8/14/2019 Vibs1_review_notes.pdf
1/86
Mechanical Vibrations I
Lecture Notes -1- 06/16/06 12:25 PM
Mechanical Vibrations I
m
k c
x(t) f(t)
Author: Allyn W. Phillips, PhD
University of Cincinnati
UC Course Nr.: 20-MECH-480Revision: 16-Jun-2006
Copyright 2001-2006
8/14/2019 Vibs1_review_notes.pdf
2/86
Mechanical Vibrations I
Lecture Notes -2- 06/16/06 12:25 PM
Course Outline
Topics
Review Dynamicso Units
o Complex Numberso Free Body Diagramso Absolute vs. Relative Motiono Differential Equationso Matrices & Linear Algebrao Kinematics
Rolling Contact Coordinate Systems
SDOF / MDOF Conceptso Newtons Method
Frequency, Damping, Mode Shape & Scaling
o Lagranges Method Energy Method (SDOF only)
o Eigenvalue / Eigenvector Solutions Superposition
o Free Vibrationso Forced Vibrations
Steady-State Transient
SDOF / MDOF Applicationso SDOF
Estimating Frequency & Damping
Half Power Log Decrement
Transmissibility Isolation
o MDOF Frequency Response Function Concepts
Stiffness / Compliance Impedance / Mobility
Modal Parameter Estimation Impedance / FRF Methods
8/14/2019 Vibs1_review_notes.pdf
3/86
Mechanical Vibrations I
Lecture Notes -3- 06/16/06 12:25 PM
Course Material Sequence
Introduction to Mechanical Vibrations [AWP #1] Newtons Method [AWP #2]
o Units [TD 0]
o Matrices & Linear Algebra [TMH A]o Free Body Diagrams [TD 2.2, 5.1]
Single and Multiple Bodyo Absolute vs. Relative Motion
Force Balance
Mass absolute acceleration Damping relative velocity Stiffness relative displacement
o Linear & Torsional Rolling Contact
o Equation of Motion [TD 5.1-3]
Single DOF MCK v. RLC Equivalence [TMH 2.6]
Homogeneous Solution of Equation of Motion [AWP #4]o Differential Equations [TMH D]
Linear, Constant Coefficient
Harmonic Functions (sin, cos) [TD 1] Complex Exponentials
o Free Vibrations [TD 2.1, 2.6] Oscillatory Motion [TD 1]
Harmonic Motion Periodic Motion
o Vibration Terminology [TD 1] Frequency Damping [TD 2.6, 2.8, 2.9]
Viscous Coulomb Structural
Mode Shape [TD 5.1] Modal Scaling
o Eigenvalue / Eigenvector Solutions [TD C] Superposition
o Effects of Changing Mass, Stiffness, Damping
o Modal Parameter Estimation Damping
Log Decrement [TD 2.7, AWP #5] Lagranges Method [TD 7.23, TMH B, AWP #6]
o Energy Terms Kinetic
8/14/2019 Vibs1_review_notes.pdf
4/86
Mechanical Vibrations I
Lecture Notes -4- 06/16/06 12:25 PM
Potential Dissipative
o Contrast with Newtons Method Global vs. Local Control Volume
o Equation of Motion Show Equivalence to Newtons Method
o Single DOF reducesEnergy Method [TD 2.3] Conservative System - only
Steady-State Solution of Equation of Motion [AWP #7]o Forced Vibrations [TD 3.1]
Harmonic Excitationo Laplace Transform [TD B]
Converts Differential Equations to Algebraic Equations Transfer Function
o Fourier Transform [TD 1.2, 13.7, DLB] Frequency Response Function [AWP #8]
MCK Variations SDOF Concept Applications [TD 3]o Harmonic Unbalance Force [TD 3.2, AWP #9]o Transmissibility [TD 3.5, AWP #10]o Isolation [TD 3.6, AWP #10]o Dampers
Mass Tuned Absorber
o Modal Parameter Estimation [AWP #12] Frequency Damping
Half Power [TD 3.10, AWP #8] Mode Shape Quadrature
MDOF Systems [AWP #11]o Equations of Motion
Matrix Solutiono Natural Frequencieso Mode Shapes
Transient Solution of Equation of Motion [TD 4, AWP #13]o Forced Vibrations
Impulse Excitation [TD 4.1]
Transient Excitation [TD 4.2]o Convolution [TMH 2.7]
Properties: F[f(t)*g(t)] = F[f(t)] F[g(t)]o Numerical Solutions [AWP #14]
Runge-Kutta
MDOF Detailed Overview (provided for reference) [AWP #15]
8/14/2019 Vibs1_review_notes.pdf
5/86
Mechanical Vibrations I
Lecture Notes -5- 06/16/06 12:25 PM
#1 Introduction to Mechanical Vibrations
All mechanical systems vibrate or undergo oscillatory motion. However, in order to
discuss the behavior of vibrating systems, it is first necessary to establish some
nomenclature. There are several terms used to describe vibrations, among which are:oscillatory motion, periodic motion, harmonic motion, period, and frequency.
