think a blast loading or a static load has more severe effect on a structure? For
answering to this question, look at the following problem.

3
 ME 4440-5540 Lecture 32

C) General Forcing Function

Example 37: The following figure showing a water tower carries 5,000 kg of water. The column has a
stiffness of 𝑘 = 5.4 × 106 𝑁/𝑚. This water tower is subjected to a blast loading (very short duration of
loading) as shown in the following graph. Find the maximum displacement 𝑢𝑚𝑎𝑥 of this water tower.
Assume this structure as an undamped system.

In this case, we have 3 regions:

1 2 3

𝑘 5.4×106 𝑟𝑎𝑑
𝜔𝑛 = √𝑚 = √ = 32. 86
5000 𝑠

2𝜋 2𝜋
𝑇 = 𝜔 = 32.86 = 0.191 𝑠𝑒𝑐
𝑛

1
 ME 4440-5540 Lecture 32

Undamped 𝜁=0
Now, to evaluate the response of system, we can use the Duhamel's integral for an undamped
case..

𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0

In this case, we have three regions for 𝑓(𝜏):

2𝐹0 𝑡0
𝜏 0<𝑡<
𝑡0 2
𝑓(𝜏) = 𝜏 𝑡0
2𝐹0 (1 − ) < 𝑡 < 𝑡0
𝑡0 2
{ 0 𝑡0 < 𝑡
𝑡0
Region 1: 0 < 𝑡 < 2

𝑡
1 2𝐹0
𝑢(𝑡) = ∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡0

𝑡0
Note: We are taking integral from 0 𝑡𝑜 𝑡 Not from 0 𝑡𝑜 ! Actually, we are interested to drive an
2
expression for the response of system caused by the force in region 1, for any moment (not only
in region 1).

If we solve the integral for region 1, we will have:

Static response

2𝐹0 1 𝑡 2𝐹0 1 𝐹0 2 1
𝑢(𝑡) = {𝑡 + [𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏)]| } = (𝑡 − 𝑠𝑖𝑛 𝜔𝑛 𝑡) = ( ) (𝑡 − 𝑠𝑖𝑛 𝜔𝑛 𝑡)
2
𝑚𝜔𝑛 𝑡0 𝜔𝑛 0 𝑘𝑡0 𝜔𝑛 𝑘 𝑡0 𝜔𝑛

Note: Always the response to general loading depends on two important quantities: 1) natural
frequency of the system (𝜔𝑛 ) 2) Duration of the forcing function (𝑡0 ). We will study later how can
reach to maximum response as a function of these two quantities (Shock Spectrum)

𝑡0
Region 2: < 𝑡 < 𝑡0
2
𝑡0
For this case, we have to add the response of system from 0 𝑡𝑜 2
to the response of system after
𝑡
time 20 .

2
 ME 4440-5540 Lecture 32

𝑡0
𝑡
1 2 2𝐹0 1 𝜏
𝑢(𝑡) = ∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏 + ∫ 2𝐹0 (1 − ) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡0 𝑚𝜔𝑛 𝑡0 𝑡0
2

2𝐹0 1 𝑡0
𝑢(𝑡) = (𝑡0 − 𝑡 + [2 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − ) − 𝑠𝑖𝑛 𝜔𝑛 𝑡])
𝑘𝑡0 𝜔𝑛 2

Region 3: 𝑡0 < 𝑡

In this case, we have two options: After time 𝑡0 , we can look at the system as an undamped free
vibration with some initial conditions (𝑢(𝑡0 ) & 𝑢̇ (𝑡0 )) or we can rewrite the Duhamel's integral
for all three regions.

𝑡0
𝑡0
1 2 2𝐹0 1 𝜏
𝑢(𝑡) = ∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏 + ∫ 2𝐹0 (1 − ) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡0 𝑚𝜔𝑛 𝑡0 𝑡0
2
1 𝑡 0
+ ∫ 𝑓(𝑡) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 𝑡0
𝑡0
𝑡0
1 2 2𝐹0 1 𝜏
𝑢(𝑡) = ∫ 𝜏 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏 + ∫ 2𝐹0 (1 − ) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0 𝑡0 𝑚𝜔𝑛 𝑡0 𝑡0
2

2𝐹0 𝑡0
𝑢(𝑡) = [2 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − ) − 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝑡0 ) ]
𝑘𝑡0 𝜔𝑛 2

For finding the maximum displacement of the system, we have to take derivative of all three
region, set them equal zero, find the value for 𝑡, put it back to equations and find the maximum
displacement for each region (if there is any) and compare them to find the maximum displacement
for the system.

1) Finding the maximum displacement for the first region

2𝐹0 1
𝑢(𝑡) = (𝑡 − 𝑠𝑖𝑛 𝜔𝑛 𝑡)
𝑘𝑡0 𝜔𝑛

𝑑𝑢 2𝐹0
= (1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡)
𝑑𝑡 𝑘𝑡0

3
 ME 4440-5540 Lecture 32

2𝐹0
For region (1), >0 and maximum value for 𝑡 is 0.015 𝑠𝑒𝑐, so
𝑘𝑡0
𝑑𝑢 𝑡 𝑑𝑢
1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡 > 0. Therefore, always > 0 for whole period of 0 < 𝑡 < 20 . That means never
𝑑𝑡 𝑑𝑡
become zero for any value of 𝑡 in the region (1) and there is no maximum displacement for the
region (1).

2) Finding the maximum displacement for the second region

2𝐹0 1 𝑡0
𝑢(𝑡) = (𝑡0 − 𝑡 + [2 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − ) − 𝑠𝑖𝑛 𝜔𝑛 𝑡])
𝑘𝑡0 𝜔𝑛 2

𝑑𝑢 2𝐹0 𝑡0
= (−1 + 2 𝑐𝑜𝑠 𝜔𝑛 (𝑡 − ) − 𝑐𝑜𝑠 𝜔𝑛 𝑡)
𝑑𝑡 𝑘𝑡0 2

Let’s first check for start and end point of the region (2).
𝑡
For 𝑡 = 20 2
𝑑𝑢 2𝐹0 𝑡0 𝑡0 𝑡0
= (−1 + 2 𝑐𝑜𝑠 𝜔𝑛 ( − ) − 𝑐𝑜𝑠 𝜔𝑛 ) > 0
𝑑𝑡 𝑘𝑡0 2 2 2
For 𝑡 = 𝑡0

𝑑𝑢 2𝐹0 𝑡0 2𝐹0
= (−1 + 2 𝑐𝑜𝑠 𝜔𝑛 (𝑡0 − ) − 𝑐𝑜𝑠 𝜔𝑛 𝑡0 ) = (−1 + 2 𝑐𝑜𝑠 (0.493) − cos(0.986)) > 0
𝑑𝑡 𝑘𝑡0 2 𝑘𝑡0

2𝐹 𝑡 𝑑𝑢 𝑑𝑢
For region (2), 𝑘𝑡0 > 0 and for any value of 𝑡 ( 20 < 𝑡 < 𝑡0 ), > 0. That means never become
0 𝑑𝑡 𝑑𝑡
zero for any value of 𝑡 in the region (2) and there is no maximum displacement for the region (2).
So, if there is any maximum displacement for the system, it has to be happen in region (3).

3) Finding the maximum displacement for the third region

First we can simplify the 𝑢(𝑡) for this region.

2𝐹0 𝑡0
𝑢(𝑡) = [2 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − ) − 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝑡0 ) ]
𝑘𝑡0 𝜔𝑛 2

We can expand these parts

4
 ME 4440-5540 Lecture 32

2𝐹0 𝑡0 𝑡0
𝑢(𝑡) = [2 𝑠𝑖𝑛 𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 − 2 𝑐𝑜𝑠 𝜔𝑛 𝑡 𝑠𝑖𝑛 𝜔𝑛 − 𝑠𝑖𝑛 𝜔𝑛 𝑡
𝑘𝑡0 𝜔𝑛 2 2
− 𝑠𝑖𝑛 𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 𝑡0 + 𝑐𝑜𝑠 𝜔𝑛 𝑡 𝑠𝑖𝑛 𝜔𝑛 𝑡0 ]

𝑡0
𝜔𝑛 (𝑟𝑎𝑑) 𝜔𝑛 𝑡0 (𝑟𝑎𝑑)
2
2𝐹0
𝑢(𝑡) = [𝑠𝑖𝑛 𝜔𝑛 𝑡[2 𝑐𝑜𝑠 (0.493) − 1 − 𝑐𝑜𝑠 (0.986)] + 𝑐𝑜𝑠 𝜔𝑛 𝑡[−2 𝑠𝑖𝑛 (0.493)
𝑘𝑡0 𝜔𝑛
+ 𝑠𝑖𝑛 (0.986)]]

A B
2𝐹0
𝑢(𝑡) = [0.208 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 0.113 𝑐𝑜𝑠 𝜔𝑛 𝑡]
𝑘𝑡0 𝜔𝑛
Harmonic
2𝐹0 Function
𝑢(𝑡) = √𝐴2 + 𝐵 2 𝑠𝑖𝑛 (𝜔𝑛 𝑡 + ∅)
𝑘𝑡0 𝜔𝑛

Instead of take the derivative of this equation and equate to zero to find the 𝑢𝑚𝑎𝑥 , we can see the
response is a harmonic function, so the maximum value would be:

2𝐹0 2 × 5000 × 9.81

𝑢𝑚𝑎𝑥 = √𝐴2 + 𝐵 2 = √0.2082 + 0.1132 = 0.004 𝑚
𝑘𝑡0 𝜔𝑛 5.4 × 106 × 0.03 × 32. 86

𝐹0 5000×9.81
𝛿𝑠𝑡𝑎𝑡𝑖𝑐 = = = 0.009 𝑚
𝑘 5.4×106

As you can see, the maximum displacement made by a blast load is even less than a half of a static
load! Therefore, it is not necessary the maximum response of a blast load (or any other short
duration dynamic loads) be higher than corresponding static displacement of system.

5
 ME 4440-5540 Lecture 33

C) General Forcing Function

General Approach-Numerical Integration:

In order to find the response under a general loading, using the analytical integral is hard and have
to be solve for each regions. A more general approach to find the response of system under a
general loading is using numerical integration.
𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0
𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 𝑡 𝑐𝑜𝑠 𝜔𝑛 𝜏 − 𝑐𝑜𝑠 𝜔𝑛 𝑡 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏
𝑚𝜔𝑛 0

𝜔𝑛 𝑡 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡
A B
𝑡 𝑡
1 1
𝑢(𝑡) = 𝑠𝑖𝑛 𝜔𝑛 𝑡 ∫ 𝑓(𝜏) 𝑐𝑜𝑠 𝜔𝑛 𝜏 𝑑𝜏 − 𝑐𝑜𝑠 𝜔𝑛 𝑡 ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 𝜏 𝑑𝜏
𝑚𝜔𝑛 0 𝑚𝜔𝑛 0

1 1
𝑢(𝑡) = {𝐴 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 𝐵 𝑐𝑜𝑠 𝜔𝑛 𝑡} = √𝐴2 + 𝐵 2 𝑠𝑖𝑛 (𝜔𝑛 𝑡 + ∅)
𝑚𝜔𝑛 𝑚𝜔𝑛

(A & B are time dependent)

You can see, always 𝑢(𝑡) can be represent as a harmonic function! So, we only need to find the A
and B. The procedure to find A & B is called numerical integration. The most common types of
numerical integration methods: 1) Central Difference method 2) Newmark-𝛽 numerical integration
3) Wilson-𝜃 scheme numerical integration.

We are not cover any of these methods in these course and you can know about them in more
advance level courses.

Second Application: Shock Spectrum/Response Spectrum

The maximum displacement under any kind of short duration loading which we call that shock
loading is a function of two variables: time duration of the load (𝑡0 ) and natural frequency of the
system (𝜔𝑛 ). Depending on these two variables, you may have a response with maximum
displacement higher than corresponding static displacement or lower than that.

Note: We only know for impact loading the factor of safety should be equal “2”, and for other
types of loading you have to evaluate the maximum value.

1
 ME 4440-5540 Lecture 33

In harmonic loading the maximum displacement was multiplication of static displacement by

𝐹
maximum value of the frequency response function (𝑢𝑚𝑎𝑥 = 0 |𝐻|𝑚𝑎𝑥 ). Shock 𝑘
Spectrum/Response Spectrum is sort of equivalent of that concept for maximum displacement in
case of a general loading. To make it clear, let’s look at the previous example again.

We found the 𝑢𝑚𝑎𝑥 like this:

𝐹0
2
𝑢(𝑡) = 𝑘 [0.208 𝑠𝑖𝑛 𝜔𝑛 𝑡 − 0.113 𝑐𝑜𝑠 𝜔𝑛 𝑡]
𝑡0 𝜔𝑛

𝐹
2 0 𝐹0
𝑘
𝑢𝑚𝑎𝑥 = 𝑡 = 0.004𝑚 , 𝛿𝑠𝑡𝑎𝑡𝑖𝑐 = = 0.009 𝑚
0 𝜔𝑛 𝑘

𝑢𝑚𝑎𝑥 𝑡0
If we plot the ( 𝐹0 ) respect to (𝑡0 = time duration of general force, 𝑇 = Natural period of the
𝑇
𝑘
2𝜋
system, 𝑇 = 𝜔 ), we will have:
𝑛

Absolute
maximum

0.44 1.25
0.4

0.8
0.157

This plot showing the maximum response respect to normalized force time duration. In this
problem, we had

2
 ME 4440-5540 Lecture 33

𝑇 = 0.191 𝑠𝑒𝑐 𝑡0
𝑡0 = 0.03 𝑠𝑒𝑐 = 0.157
𝑇

The response spectrum is a plot that can give you quickly the maximum dynamic response for any
𝑡 𝑢𝑚𝑎𝑥
values of ( 𝑇0 ). This plot give you lots of information. From this plot, the ( 𝐹0 ) would be equal
𝑘
𝑡0 𝑡
0.44 for = 0.157. For this type of loading which was discussed in this example, for ( 𝑇0 ) = 0.4,
𝑇
the maximum response of system would be equal to static displacement. In the other words, for
𝑡
any shock loading with the shape look like this example, as long as ( 𝑇0 ) ≤ 0.4, the dynamic
response would never exceed the corresponding static displacement! Also, from this graph, we can
𝑡
see for ( 0 ) = 0.8, we will have the absolute maximum dynamic response! Form this graph, we
𝑇
𝑡
can quickly find what would be the critical time duration for force (𝑡0 ). Also, for ( 𝑇0 ) > 0.8, the
dynamic response would be oscillate between 1-1.25 of static displacement. The shock spectrum
is available for a wide range of short duration loadings and these plots can be used for all kinds of
practical research purposes.

Example 38: Set up Shock Spectrum/Response Spectrum for following forcing function.

𝑡
1
𝑢(𝑡) = ∫ 𝑓(𝜏) 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝜏) 𝑑𝜏
𝑚𝜔𝑛 0

𝐹0
𝜏 0 < 𝑡 < 𝑡0
𝑓(𝜏) = { 𝑡0
𝐹0 𝑡0 < 𝑡

3
 ME 4440-5540 Lecture 33

If we solve it for both regions, we will find the responses equal to:
Region 1:
𝐹0 𝑡 𝑠𝑖𝑛 𝜔𝑛 𝑡
𝑢(𝑡) = ( − )
𝑘 𝑡0 𝜔𝑛 𝑡0
Region 2:
𝐹0 𝑠𝑖𝑛 𝜔𝑛 (𝑡 − 𝑡0 ) 𝑠𝑖𝑛 𝜔𝑛 𝑡
𝑢(𝑡) = (1 + − )
𝑘 𝜔𝑛 𝑡0 𝜔𝑛 𝑡0
𝑡
Now, we want to plot the 𝑢𝑚𝑎𝑥 vs ( 𝑇0 ). We can find the 𝑢𝑚𝑎𝑥 by taking derivative of previous
equations, equate to zero, find 𝑡 values, substitute these values back in 𝑢(𝑡) to find the 𝑢𝑚𝑎𝑥 . After
solving for both regions, the 𝑢𝑚𝑎𝑥 would be:

𝐹0 √2(1 − 𝑐𝑜𝑠 𝜔𝑛 𝑡0 )
𝑢𝑚𝑎𝑥 = [1 + ]
𝑘 𝜔𝑛 𝑡0

Then, based on the 𝑢𝑚𝑎𝑥 , we can plot the shock spectrum for this type of force and use it for other
experiments (e.g. earthquake, blast loads, etc).

4
 ME 4440-5540 Lecture 34

Two Degree of Freedom Systems

For single degree of freedom systems we assumed that the overall deformation of structure can be
demonstrate by looking at a single location in the structure. In general this assumption is not true!
Especially if you have vibration that goes to higher frequency of the excitation (e.g. a long beam
which is moving very fast).

Single Degree of Freedom System

Multi Degree of Freedom System

In fact, any structure made of infinite number of D.O.F. To solve this problem, one way is
considering structure as continues system and study the vibration of continues systems (it is a little
difficult to study). The alternative option would be making an assumption that the structure made
of finite degrees of freedom (finite number of lumped masses) which each one move in different
directions (multi degree of freedom system). The M.D.O.F. give you more accurate and better
understanding about the way structure deform. The study of M.D.O.F. is required using matrix
analysis. To make it easier and more understandable, before talking about M.D.O.F. systems, let’s
start with two D.O.F. systems.
Study of two D.O.F. systems is easy way to introduce the fundamental concepts of M.D.O.F.
systems. For instance “Modal Coordinates”, “Orthogonality of Modes”, “Modal transformation”,
“Mode Shapes”, etc.