Oscillatory motionis any pattern of motion where the system under observation moves
back and forth across some equilibrium position, but does not necessarily have anyparticular repeating pattern.
Periodic motionis a specific form of oscillatory motion where the motion pattern repeatsitself with a uniform time interval. This uniform time interval is referred to as theperiod
and has units of seconds per cycle. The reciprocal of the period is referred to as the
frequencyand has units of cycles per second. This unit combination has been given a
special unit symbol and is referred to asHertz(Hz).
Harmonic motionis a specific form of periodic motion where the motion pattern can be
describe by either a sine or cosine. This motion is also sometimes referred to as simpleharmonic motion. Because the sine or cosine technically uses angles in radians, the
frequency term expressed in the units radians per second ( secrad ). This is sometimes
referred to as the circular frequency. The relationship between the frequency in Hz and
the frequency in secrad is simply the relationship,2 rad
cycle .
Natural frequencyis the frequency at which an undamped system will tend to oscillate
due to initial conditions in the absence of any external excitation. Because there is no
damping, the system will oscillate indefinitely.
Damped natural frequencyis frequency that a damped system will tend to oscillate due to
initial conditions in the absence of any external excitation. Because there is damping in
the system, the system response will eventually decay to rest.
Resonanceis the condition of having an external excitation at the natural frequency of the
system. In general, this is undesirable, potentially producing extremely large systemresponse.
Degrees-of-freedomis the number of independent coordinates necessary to describe them
configuration (state) of a system. It can also be expressed as the number of coordinatesused to describe the configuration of a system minus the number of independent
constraint equations between those coordinates. For example, a point in space has three
degrees-of-freedom ( , , )x y z and a rigid body in space has six ( , , , , , )x y zx y z .
8/14/2019 Vibs1_review_notes.pdf
6/86
Mechanical Vibrations I
Lecture Notes -6- 06/16/06 12:25 PM
A decibelis a log scale measure of relative power. Specifically, it is ten times the log
base ten of the power to a reference power, 1010logref
PdB
P
=
. If the quantity being
referenced is not power based, then the quantities are first squared, yielding2
10 210log
ref
VdBV
=
or 1020logref
VdBV
=
.
An octaveis a frequency range where the upper frequency is twice the lower frequency.
A decadeis a frequency range where the upper frequency is ten times the lowerfrequency.
Bandwidthandfrequency spanboth terms describe a range of frequencies from low to
high. Bandwidth is particularly associated with the energy of the response of a systemand in this course is used in reference to damping (see section eight.)
Having established a small set of basic nomenclature, it is now possible to provide a briefoverview of the direction of this course. In this course, Mechanical Vibrations I, much of
the effort will be spent upon bringing together the many concepts and techniques
developed in previous courses and applying them to the solution of vibrations problems.
Note that, it is possible to solve simple, but representative, problems utilizing only
information previously learned. To help clarify which courses and knowledge are used,the following introductory example has been developed.