1
 ME 4440-5540 Lecture 34

The following picture shows the most general case for a 2 D.O.F system. This system includes two
masses (𝑚1 & 𝑚2 ) and each of these masses subjected to separate forcing function 𝑓1 (𝑡) & 𝑓2 (𝑡).
Because of these forcing functions, each mass is experiencing a displacement respect to a base (𝑢1
& 𝑢2 ). Also, this base is moving respect to a reference frame. The 𝑢1 & 𝑢2 are relative
displacement for masses (respect to the base) and we can show the absolute displacement of masses
with (𝑧1 & 𝑧2 ) (respect to the reference frame).

In order to drive the equation of motion for this 2 D.O.F. system, there are two approaches: 1)
Newton’s law 2) Lagrange’s Formula

1) Newton’s law (Dynamic equilibrium, F.B.D, and D’Alembert force, etc.)

In this case we have two free body diagrams.

1 2

For each of these free body diagrams, we can write the equation of equilibrium.

2
 ME 4440-5540 Lecture 34

1) 𝑚1 𝑧̈1 = 𝑓1 + 𝑓𝑘2 + 𝑓𝑐2 − 𝑓𝑘1 − 𝑓𝑐1

2) 𝑚2 𝑧̈2 = 𝑓2 − 𝑓𝑘2 − 𝑓𝑐2

𝑓𝑘1 = 𝑘1 𝑢1 , 𝑓𝑘2 = 𝑘2 (𝑢2 − 𝑢1 ) , 𝑓𝑐1 = 𝑐1 𝑢̇ 1 , 𝑓𝑐2 = 𝑐2 (𝑢̇ 2 − 𝑢̇ 1 )

𝑧1 = 𝑢1 + 𝑦 & 𝑧2 = 𝑢2 + 𝑦

𝑚1 𝑢̈ 1 + (𝑐1 + 𝑐2 )𝑢̇ 1 − 𝑐2 𝑢̇ 2 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 𝑓1 − 𝑚1 𝑦̈ Equations of motion in

terms of relative
𝑚2 𝑢̈ 2 − 𝑐2 𝑢̇ 1 + 𝑐2 𝑢̇ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 𝑓2 − 𝑚2 𝑦̈ displacements (𝑢1 &𝑢2 )

1. 𝑦̈ is input acceleration

Practice: Write the equation of motion in term of absolute displacement.

𝑚1 𝑧̈1 + (𝑐1 + 𝑐2 )𝑧̇1 − 𝑐2 𝑧̇2 + (𝑘1 + 𝑘2 )𝑧1 − 𝑘2 𝑧2 = 𝑓1 + 𝑘1 𝑦 + 𝑐1 𝑦̇ Equations of motion in

terms of absolute
𝑚2 𝑧̈2 − 𝑐2 𝑧̇1 + 𝑐2 𝑧̇2 − 𝑘2 𝑧1 + 𝑘2 𝑧2 = 𝑓2
displacements (𝑧1 &𝑧2 )

In terms of applied forces

For S.D.O.F system, we had one equation in terms of one variable, so we use homogenous
deferential equation of second order and solve it directly. But in this case, we have two equations
and two variables (𝑢1 & 𝑢2 ) and in each of them we have both variables (they are coupled).
Therefore, we cannot solve these equations independently/directly.
In this case, it is possible to transform these two equations to another coordinate system where
these equations become decoupled (each equation only in term of one variable) and solve the
equations there and return them back to original coordinate system.
Note: In the S.D.O.F. we only have a single natural frequency for system. In multi degree of
freedom systems, system doesn’t have just one fundamental frequency but can vibrate under the
action of load which excite the system in multi frequencies (the first frequency called fundamental
frequency, and higher frequency called second mode, third mode,…).

3
 ME 4440-5540 Lecture 34

Example 39: The following figure showing a two story building. The columns are rigidly
connected to the masses 𝑚1 & 𝑚2 . The columns for each level of this building have same stiffness.
Find the equation of motion for this building.

This is an undamped two degree of freedom system (𝑐 = 0). First of all, let’s find the equivalent
stiffness for each level of this building. Columns are parallel to each other, so:

12𝐸1 𝐼1 24𝐸1 𝐼1
For level 1: 𝑘1 = 3 𝑘𝑒𝑞1 =
𝑙1 𝑙1 3

12𝐸2 𝐼2 24𝐸2 𝐼2
For level 2: 𝑘2 = 𝑘𝑒𝑞2 =
𝑙2 3 𝑙2 3

In this case, we only look at the free vibration of the system. Also, we don’t have a base motion.
So, based on these information and from the general formulation for two degree of freedom
systems, the equations of motion for this system would be:

0 0 0 0
𝑚1 𝑢̈ 1 + (𝑐1 + 𝑐2 )𝑢̇ 1 − 𝑐2 𝑢̇ 2 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 𝑓1 − 𝑚1 𝑦̈
0 0 0 0
𝑚2 𝑢̈ 2 − 𝑐2 𝑢̇ 1 + 𝑐2 𝑢̇ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 𝑓2 − 𝑚2 𝑦̈

𝑚1 𝑢̈ 1 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 0
Equation of motion
𝑚2 𝑢̈ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 0

However, in this case, we had translational motion for both movements, what happened if we
have one translational and one rotational motion?

4
 ME 4440-5540 Lecture 35

Two Degree of Freedom Systems

Example 40: The following figure showing the model of a car. Suspension system of this car
showing with two springs with stiffness of 𝑘1 & 𝑘2 . This car experience both translational and
rotational motions. Find the equation of motion for this car.

(CG: Center of Gravity of the car)

Free Body Diagram:

1
 ME 4440-5540 Lecture 35

To set up the equation of motion, we will have:

∑ 𝐹𝑦 = 0 1

∑ 𝑀𝐶𝐺 = 0 2

𝑓𝑘1 = 𝑘1 𝑑1 & 𝑓𝑘2 = 𝑘2 𝑑2

𝑚1 𝑢̈ 1 + 𝑓𝑘1 + 𝑓𝑘2 = 0 1

𝐼𝑢̈ 2 + 𝑓𝑘1 . 𝑙1 − 𝑓𝑘2 . 𝑙2 = 0 2

However, we want to express the above equations in terms of 𝑢1 & 𝑢2 .

𝑑1 −𝑑2 𝑢1 −𝑑2 𝑙2
𝜃 = 𝑢2 = & =
𝑙 𝑑1 −𝑑2 𝑙

We can use these two relations to write the equation of motion in terms of 𝑢1 & 𝑢2 .
𝑑1 = 𝑢1 − 𝑢2 𝑙2 + 𝑢2 𝑙 𝑑1 = 𝑢1 + 𝑢2 𝑙1
𝑑2 = 𝑢1 − 𝑢2 𝑙2
𝑓𝑘1 = 𝑘1 𝑑1 = 𝑘1 (𝑢1 + 𝑢2 𝑙1 ) & 𝑓𝑘2 = 𝑘2 (𝑢1 − 𝑢2 𝑙2 )

𝑚1 𝑢̈ 1 + (𝑘1 + 𝑘2 )𝑢1 + (𝑘1 𝑙1 − 𝑘2 𝑙2 )𝑢2 = 0

Equation of motion
𝐼𝑢̈ 2 + (𝑘1 𝑙1 − 𝑘2 𝑙2 )𝑢1 + (𝑘1 𝑙1 2 + 𝑘2 𝑙2 2 )𝑢2 = 0

2
 ME 4440-5540 Lecture 35

1) Lagrange’s Equation
This method is very general and you don’t need a F.B.D. and doesn’t matter if we are working
with flexible structure or rigid body. Lagrange’s dynamics, itself is an advance course and here
we will briefly talk about this approach (we will not go through all details, steps, energy
methods, etc.).

If you have a system with 𝑞𝑖 number of degrees of freedom (generalized coordinates), if you
use this approach, you can drive equation of motion for all degrees of freedom. Lagrange
stablish by “L” which is equal to difference between kinetic energy and potential energy.

𝐿 = 𝑘𝑒 − 𝑝𝑒
Following formula which is derived based on energy method can be used directly to drive the
equation of motion.

𝑑 𝜕𝐿 𝜕𝐿
( )− =0
𝑑𝑡 𝜕𝑞̇ 𝑖 𝜕𝑞𝑖

Note: for two degrees of freedom system 𝑖 = 1,2, and “𝑞” is equivalent “𝑢” (displacement in
this case)
Note: This approach same as all other energy methods is only applicable to conservative
systems (no damping or damping is negligible).

Example 41: The following figure shows a vehicle (𝑚1 ) connected to a wall with a spring with
the stiffness of 𝑘 and moving along direction 𝑢1 . Mass (𝑚2 ) with a massless rod with length
of 𝑙 is attached to this vehicle. When the vehicle is moving, 𝑚2 is also moving (𝑢2 ). In this
case we have two degrees of freedom (two generalized coordinates). Find two equations of
motion for this system.

3
 ME 4440-5540 Lecture 35

First, we have to set up kinetic and potential energies.

In this case, we have both translational and rotational motions. To write the kinetic &
potential energies, we will find the translational coordinates of rotational motion of 𝑚2 (𝑧𝑥
& 𝑧𝑦 )

1 1 1
𝑘𝑒 = 2 𝑚𝑣 2 = 2 𝑚1 𝑢̇ 1 2 + 2 𝑚2 (𝑧̇𝑥 2 + 𝑧̇𝑦 2 )

𝑧𝑥 = 𝑢1 + 𝑙 𝑠𝑖𝑛 𝑢2 𝑧̇𝑥 = 𝑢̇ 1 + 𝑙 𝑢̇ 2 𝑐𝑜𝑠 𝑢2

𝑧𝑦 = 𝑙 𝑐𝑜𝑠 𝑢2 𝑧̇𝑦 = −𝑙 𝑢̇ 2 𝑠𝑖𝑛 𝑢2

1 1
𝑘𝑒 = 2 𝑚1 𝑢̇ 1 2 + 2 𝑚2 [(𝑢̇ 1 + 𝑙 𝑢̇ 2 cos 𝑢2 )2 + (−𝑙 𝑢̇ 2 sin 𝑢2 )2 ]

1 1
𝑘𝑒 = 2 𝑚1 𝑢̇ 1 2 + 2 𝑚2 (𝑢̇ 1 2 + 𝑙 2 𝑢̇ 2 2 + 2𝑙𝑢̇ 1 𝑢̇ 2 cos 𝑢2 )

ℎ = 𝑙 − 𝑙 cos 𝑢2 = 𝑙(1 − cos 𝑢2 )

1 1
𝑝𝑒 = 2 𝑘𝑥 2 + 𝑚𝑔ℎ = 2 𝑘𝑢1 2 + 𝑚2 𝑔𝑙(1 − cos 𝑢2 )

𝐿 = 𝑘𝑒 − 𝑝𝑒
𝑑 𝜕𝐿 𝜕𝐿
( )− =0
𝑑𝑡 𝜕𝑞̇ 𝑖 𝜕𝑞𝑖
𝑖 = 1:
𝜕𝐿
= 𝑚1 𝑢̇ 1 + 𝑚2 𝑢̇ 1 + 𝑚2 𝑙𝑢̇ 2 cos 𝑢2 (There is no 𝑢̇ 1 in 𝑝𝑒)
𝜕𝑢̇ 1
1
𝜕𝐿
If 𝑢2 be a small angle = 𝑚1 𝑢̇ 1 + 𝑚2 𝑢̇ 1 + 𝑚2 𝑙𝑢̇ 2 cos 𝑢2 ≅ (𝑚1 + 𝑚2 )𝑢̇ 1 + 𝑚2 𝑙𝑢̇ 2
𝜕𝑢̇ 1
𝑑 𝜕𝐿
( ) = (𝑚1 + 𝑚2 )𝑢̈ 1 + 𝑚2 𝑙𝑢̈ 2
𝑑𝑡 𝜕𝑢̇ 1

4
 ME 4440-5540 Lecture 35

𝜕𝐿
= −𝑘𝑢1 (There is no 𝑢1 in 𝑘𝑒, “−” sign because of 𝐿 = 𝑘𝑒 − 𝑝𝑒)
𝜕𝑢1

𝑑 𝜕𝐿 𝜕𝐿
( )− = (𝑚1 + 𝑚2 )𝑢̈ 1 + 𝑚2 𝑙𝑢̈ 2 + 𝑘𝑢1 = 0
𝑑𝑡 𝜕𝑢̇ 1 𝜕𝑢1

𝑖 = 2:
𝜕𝐿
= 𝑚2 𝑙 2 𝑢̇ 2 + 𝑚2 𝑙 𝑢̇1 cos 𝑢2 (There is no 𝑢̇ 2 in 𝑝𝑒)
𝜕𝑢̇ 2
1
𝜕𝐿
If 𝑢2 be a small angle = 𝑚2 𝑙 2 𝑢̇ 2 + 𝑚2 𝑙 𝑢̇1 cos 𝑢2 ≅ 𝑚2 𝑙 2 𝑢̇ 2 + 𝑚2 𝑙 𝑢̇1
𝜕𝑢̇ 2
𝑑 𝜕𝐿
( ) = 𝑚2 𝑙 2 𝑢̈ 2 + 𝑚2 𝑙𝑢̈ 1
𝑑𝑡 𝜕𝑢̇ 2

𝜕𝐿
= −𝑚2 𝑙𝑢̇ 1 𝑢̇ 2 𝑠𝑖𝑛 𝑢2 − 𝑚2 𝑔𝑙 𝑠𝑖𝑛 𝑢2
𝜕𝑢2 0 𝑢2 𝑢2
𝜕𝐿
If 𝑢2 be a small angle = −𝑚2 𝑙𝑢̇ 1 𝑢̇ 2 𝑠𝑖𝑛 𝑢2 − 𝑚2 𝑔𝑙 𝑠𝑖𝑛 𝑢2 ≅ −𝑚2 𝑔𝑙𝑢2
𝜕𝑢2

Multiplication of two small numbers

𝑑 𝜕𝐿 𝜕𝐿
( )− = 𝑚2 𝑙 2 𝑢̈ 2 + 𝑚2 𝑙𝑢̈ 1 + 𝑚2 𝑔𝑙𝑢2 = 0
𝑑𝑡 𝜕𝑢̇ 2 𝜕𝑢2

In this case, if we used Newton’s law, it will be so much harder to solve!

5
 ME 4440-5540 Lecture 36

Two Degree of Freedom Systems

Matrix form representation of the equation of motion:
The Equations of motion in terms of relative displacements (𝑢1 &𝑢2 ) can be write as a matrix
form:
𝑚1 𝑢̈ 1 + (𝑐1 + 𝑐2 )𝑢̇ 1 − 𝑐2 𝑢̇ 2 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 𝑓1 − 𝑚1 𝑦̈
𝑚2 𝑢̈ 2 − 𝑐2 𝑢̇ 1 + 𝑐2 𝑢̇ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 𝑓2 − 𝑚2 𝑦̈

𝑚1 0 𝑢̈ 1 𝑐 + 𝑐2 −𝑐2 𝑢̇ 1 −𝑘 + 𝑘2 −𝑘2 𝑢1 𝑓 − 𝑚1 𝑦̈
[ ]{ } + [ 1 ]{ } + [ 1 ] {𝑢 } = { 1 }
0 𝑚2 𝑢̈ 2 −𝑐2 𝑐2 𝑢̇ 2 −𝑘2 𝑘2 2 𝑓2 − 𝑚2 𝑦̈

Mass Acceleration Damping Velocity Stiffness Displacement Forcing

Matrix Vector Matrix Vector Matrix Vector Function
([𝑀] or 𝑀) 𝑢̇ ([𝐾] or 𝑘) 𝑢 Vector
~ 𝑢̈
~
([𝐶] or 𝐶)
~ ~ ~ ~ 𝑓
(Diagonal matrix) ~

So, we can write the equation of motion for M.D.O.F systems like the scalar form that we
had for S.D.O.F system. However, you have to remember, it is a matrix equation and it is a
coupled system of differential equation and you have transform it to the normal coordinate to
decoupled and then solve it.

𝑀. 𝑢̈ + 𝐶𝑢̇ + 𝐾𝑢 = 𝑓
~ ~ ~~ ~~ ~

State-Space formulation: The second order differential equation is not easy to solve, so we
will do a variable substitution and convert it to the first order differential equation.

Just imagine we have:

𝑢1 𝑢̇ 𝑢̈
𝑢 = {𝑢 } , 𝑢̇ = { 1 } , 𝑢̈ = { 1 }
2 𝑢̇ 2 𝑢̈ 2

If we define Z vector as:

𝑧1 𝑢1 𝑢̇ 1 𝑢̈
𝑍 = {𝑧 } , 𝑧1 = {𝑢 } , 𝑧2 = { } 𝑧̇2 = { 1 }
2 2 𝑢̇ 2 𝑢̈ 2

Note: From matrix algebra multiplication of a matrix to its invers would be equal to an
identity matrix.