8/14/2019 Vibs1_review_notes.pdf
7/86
Mechanical Vibrations I
Lecture Notes -7- 06/16/06 12:25 PM
Introductory Example
For the figure given, assume that the object rollswithout slipping. The objective will be to develop
the equation of motion and determine the naturalfrequency of the system. (A more completeexample would involve a general multi-degree of
freedom system, where the objective would be to
determine; the equations of motion, the systempole, mode shapes, and steady-state response.)
Step 1: Draw the free-body diagram for each independent rigid body in the system. Be
sure to include allexternal forces and moments and their points of application. (Recallthat the internal force resulting from the spring is due to the actual relativemotion of the
ends of the spring. Therefore the coordinatexis measured from thefree lengthof the
spring.) [Mechanics II, Kinematics & Dynamics]
Step 2: Write down all the force and
moment balance equations. [MechanicsII, Kinematics & Dynamics]
) fxF f k x f mx =
) 0yF N mg =
) f cgcg f r f r J + =
Step 3: Identify all relevant coordinateconstraint equations. [Mechanics II, Kinematics & Dynamics]
r= and therefore x r= & x r=
Step 4: Choose the desired independent coordinates and eliminate all the dependent
coordinates by substituting the relevant constraint equations. [Mechanics II, Kinematics
& Dynamics]
) fxF f k x f mx =
) 0yF N mg =
) f cgcgM f r f r J r+ =
Step 5: Evaluate the set of simultaneous differential equations to yield the equations ofmotion. [Mechanics II, Kinematics & Dynamics]
( )f t
( )x t
( )t
k
, cgm J
r
Rolls w/o slipping
=
f
k x
ff
mxcgJ
N
mg
8/14/2019 Vibs1_review_notes.pdf
8/86
Mechanical Vibrations I
Lecture Notes -8- 06/16/06 12:25 PM
)yF N mg=
) 2cgfcgJ
f f xr
=
) 2& cgx cgJ
F M f k x f x mxr
=
Step 5a: Identify the actual equation of motion.
22
cgJm x k x f
r
+ + =
Step 6: Solve the simultaneous set of differential equations of motion. The total solution
( ( )Tx t ) involves the sum of two parts: the particular solution ( ( )p t ) and the
homogeneous solution ( ( )hx t ). [Differential Equations]
Step 6a: The particular solution involves knowing the exact form of the forcing function
( ( )f t ). [Differential Equations]
Step 6b: A second order homogeneous constant coefficient differential equation hascomplex exponentials as its solution. [Differential Equations, Numerical Methods]
( ) th t Ce= and therefore ( ) thx t C e
= & 2( ) th t C e=
2
20cg tJm k Ce
r
+ + =
therefore the system poles (eigenvalues) are:
1,2
2
cg
kj
Jm
r
=
+
and the homogeneous solution is:
1 21 2( ) t th t C e C e = +
Step 6c: Formulate the complete solution ( ( ) ( ) ( )T p ht x t x t = + ) and evaluate the initial
conditions to eliminate the constants of integration. [Differential Equations]
8/14/2019 Vibs1_review_notes.pdf
9/86
Mechanical Vibrations I
Lecture Notes -9- 06/16/06 12:25 PM
Note: if this example had involved more than one independent dynamic coordinate, the
solution would have additionally required the solution of an eigenvalue/eigenvectorproblem. [Differential Equations, Numerical Methods]
8/14/2019 Vibs1_review_notes.pdf
10/86
Mechanical Vibrations I
Lecture Notes -10- 06/16/06 12:25 PM
#2 - Newtons Method ( -or- d'Alemberts Method )
Force Balance
GF M q= ( -or- 0GF M q = )
Moment Balance
/P P G P PJ r M q= + ( -or- / 0P P G P PM J r M q = )
Typical problems with Newtons ( -or- d'Alemberts ) formulation:
Be sure to establish the number of degrees of freedom first and formulate allterms in only those variables. Clearly identify which degrees of freedom are
relative coordinates versus absolute coordinates. Also, clearly identify what willbe the positive direction of motion for each coordinate. Watch out for
rotational/translational problems. State any constraint relationships that relate
independent and dependent coordinates.