1
 ME 4440-5540 Lecture 36

1 ⋯ 0
𝐴. 𝐴−1 = [𝐼] = [ ⋮ ⋱ ⋮]
0 ⋯ 1

Based on above information, if we multiply the equation of motion by 𝑀−1:

𝑢̈ + 𝑀−1 𝐶𝑢̇ + 𝑀−1 𝐾𝑢 = 𝑀−1 𝑓

~ ~ ~~ ~ ~~ ~ ~
𝑢̈ = −𝑀−1 𝐶𝑢̇ − 𝑀 −1 𝐾𝑢 + 𝑀−1 𝑓
~ ~ ~~ ~ ~~ ~ ~
Now, if we use State-Space formulation, we can convert it to the first order differential
equation. Then, just we need to take one integration to find 𝑧1 & 𝑧2 .

𝑧̇ = −𝑀−1 𝐶𝑧2 − 𝑀−1 𝐾𝑧1 + 𝑀−1 𝑓

~2 ~ ~~ ~ ~~ ~ ~
In general form we can write it:
𝑢1
𝑧1 𝑢2 0 𝐼 0
𝑧 = {𝑧 } = {𝑢̇ } , 𝑧̇ = 𝐴𝑧 + 𝐵𝑓 , 𝐴 = [ −1 ] , 𝐵 = [ −1 ]
~ 2 1 ~ ~~ ~~ ~ −𝑀−1 𝐾 −𝑀 𝐶 ~ 𝑀
𝑢̇ 2
MDOF: In this part we will see all tools to cover the concepts and formulation for two degree
of freedom system which would be same formulation for multi degree of freedom system (just
matrix has larger dimension). We have Undamped, damped, free vibration and forced vibration
same as S.D.O.F. system but in this course we just cover Undamped free vibration case.
Undamped Free Vibration:

From before, the equation of motions would be:

𝑚1 𝑢̈ 1 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 0
𝑚2 𝑢̈ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 0
These are two simultaneously linear homogeneous differential equations. In S.D.O.F. system,
we assumed the solution would be a harmonic function, took first and second derivative of
that and put it back in the equation and try to find constant value “𝐴” and “𝜔”. We will do
exactly same thing for two D.O.F. system, with only difference we have two harmonic
functions.

2
 ME 4440-5540 Lecture 36

𝑢1 = 𝐴1 𝑠𝑖𝑛 𝜔𝑡 & 𝑢2 = 𝐴2 𝑠𝑖𝑛 𝜔𝑡

If we take first and second derivatives and put them back in the equations of motion, we will
have a homogenous set of equations.

𝑚1 (−𝜔2 𝐴1 )𝑠𝑖𝑛 𝜔𝑡 + (𝑘1 + 𝑘2 )𝐴1 𝑠𝑖𝑛 𝜔𝑡 − 𝑘2 𝐴2 𝑠𝑖𝑛 𝜔𝑡 = 0

𝑠𝑖𝑛 𝜔𝑡 ≠ 0
𝑚2 (−𝜔2 𝐴2 )𝑠𝑖𝑛 𝜔𝑡 − 𝑘2 𝐴1 𝑠𝑖𝑛 𝜔𝑡 + 𝑘2 𝐴2 𝑠𝑖𝑛 𝜔𝑡 = 0

(−𝜔2 𝑚1 + 𝑘1 + 𝑘2 )𝐴1 − 𝑘2 𝐴2 = 0
{
−𝑘2 𝐴1 + (−𝜔2 𝑚2 + 𝑘2 )𝐴2 = 0

From matrix algebra there is no unique solution for homogeneous set of simultaneous
equations. In other words, you cannot solve for two variables 𝐴1 & 𝐴2 because they are not
independent. Only we can do is finding the ratios of those two variables!

There are many different ways that we can solve system of homogeneous, simultaneous
equation. One way of solving a set of homogeneous, simultaneous equation is Cramer’s rule.
Example:
𝑎1 𝑥1 + 𝑎2 𝑥2 = 0
{
𝑏1 𝑥1 + 𝑏2 𝑥2 = 0

𝑎1 𝑎2 𝑥1 0 𝑑𝑒𝑡|𝐴𝑖 |
[𝑏 𝑏2 ] {𝑥2 } = {0} 𝑥𝑖 =
1 𝑑𝑒𝑡|𝐴|

𝐴 0
0 𝑎1 We will have trivial solution, for
𝑑𝑒𝑡| |
For instance: 𝑥2 =
0 𝑏1 𝑥1 & 𝑥2 = 0 or non-trivial
𝑎 𝑏1
𝑑𝑒𝑡| 1 | 𝑎 𝑏1
𝑎2 𝑏2 solution for: 𝑑𝑒𝑡 | 1 |=0
𝑎2 𝑏2
0
0 𝑎1
𝑑𝑒𝑡 | | 0
0 𝑏1 That means 𝑥2 ≠ 0
𝑥2 = =
𝑎 𝑏1 0 0
𝑑𝑒𝑡 | 1 |
𝑎2 𝑏2

For our case:

−𝜔2 𝑚1 + 𝑘1 + 𝑘2 −𝑘2
𝑑𝑒𝑡 | 2 |=0
−𝑘2 −𝜔 𝑚2 + 𝑘2

3
 ME 4440-5540 Lecture 36

This determinant consists of some system parameters (𝑚1 , 𝑚2 , 𝑘1 , 𝑘2 ), and it is the function of
frequency (𝜔). We show this equation with 𝐷(𝜔) = 0 and it is named: frequency equation.
This would be a quadratic equation in terms of (𝜔2 ) and it will have two solutions for
frequencies corresponding to this two degree of freedom system.

(−𝜔2 𝑚1 + 𝑘1 + 𝑘2 )(−𝜔2 𝑚2 + 𝑘2 ) − 𝑘2 2 = 0

.
.
.

1 𝑘1 + 𝑘2 𝑘2 1 𝑘1 + 𝑘2 𝑘2 2 𝑘1 𝑘2
𝜔1,2 2
= ( + )∓√ ( + ) −
2 𝑚1 𝑚2 4 𝑚1 𝑚2 𝑚1 𝑚2

These roots are the frequency associated with those two degrees of freedom. Now to solve for
𝐴1 & 𝐴2 , we need to substitute these 𝜔 values to the main equations. However, when we have
a homogenous set of equation, the equations are not independent! So, you can use only one of
those equations and divide it by one of the variables (𝐴1 or 𝐴2 ) and we just able to find the
𝐴 𝐴
ratio of those two variables (relative value of those variables not the actual value (𝐴1 or 𝐴2 )).
2 1
𝐴2 𝐴
For instance for 𝜔1 you will find a value for 𝐴 and for 𝜔2 you will find another value for 𝐴2 .
1 1
𝐴2
Usually, we normalized with putting 𝐴1 = 1. Each frequency shows how the structure
𝐴1
deformed based on that frequency. The displacement based on 𝜔1 and its corresponding ratio
𝐴 𝐴
of 𝐴2 is named “first mode” and displacement related to 𝜔2 and its corresponding ratio of 𝐴2 is
1 1
named “second mode”. In two degree of freedom we have two frequencies which are the
natural frequencies of the system. The first frequency is called “fundamental frequency” and
higher frequencies correspond to higher modes.

Example: If we have a beam with two lump mass the first and second modes would be:

First Mode Second Mode

4
 ME 4440-5540 Lecture 37

Two Degree of Freedom Systems

Example 42: There is a two story building with the weight of 𝑊1 & 𝑊2 for first and second
floors respectively. The columns in each floor has same stiffness. Assume 𝑊1 = 𝑊2 = 𝑊, 𝐼1 =
2𝐼2 = 𝐼 (second moment of area for the columns), and 𝑙1 = 𝑙2 = 𝑙.