Evaluate the static balance for the problem in order to determine whether theorientation of the system in the gravitational field will effect the equations of
motion (Are the weights of the objects balanced by an initial static deflection in
the springs?). When in doubt, perform a static force balance to determine theappropriate constraint equation.
For displacement, velocity and acceleration terms, be sure to develop absoluteorrelativedisplacement, velocity and accelerations of appropriate points as required.
Watch out for 2-D and 3-D vector motions. Be sure to draw the appropriate free body diagrams for each mass (or combination
of masses) in the system.
o Whenever the system is separated in order to draw a free body diagram,replace the separation with the appropriate internal forces/moments (equal
and opposite forces/moments on each side of the separation).
o Do not move forces/moments arbitrarily from one mass to another. Theinternal forces account for the effects of one mass on another.
Develop one equation of motion for each degree of freedom of the system usingNewton's ( -or- d'Alemberts ) method. Be sure to watch for moving reference
frame issues. Also, check that the units are the same for each term in an equation
(Forces + Moments: NOT!) If necessary, once the exact equations of motion have been determined, linearize
the equations of motion by neglecting nonlinear terms in the equations of motion.Note that the linear equations of motion may not adequately describe the original
equations of motion if some of the terms that have been neglected are not
insignificant.
P
G
8/14/2019 Vibs1_review_notes.pdf
11/86
Mechanical Vibrations I
Lecture Notes -11- 06/16/06 12:25 PM
Newton Example A
In order to solve Newtons Equation,
iF mx= ,
for the given single degree-of-freedomsystem, it is first necessary to draw the
complete free-body diagram. This involves
identifying allinternal and external forcesacting upon the degree-of-freedom.
(Recall that the internal force resulting
from the spring is due to the actual relativemotion of the ends of the spring. Thereforethe coordinatexis measured from thefree lengthof the spring.)
All the internaland externalforcesare combined to yield the actual
equation of motion.
f kx mg cx mx =
collecting all the coordinate based
terms to one side of the equation
and all the externally applied forcesto the other yields
mx cx kx f mg + + =
Since the mgterm ends up on the right hand side with the external force, the coordinate
x can be redefined to be from thestaticequilibrium position. As a result, the mgterm
can be eliminated as shown.
Evaluating the static equilibrium equation (i.e. 0x= , 0x= & 0f = .) yields,
kx mg =
Therefore the static deflection ( ) can be evaluated as
mg
k =
The coordinate system reference can be shifted to the static equilibrium position bydefining
m
k c
x(t) f(t)
k x
m
cx
= m
mx
mg
f
8/14/2019 Vibs1_review_notes.pdf
12/86
Mechanical Vibrations I
Lecture Notes -12- 06/16/06 12:25 PM
sx x=
Since the static deflection ( ) is constant, obviously the velocity and acceleration isunchanged by the coordinate shift.
sx x= and s x=
Substituting into the original solution,
( )s s smx cx k x f mg + + + =
But recall that k mg = , therefore
s s smx cx kx f + + =
Note: Often the solution solved about the static equilibrium position will eliminate the
mg term when there are springs to carry the load. When this occurs, the mg term will
end up on the right hand side of the equation with the applied external forces.
8/14/2019 Vibs1_review_notes.pdf
13/86
Mechanical Vibrations I
Lecture Notes -13- 06/16/06 12:25 PM
Newton Example B
Again, in order to solve Newtons Equation,
0 cgJ L mL = + ,
for the given single degree-of-freedomsystem, it is first necessary to draw the
complete free-body diagram. This involves
identifying allinternal and external forcesacting upon the degree-of-freedom. In this
case, there is no specific external moment applied at point o. Also, recall the parallel
axis theorem, 20 cgJ J mL= + . (See above for the general expression.)
All the internaland externalmoments arecombined to form the actual equation of motion.
Notice that neither the support force T nor the
centripetal acceleration force 2mL contribute
to the moment about point o. (However Twould contribute to the moments about the cg.)
( )2sin cgmgL J mL = +
collecting all the coordinate based terms to one side of the equation and all the externally
applied moments to the other yields
( )2 sin 0cgJ mL mgL + + =
This time, since the mgterm ends up on the left hand side (multiplied by the coordinate),
it is not possible to eliminate the mgterm from the solution. So the exactequation of
motion remains,
( )2 sin 0cgJ mL mgL + + =
However, because the sinmgL term is non-linear in the desired coordinate, it ispossible to reduce the solution to a linearapproximation by using the first order
McClaurin Series expansion, sin for 1 , as follows.
( )2 0cgJ mL mgL + + =
,cgm J
L
=
cgJ
mg
T
mL2mL
8/14/2019 Vibs1_review_notes.pdf
14/86
Mechanical Vibrations I
Lecture Notes -14- 06/16/06 12:25 PM
#3 - The Eigenvalue Problem
In many areas of engineering, a mathematical concept referred to as the eigenvalueproblem arises. While the general application of eigenvalue methods has been used to
solve a wide variety of mathematical, scientific and engineering problems, in this course,we will focus upon its application to the area of mechanical vibrations. Because of this,
we will not examine the eigenvalue/eigenvector problem in detail, but only its most basiccharacteristics.
Mathematically, the eigenvalue problem can be expressed as:
[ ]{ } { }A x x=
In other words, given a matrix [ ]A , the object is to find a vector { } that when multiplied
times the matrix [ ]A does not change direction, but only magnitude (and possibly sense.)If this vector can be found, it is referred to as an eigenvector of [ ]A , and the scalar ,which describes the change in magnitude and sense, is referred to as an eigenvalue of
[ ]A . (Note that in general, there will be as many eigenvalue/eigenvector pairs as the sizeof the matrix. Also, it is only defined for square matrices.)
The solution is essentially a two-step process. First identify the eigenvalues of the matrix,
and then solve for their associated eigenvectors.
Solving for the eigenvalues involves manipulating the equation form as follows.
[ ]{ } { } { }0A x x =
[ ]{ } [ ]{ } { }0A x I x =
[ ] [ ] { } { }0A I x =
At this point, we recognize the form as a null space solution problem. There are three
ways in which the equation can be satisfied, two of them, [ ] [ ] [ ]0A I = and
{ } { }0x = , are trivial solutions and are rejected. The third is that the matrix [ ] [ ]A I
is singular (or rank deficient) for certain values of . These values of are theeigenvalues of the matrix [ ]A . To solve for the s, recognize that a singular (or rankdeficient) matrix has a determinant of zero.
[ ] [ ] 0A I =
8/14/2019 Vibs1_review_notes.pdf
15/86
Mechanical Vibrations I
Lecture Notes -15- 06/16/06 12:25 PM
Expanding this determinant results in a polynomial, known as the characteristic
polynomial, whose roots are the eigenvalues .
After identifying the eigenvalues, go back and solve the linear equations,
[ ] [ ] { } { }0A I x = , for each eigenvalue . However there is a problem, since the
coefficient matrix is singular (or rank deficient) when evaluated at the eigenvalues, wecannot simply solve the equation set. In general, the matrix will be rank deficient by one,
indicating that one variable is free, so the most straight-forward solution approach is to
simply set one of the vector { } elements equal to one and solve as normal. (We will not
get into the issues involved in the repeated root problem. In other words, where two or
more of the eigenvalues are equal.)
8/14/2019 Vibs1_review_notes.pdf
16/86
Mechanical Vibrations I
Lecture Notes -16- 06/16/06 12:25 PM
Eigenvalue/Eigenvector Example
Given the following matrix, find its associated eigenvalues and eigenvectors.
[ ] 4 12 5
A =
First identify the eigenvalues,
[ ] [ ]4 1 1 0 4 1
02 5 0 1 2 5
A I
= = =
4 1(4 )(5 ) ( 1)( 2) 0
2 5
= =
2 9 18 0 + =
Which has the roots, 1 3 = and 2 6 = .