Free Body Diagram:

12𝐸𝐼
𝑘1 = 2 ( )
𝑙3
𝑘1 12𝐸𝐼
𝑘2 = =( )
2 𝑙3

1
 ME 4440-5540 Lecture 37

From the equation of motion that we have for undamped-free vibration:

𝑚1 𝑢̈ 1 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 0
𝑚2 𝑢̈ 2 − 𝑘2 𝑢1 + 𝑘2 𝑢2 = 0

𝑢̈ 1 + 3𝐵𝑢1 − 𝐵𝑢2 = 0 12𝐸𝐼𝑔

{ , 𝐵=( )
𝑢̈ 2 + 𝐵𝑢2 − 𝐵𝑢1 = 0 𝑊𝑙3

𝑢1 = 𝐴1 𝑠𝑖𝑛 𝜔𝑡
𝑢2 = 𝐴2 𝑠𝑖𝑛 𝜔𝑡
−𝐴1 𝜔2 + 3𝐵𝐴1 − 𝐵𝐴2 = 0
{
−𝐴2 𝜔2 + 𝐵𝐴2 − 𝐵𝐴1 = 0
If we substitute 𝑢1 & 𝑢2 in equations of motion and find the determinant, we will have:
2
𝑑𝑒𝑡 |−𝜔 + 3𝐵 −𝐵 | = 0
−𝐵 −𝜔2 + 𝐵

We need to solve for the roots of this determinant to find the frequencies of system.
𝜔4 − 4𝐵𝜔2 + 2𝐵 2 = 0

0.586 𝐵
𝜔2 = {
3.414 𝐵

𝐸𝐼𝑔
* As a numerical example, let’s assume 𝑊𝑙3 = 2, then we will have:

𝜔1 = 3.75 𝑟𝑎𝑑/𝑠
𝜔2 = 9.05 𝑟𝑎𝑑/𝑠
Now, for finding 𝐴1 & 𝐴2 , we have to substitute these 𝜔 values in one of the equation of
motions (two equations are dependent) and divide it by one of the variables (𝐴1 ).
𝐴
−𝐴1 𝜔2 + 3𝐵𝐴1 − 𝐵𝐴2 = 0 −𝜔2 + 3𝐵 − 𝐵 (𝐴2 ) = 0
1

−𝜔2 +3𝐵
If we take 𝐴1 = 1 −𝜔 2 +3𝐵
𝐴2 = 𝐴1 𝐴2 =
𝐵 𝐵

𝐴1 = 1
𝜔1 2 = 0.586 𝐵 {
𝐴2 = 2.414
𝐴1 = 1
𝜔2 2 = 3.414 𝐵 {
𝐴2 = −0.414
2
 ME 4440-5540 Lecture 37

𝜔1 and 𝜔2 are the natural frequencies and 𝐴1 & 𝐴2 are the shapes corresponding to each
natural frequencies which we call them mode shapes.

Node

Mode Shape 1 Mode Shape 2

Node: There is point along the structure that doesn’t move (stay stationary). In other words, node
is a point on the structure which is experiences zero deformation (is not subjected to any dynamic
force, no displacement, no stresses, etc). The number of nodal points is depends on what modes of
vibration we are consider, higher modes having more nodal points.

Summary of Steps:
1) Set up the equation of motion
2) Consider a solution of type 𝑢𝑖 = 𝐴𝑖 𝑠𝑖𝑛 𝜔𝑡 (for two D.O.F. 𝑖 = 1,2) and substitute it in
equation of motion
3) Consider the determinant of coefficient for part (2) equal zero (𝐷(𝜔) = 0 , frequency
equation)
4) Find the roots of frequency equation (frequency of the system)
5) Substitute these roots in one of the equations in part (2) and for each value of (𝜔) find
𝐴
the corresponding relative deformed shapes (i. e. 𝐴2 )
1
6) Plot 𝐴1 & 𝐴2 (𝐴1 = 1) for each 𝜔 (mode shapes)

3
 ME 4440-5540 Lecture 37

Mode Shape
For a free vibration of an undamped system, if we follow the steps:

1) 𝑀. 𝑢̈ + 𝐾𝑢 = 0
~ ~ ~~ ~
𝑀 : Mass matrix
~
𝐾 : Stiffness matrix
~
𝑢 : Vector of displacement
~
2) 𝑢𝑖 = ∅𝑖 𝑠𝑖𝑛 𝜔𝑡 (We use ∅𝑖 rather than 𝐴𝑖 )

3) Substitute (2) in to (1) and set the eigenvalue problem

(𝐾 − 𝜔2 . 𝑚). {Փ𝑖 } = {0}

~ ~
Note: From the matrix algebra the following equation is named eigenvalue equation:
(𝐴 − 𝜆. 𝐼). {𝑥} = {0}
~ ~
We can rewrite the equation from (3) by pre-multiply by 𝑚−1:

Eigenvalues
(𝑚−1 𝐾). {Փ𝑖 } = 𝜔2 . {Փ𝑖 }
~ ~
∅
{Փ𝑖 } = { 1𝑖 }
∅2𝑖
D.O.F.
Frequency

Eigenvectors
(Mode shapes)

For instance ∅11 is shape for D.O.F. “1” for mode frequency “1”. In other words, for two degree
of freedom system, we will have ∅11 & ∅21 for 𝜔1 and ∅21 & ∅22 for 𝜔2 as we discussed before.
The 𝜔′𝑠 (frequencies) are “Eigenvalues” and ∅′𝑠 (Mode Shapes) are “Eigenvectors”.
∅11
𝜔1 : { }
∅21
∅21
𝜔2 : { }
∅22

{Փ} = [{Փ1 }, {Փ2 }, … . , {Փ𝑛 }]

Mode shape matrix

4
 ME 4440-5540 Lecture 37

4) From (𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡𝑠) = 0 , find the 𝜔′𝑠 (frequencies)

5) Substitute the 𝜔′𝑠 in equation (3), and find corresponding mode shape (∅′𝑠)

∅
{Փ𝑖 } = { 1𝑖 }
∅2𝑖

5
 ME 4440-5540 Lecture 38

Two Degree of Freedom Systems

Properties of Mode Shapes:
1. Mode Shapes are invariant with respect to a constant. That means if you have a mode shape
(modal Metrix) Փ𝑖 , it would be same as 𝑐 × Փ𝑖 .
∅1𝑖 𝑐∅
{Փ𝑖 } = { } {Փ𝑖 } = { 1𝑖 }
∅2𝑖 𝑐 ∅2𝑖
2.
2.1. Mode shapes are orthogonal with respect to each other relative to mass or stiffness
matrices. Based on definition of orthogonality, we will have:

{Փ1 }𝑇 [𝑚]{Փ2 } = 0
~
{Փ1 }𝑇 [𝑘]{Փ2 } =0
~
And in general:
{Փ𝑖 }𝑇 [𝑚]{Փ𝑗 } = 0
~
(𝒊 ≠ 𝒋)
{Փ𝑖 }𝑇 [𝑘]{Փ𝑗 } =0
~

2.2. Mode shapes are not orthogonal with respect to themselves & relative to mass or
stiffness matrices. However, in relation to mass matrix that would be equal to a constant
matrix (a diagonal matrix) and for stiffness would be equal to multiplication of that
constant matrix to 𝜔2 .
In general:

𝑀1 0
{Փ𝑖 }𝑇 [𝑚]{Փ𝑖 } = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 = [ ⋱ ]
(𝒊 ≠ 𝒋)
0 𝑀𝑖

𝑀1 0
{Փ𝑖 }𝑇 [𝑘]{Փ𝑖 } = 𝜔2 [ ⋱ ]
0 𝑀𝑖

𝑀1 0
[ ⋱ ] = 𝐺𝑒𝑛𝑒𝑟𝑎𝑙𝑖𝑧𝑒𝑑 𝑀𝑎𝑠𝑠 𝑀𝑎𝑡𝑟𝑖𝑥 (𝑚𝑜𝑑𝑎𝑙 𝑀𝑎𝑠𝑠 𝑀𝑎𝑡𝑟𝑖𝑥)
0 𝑀𝑖

1
 ME 4440-5540 Lecture 38

Note: The property # 2 is very important and in order to verify if the mode shape and frequency
that we calculated for the system are correct the relationships in the property number #2 must
hold.
Normalization of Mode Shapes
Mode shape represents the quality shape of the deformation. There are several approaches to
normalize the mode shapes.
1. Set ∅1 = 1 & find other relative displacement with respect to that.

𝑥1
𝑥1 = 1 𝑥2 =?
𝑥2

2. Consider the maximum value of each mode & set that equal to 1.

3. Set modal mass corresponding to each mode equal to 1.

1

𝑀1 0
Find the relation for the rest of them
{Փ𝑖 }𝑇 [𝑘]{Փ𝑖 } = [ ⋱ ]
0 𝑀𝑖

Note: Between these three approaches, last one is most commonly used especially in finite element
code because need less computation.

Example 43: Consider the following two degree of freedom system. This system is consist of a
rigid bar attached to a spring at one end and pivot at the other end. A mass (𝑚) with a spring is
connected to the bar in the point with a distance of 𝑙1 from the pivot. The bar has a mass of 𝑚𝑏𝑎𝑟
and total length of 𝑙2 . Whole system start oscillating. Find the mode shapes, corresponding
frequencies, and check if the mode shape are orthogonal with respect to each other. Assume 𝑦 <
𝑙1 𝜃 (the spring is in compression).
There are some relations between parameters in this problem (a, b, c are constant values):
𝑙1 = 𝑎𝑙2 = 𝑙
𝑘2 = 𝑏𝑘1
𝑚𝑏𝑎𝑟 = (3𝑚). 𝑐
𝑘1
Ω2 = 𝑚

2
 ME 4440-5540 Lecture 38

Free body diagram:

1 2
A

Step 1: Equation of Motions

1)
𝑚𝑦̈ − 𝑓𝑘1 = 0

𝑓𝑘1 = 𝑘1 (𝑙1 𝜃 − 𝑦)

𝑚𝑦̈ + 𝑘1 𝑦 − 𝑘1 𝑙1 𝜃 = 0 1st Equation of motion

2)
∑ 𝑀𝐴 = 0

𝐼𝜃̈ + 𝑓𝑘1 (𝑙1 ) + 𝑓𝑘2 (𝑙2 ) = 0

1
𝑓𝑘2 = 𝑘2 (𝑙2 𝜃) & 𝐼 = 3 𝑚𝑏𝑎𝑟 𝑙2 2

𝐼𝜃̈ + 𝑘1 (𝑙1 𝜃 − 𝑦)𝑙1 + 𝑘2 (𝑙2 𝜃)𝑙2 = 0

𝐼𝜃̈ + (𝑘1 𝑙1 2 + 𝑘2 𝑙2 2 )𝜃 − 𝑘1 𝑙1 𝑦 = 0 2nd Equation of motion

3
 ME 4440-5540 Lecture 38

Step 2: Find the mode shapes

𝑚. 𝑢̈ + 𝑘𝑢 = 0
~ ~ ~~ ~
𝑚 0
𝑚=[ ]
~ 0 𝐼
𝑘1 −𝑘1 𝑙1
𝑘=[ ]
~ −𝑘1 𝑙1 𝑘1 𝑙1 2 + 𝑘2 𝑙2 2
𝑢1 𝑦
𝑢 = {𝑢 } = { }
2 𝜃
Based on the given relation between parameters, equations of motion would be:
𝑚𝑦̈ + 𝑘1 𝑦 − 𝑘1 𝑙1 𝜃 = 0 𝑙1 = 𝑎𝑙2 = 𝑙
𝑦̈ + Ω2 𝑦 − Ω2 𝑙𝜃 = 0 𝑘2 = 𝑏𝑘1
𝑚𝑏𝑎𝑟 = (3𝑚). 𝑐
𝐼𝜃̈ + (𝑘1 𝑙1 2 + 𝑘2 𝑙2 2 )𝜃 − 𝑘1 𝑙1 𝑦 = 0 𝑘1
Ω2 = 𝑚
2
𝑎 +𝑏 𝑎 2
After simplification 𝜃̈ + Ω2 ( 𝑐 ) 𝜃 − Ω2 ( 𝑐𝑙 ) 𝑦 = 0 1
𝐼 = 3 𝑚𝑏𝑎𝑟 𝑙2 2

𝑦̈ + (Ω2 )𝑦 − (Ω2 𝑙)𝜃 = 0

{ ̈ 2
𝑎2 2
𝑎2 + 𝑏
𝜃 − Ω ( )𝑦 + Ω ( )𝜃 = 0
𝑐𝑙 𝑐

𝑢̈ 1 + (Ω2 )𝑢1 − (Ω2 𝑙)𝑢2 = 0

{ 2
𝑎2 2
𝑎2 + 𝑏
𝑢̈ 2 − Ω ( ) 𝑢1 + Ω ( ) 𝑢2 = 0
𝑐𝑙 𝑐

In the matrix form:

1 −𝑙 𝑢
1 0 𝑢̈ 1 𝑎 + 𝑏] { 1 } = 0
[ ] { } + Ω [ 𝑎2
2 2
𝑢2
0 1 𝑢̈ 2 − ~
𝑐𝑙 𝑐
𝑚
~ 𝑘
~

4
 ME 4440-5540 Lecture 38

Step 3: Find the corresponding frequencies

𝑑𝑒𝑡(𝐾 − 𝜔2 . 𝑚) = 0 Find 𝜔′𝑠
~ ~

1 −𝑙 1 − 𝜔2 −𝑙
1 0
𝑑𝑒𝑡 (Ω [ 𝑎2
2 2
𝑎 + 𝑏] − 𝜔2 [ ]) = 𝑑𝑒𝑡 (Ω2 [ 𝑎2 2
𝑎 +𝑏 ]) = 0
− 0 1 − − 𝜔2
𝑐𝑙 𝑐 𝑐𝑙 𝑐
1 − 𝜔2 −𝑙
2
Ω ≠0 𝑑𝑒𝑡 ([ 𝑎2 𝑎2 +𝑏 ]) = 0
− 𝑐𝑙 − 𝜔2
𝑐

After simplification:

4
𝑎2 + 𝑏 + 𝑐
2
𝑏
𝜔 −𝜔 ( )+ = 0
𝑐 𝑐
To be more realistic, let’s give some values to the parameters:
𝑙1 = 𝑙 = 12.5 𝑐𝑚 = 0.125 𝑚
𝑙2 = 50 𝑐𝑚 = 0.5 𝑚
𝑚 = 1 𝑘𝑔
𝑚𝑏𝑎𝑟 = 0.3 𝑘𝑔
𝑘1 = 34 𝑁/𝑚
𝑘2 = 17 𝑁/𝑚
𝐼 = 0.025 𝑚4
𝑘1
Ω2 = = 34
𝑚
𝑙
𝑎 = 𝑙 = 0.25
2

𝑏 = 0.5
𝑚𝑏𝑎𝑟
𝑐= = 0.1
3

𝑎2 +𝑏+𝑐
= 6.625
𝑐

𝑎2 +𝑏+𝑐 𝑏
𝜔4 − 𝜔2 ( )+𝑐 =0 𝜔4 − 6.625 𝜔2 + 5 = 0 𝜔2 = 0.869 , 5.756
𝑐

𝜔1 = 0.932 & 𝜔2 = 2.4

5
 ME 4440-5540 Lecture 38

Step 3: Find the mode shape

𝑦
1 − 𝜔2 −𝑙 𝑢1
([ 𝑎2 2
𝑎 +𝑏 2
]) {𝑢 } = 0
− −𝜔 2
𝑐𝑙 𝑐
𝜃

𝜔1 = 0.932 𝑦=1 , 𝜃 = 1.048 𝑚𝑜𝑑𝑒 𝑠ℎ𝑎𝑝𝑒 1

𝜔2 = 2.4 𝑦=1 , 𝜃 = −38.048 𝑚𝑜𝑑𝑒 𝑠ℎ𝑎𝑝𝑒 2

Step 4: Check if the mode shapes are orthogonal with respect to each other

∅11 1
{Փ1 } = { }={ }
∅12 1.048

∅21 1
{Փ2 } = { }={ }
∅22 −38.048

{Փ}𝑇 [𝑚]{Փ}
~ ~ ~

1 1.048 1 0 1 1 1.027 0
[ ][ ][ ]=[ ]
1 −38.048 0 0.025 1.048 −38.048 0 40

𝐺𝑒𝑛𝑒𝑟𝑎𝑙𝑖𝑧𝑒𝑑 𝑀𝑎𝑠𝑠
𝑀𝑎𝑡𝑟𝑖𝑥

{Փ1 }𝑇 [𝑚]{Փ2 }

[1 1.048] [1 0 1
][ ] = [0]
0 0.025 −38.048
Both conditions are satisfied and the mode shape are orthogonal with respect to each other.

6
 ME 4440-5540 Lecture 39

Finding the Solution for Free Vibration 2nd DOF of an Undamped System

Look at following system.

𝑢1 (𝑡) 𝑢2 (𝑡)

Free body diagram:

𝑢1 (𝑡) 𝑢2 (𝑡)

𝑘1 𝑢1 𝑘2 (𝑢2 − 𝑢1 ) 𝑘3 𝑢2

Let’s go over all steps for this problem.