Next identify the eigenvector associated with each eigenvalue.
[ ] [ ] { } { }0A I x =
1 1
2 2
4 1 1 0 4 1 0
2 5 0 1 2 5 0
x x
x x
= =
For the first eigenvalue, 1 3 = ,
11
21
4 1 0
2 5 0
x
x
=
1 1
2 2
4 3 1 1 1 0
2 5 3 2 2 0
x x
x x
= =
Notice the rank deficient nature of the resulting coefficient matrix.
Set one element of the vector { }x equal to one (unity) to solve. For this example, we will
use the first element. (Note that we could actually set it equal to any arbitrary valueexcept zero and still solve. Occasionally an eigenvector will contain a zero, but we cannot
arbitrarily pick that value. Finally, if after picking a value, the resulting equations cannot
be solved, simply try again picking a different element.)
Letting 1 1x = ,
8/14/2019 Vibs1_review_notes.pdf
17/86
Mechanical Vibrations I
Lecture Notes -17- 06/16/06 12:25 PM
2
11 1 0
2 2 0x
=
We can solve either equation for2
. Lets pick the first for simplicity.
[ ] 22
11 1 1 0x
x
= =
Which yields, 2 1x = . Therefore, the first eigenvalue/eigenvector pair is:
1 3 = , { }11
1x
=
Repeating the process for the second eigenvalue, 2 6 = , yields,
12
22
4 1 0
2 5 0
x
x
=
1 1
2 2
4 6 1 2 1 0
2 5 6 2 1 0
x x
x x
= =
Again letting 1 1x = , and solving for 2x , yields,
2
12 1 0
2 1 0x =
[ ] 22
12 1 2 0x
x
= =
Which yields, 2 2x = . Therefore, the second eigenvalue/eigenvector pair is:
2 6 = , { }21
2x
=
8/14/2019 Vibs1_review_notes.pdf
18/86
Mechanical Vibrations I
Lecture Notes -18- 06/16/06 12:25 PM
Final Notes
Looking at the form of the eigenvalue problem one more time, it can be shown that it canactually be used to solve certain first order differential equations directly, those of the
form 'y Ay= . Assuming that the vector { } is of the form { } { }( ) tx t X e= , then theeigenvalue problem can be written as.
[ ]{ } { }A x x=
[ ]{ } { }( ) ( )A x t x t=
[ ]{ } { }t tA X e X e =
[ ]{ } { }A X X=
Which is really no different than the example problem solved above.
There are a few issues, which will be covered later, applying this method to the second
order differential equations that arise in structural dynamics, specifically the need to use
state-space expansion to convert the set second order differential equations to a larger
system of first order differential equations.
Finally, when using MATLAB to solve eigenvalue problems, while the eigenvalues
should be the same, the eigenvector will in general be scaled differently. MATLAB usesunity vector length for its scaling.
8/14/2019 Vibs1_review_notes.pdf
19/86
Mechanical Vibrations I
Lecture Notes -19- 06/16/06 12:25 PM
#4 - Homogeneous Solution of the Equation of Motion
For the single degree offreedom (SDOF) system
shown, the equation of motionis:
( ) ( ) ( ) ( )mx t cx t k x t f t + + =
This differential equation canbe solved by any of several
methods, including: Laplace Transforms, Fourier Transforms, or by an assumed solution.
In this case, assuming a solution appropriate for a second-order, constant coefficientdifferential equation.
The general multi-degree of freedom problem can be formulated in matrix form as:
[ ]{ } [ ]{ } [ ]{ } { }( ) ( ) ( ) ( )x t C x t K x t f t+ + =
Using a solution appropriate for a linear, constant matrix coefficient, differential
equation, (namely complex exponentials), we have:
{ } { }( ) stx t X e= & { } { }( ) stf t F e= NOTE: thatX,Fandsare complex scalars.