For this system, the equations of motion will be:
𝑚1 𝑢̈ 1 + (𝑘1 + 𝑘2 )𝑢1 − 𝑘2 𝑢2 = 0
𝑚2 𝑢̈ 2 − 𝑘2 𝑢1 + (𝑘2 + 𝑘3 )𝑢2 = 0
If 𝑚1 & 𝑚2 can oscillate harmonically with same frequency and phase but different amplitude
then we can say:
𝑢1 = Փ1 𝑐𝑜𝑠 (𝜔𝑡 + 𝜃)
𝑢2 = Փ2 𝑐𝑜𝑠 (𝜔𝑡 + 𝜃)
Փ1 & Փ2 : The maximum amplitudes of 𝑢1 & 𝑢2
With replacing these equations in the equations of motion and simplifying, we will reach:
[−𝑚1 𝜔2 + (𝑘1 + 𝑘2 )]Փ1 − 𝑘2 Փ2 = 0
−𝑘2 Փ1 + [−𝑚2 𝜔2 + (𝑘2 + 𝑘3 )]Փ2 = 0
As you seen before, for finding Փ1 & Փ2 , we have to write these equations and make
determinant of coefficients equal zero.
[−𝑚1 𝜔2 + (𝑘1 + 𝑘2 )] −𝑘2
𝑑𝑒𝑡 [ 2 ]=0
−𝑘2 [−𝑚2 𝜔 + (𝑘2 + 𝑘3 )]

1
 ME 4440-5540 Lecture 39

From that determinant, we can find two natural frequencies (𝜔) for system:
1 (𝑘1 + 𝑘2 )𝑚2 + (𝑘2 + 𝑘3 )𝑚1
𝜔1 2 , 𝜔2 2 = [ ]
2 𝑚1 𝑚2
1 (𝑘1 + 𝑘2 )𝑚2 + (𝑘2 + 𝑘3 )𝑚1 2 (𝑘1 + 𝑘2 )(𝑘2 + 𝑘3 ) − 𝑘2 2
± √[ ] −4[ ]
2 𝑚1 𝑚2 𝑚1 𝑚2

For each of these natural frequencies, we will have different mode shapes (Փ1 & Փ2 ).
∅11
𝜔1 {Փ1 } = { }
∅12
∅21
𝜔2 {Փ2 } = { }
∅22
However, two equations of motions are homogenous set of equation and they are not independent!
∅ ∅
So, we are only able to find the ratio of 𝑟1 = ∅12 & 𝑟2 = ∅22 .
11 21

∅12 [−𝑚1 𝜔1 2 + (𝑘1 + 𝑘2 )] 𝑘2

𝑟1 = = =
∅11 𝑘2 [−𝑚2 𝜔1 2 + (𝑘2 + 𝑘3 )]
∅22 [−𝑚1 𝜔2 2 + (𝑘1 + 𝑘2 )] 𝑘2
𝑟2 = = =
∅21 𝑘2 [−𝑚2 𝜔2 2 + (𝑘2 + 𝑘3 )]

So, the normal modes of vibration corresponding 𝜔1 & 𝜔2 will be:

∅11 ∅
{Փ1 } = { } = { 11 }
∅12 𝑟1 ∅11
Modal vectors
∅ ∅
{Փ2 } = { 21 } = { 21 }
∅22 𝑟2 ∅21
Then, solution for each mode will be:

∅11 𝑐𝑜𝑠 (𝜔1 𝑡 + 𝜃1 ) 1st mode

𝑢1 = { }
𝑟1 ∅11 𝑐𝑜𝑠 (𝜔1 𝑡 + 𝜃1 )
∅ 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )
𝑢2 = { 21 } 2nd mode
𝑟2 ∅21 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )
The general solution can be obtained by a linear superposition of the two normal modes:
𝑢 = 𝑐1 𝑢1 + 𝑐2 𝑢2

2
 ME 4440-5540 Lecture 39

𝑐1 & 𝑐2 are constant and can be combine with ∅11 & ∅21 values (constants).
𝑢1 ∅ 𝑐𝑜𝑠 (𝜔1 𝑡 + 𝜃1 ) + ∅21 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )
𝑢 = {𝑢 } = { 11 }
2 𝑟1 ∅11 𝑐𝑜𝑠 (𝜔1 𝑡 + 𝜃1 ) + 𝑟2 ∅21 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )
However, we still need to find the value for ∅11 , ∅21 , 𝜃1 , and 𝜃2 . To obtain these values we have
to use two initial conditions for each mass (the rest of steps is in the book). The final results will
be:

1 {−𝑟2 𝑢̇ 1 (0) + 𝑢̇ 2 (0)}2

∅11 = √{𝑟2 𝑢1 (0) − 𝑢2 (0)}2 +
(𝑟2 − 𝑟1 ) 𝜔1 2

1 {𝑟 𝑢̇ (0) − 𝑢̇ 2 (0)}2
∅21 = √{−𝑟1 𝑢1 (0) + 𝑢2 (0)}2 + 1 1
(𝑟2 − 𝑟1 ) 𝜔2 2

−𝑟2 𝑢̇ 1 (0) + 𝑢̇ 2 (0)

𝜃1 = 𝑡𝑎𝑛−1 [ ]
𝜔1 [𝑟2 𝑢1 (0) − 𝑢2 (0)]

𝑟1 𝑢̇ 1 (0) − 𝑢̇ 2 (0)
𝜃2 = 𝑡𝑎𝑛−1 [ ]
𝜔2 [−𝑟1 𝑢1 (0) + 𝑢2 (0)]

If in this example we have 𝑘1 = 30, 𝑘2 = 5, 𝑘3 = 0, 𝑚1 = 10, 𝑚2 = 1, and initial conditions of

𝑢1 (0) = 1, 𝑢2 (0) = 𝑢̇ 1 (0) = 𝑢̇ 2 (0) = 0,

[−𝑚1 𝜔2 + (𝑘1 + 𝑘2 )] −𝑘2 Փ 0

[ 2 ] { 1} = { }
−𝑘2 [−𝑚2 𝜔 + (𝑘2 + 𝑘3 )] Փ2 0

2
[−10𝜔 + 35 −5 ] {Փ1 } = {0}
−5 −𝜔2 + 5 Փ2 0

2
𝑑𝑒𝑡 [−10𝜔 + 35 −5 ] = 0
−5 −𝜔2 + 5

10𝜔4 − 85𝜔2 + 150 = 0 𝜔1 2 = 2.5 , 𝜔2 2 = 6 𝜔1 = 1.5811 , 𝜔2 = 2.4495

∅12 [−𝑚1 𝜔1 2 + (𝑘1 + 𝑘2 )] [−10 × 2.5 + (35)]

𝑟1 = = = =2
∅11 𝑘2 5

3
 ME 4440-5540 Lecture 39

∅22 [−𝑚1 𝜔2 2 + (𝑘1 + 𝑘2 )] [−10 × 6 + (35)]

𝑟2 = = = = −5
∅21 𝑘2 5

Normal modes will be:

∅11 ∅ 1
{Փ1 } = { } = { 11 } = { } ∅11
∅12 𝑟1 ∅11 2
∅21 ∅ 1
{Փ2 } = { } = { 21 } = { } ∅21
∅22 𝑟2 ∅21 −5
Then, solution for each mode will be:

∅ 𝑐𝑜𝑠 (1.5811𝑡 + 𝜃1 ) + ∅12 𝑐𝑜𝑠 (2.4495𝑡 + 𝜃2 )

𝑢1 = { 11 } 1st mode
𝑟1 ∅11 𝑐𝑜𝑠 (1.5811𝑡 + 𝜃1 )
∅ 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )
𝑢2 = { 21 } 2nd mode
𝑟2 ∅21 𝑐𝑜𝑠 (𝜔2 𝑡 + 𝜃2 )

𝑢1 ∅ 𝑐𝑜𝑠 (1.5811𝑡 + 𝜃1 ) + ∅21 𝑐𝑜𝑠 (2.4495𝑡 + 𝜃2 )

𝑢 = {𝑢 } = { 11 }
2 2∅11 𝑐𝑜𝑠 (1.5811𝑡 + 𝜃1 ) − 5∅21 𝑐𝑜𝑠 (2.4495𝑡 + 𝜃2 )

𝑢1 (0) = 1 = ∅11 𝑐𝑜𝑠 𝜃1 + ∅21 𝑐𝑜𝑠 𝜃2

𝑢2 (0) = 0 = 2∅11 𝑐𝑜𝑠 𝜃1 − 5∅21 𝑐𝑜𝑠 𝜃2
𝑢̇ 1 (0) = 0 = −1.5811∅11 𝑠𝑖𝑛 𝜃1 − 2.4495∅21 𝑠𝑖𝑛 𝜃2
𝑢̇ 2 (0) = 0 = −3.1622∅11 𝑠𝑖𝑛 𝜃1 − 12.2475∅21 𝑠𝑖𝑛 𝜃2

5 2
∅11 𝑐𝑜𝑠 𝜃1 = 7 , ∅21 𝑐𝑜𝑠 𝜃2 = 7 , ∅11 𝑠𝑖𝑛 𝜃1 = 0 , ∅21 𝑠𝑖𝑛 𝜃2 = 0
5 2
∅11 = 7 , ∅21 = 7 , 𝜃1 = 0 , 𝜃2 = 0

5 2
𝑢1 (𝑡) = 𝑐𝑜𝑠 1.5811𝑡 + 𝑐𝑜𝑠 2.4495𝑡
7 7
10 10
𝑢2 (𝑡) = 𝑐𝑜𝑠 1.5811𝑡 − 𝑐𝑜𝑠 2.4495𝑡
7 7