By focusing on the homogeneous solution (i.e. 0f = ), we first evaluate the derivatives of
the assumed solution,{ } { }( ) stt X e= :
{ } { } { }( ) ( ) stddt x t x t X se= =
{ } { } { }2
2
2( ) ( ) stddt
x t x t X s e= =
By substituting the assumed velocity and acceleration into the original matrix expression
and collecting terms, we have: (NOTE: ste is non zero for all time and so may be
eliminated.)
[ ] [ ] [ ] { } { }2
0M s C s K X + + =
We should recognize this as a form of eigenvalue problem (or a null space solution
involvings.) Also, recognize that the solution { } { }0X = is the trivial solution and does
not contribute any useful information. Therefore, since the system matrix becomes rank
deficient for certain values ofs, the problem reduces to a polynomial root solver. (Thedeterminant of the system matrix is zero.)
m
k
c
x(t)f(t)
8/14/2019 Vibs1_review_notes.pdf
20/86
Mechanical Vibrations I
Lecture Notes -20- 06/16/06 12:25 PM
[ ] [ ] [ ]2 0M s C s K+ + =
This determinant results in a polynomial whose roots are the poles of the system. This isknown as the characteristic polynomial. Often the determinant notation is dropped (but
still written equal to thescalarzero implying a determinant) and the matrix characteristicpolynomial is expressed as,
[ ] [ ] [ ]2 0M s C s K+ + =
Returning to the scalar (1 DOF) problem as initially presented, we have:
2 0ms cs k + + =
Since the determinant of a scalar is simply the scalar itself, the characteristic polynomial
is simply:
2 0ms cs k + + =
Since this is a quadratic equation, there are two roots, namely:
2
1,2
4
2
c c mk
m
=
This results in the homogeneous (transient) solutions form of:
1 2
1 2( ) t t
x t Ae A e = +
The characteristic behavior of this system solution is determined by the relationship of m,
c, & k.
2 4c mk> Two distinct real roots.2 4c mk= Two equal real roots.
2 4c mk< One pair of complexconjugate roots.
For the condition of 2 4c mk> , no oscillation occurs. This is called over-damped. Sinceno oscillation occurs, this is not particularly interesting from a vibration point of view
and we will not investigate it any further.
For the condition of 2 4c mk= , this is the boundary between oscillatory and non-oscillatory behavior. This is called critically-damped. We will return to this later.
8/14/2019 Vibs1_review_notes.pdf
21/86
Mechanical Vibrations I
Lecture Notes -21- 06/16/06 12:25 PM
The interesting oscillatory behavior occurs when 2 4c mk< . This is called under-damped.When a system is under-damped, the system will respond to initial conditions with
harmonic oscillatory behavior.
First, we recognize the roots of the characteristic equation as the system poles ( r ). For
the given SDOF system, the roots have the following form:
2
1,22 2
r
c k cj
m m m
= =
Examining the two parts (real and imaginary) of the poles, we can write:
r r rj =
Where,
2rc
m
= is the damping for mode r(for single degree-of-freedom systems, it isoften written as d ) and
2
2r
k c
m m
=
is the dampednatural frequency for mode r(for single
degree-of-freedom system, it is often written as d )
Since 1 and 2 are complex conjugates, the homogeneous solution for the SDOF
problem can be written as:
**( ) t tx t Xe X e = +
If we let 0c , then rk
m . This is called the undampednatural frequency, written
as r (for single degree-of-freedom system, it is often written as n .) This is the
frequency at which the system will oscillate if there is no damping in the system.
Returning to the critically damped condition, where 2 4c mk= , we can now define a fewmore vibration parameters.
Defining the critical damping as:
2cc mk=
Allows us to define the damping ratioas:
c
c
c =
We can now express the damping and damped natural frequency as functions of the
undamped natural frequency and damping ratio, as follows:
8/14/2019 Vibs1_review_notes.pdf
22/86
Mechanical Vibrations I
Lecture Notes -22- 06/16/06 12:25 PM
2r r r
c
m = =
2
21
2
r r r
k c
m m
= =
The single degree of freedom equation can now be written as:
2 ( )( ) 2 ( ) ( )r r rf t
x t x t x tm
+ + =
While each of these parameters has been developed from a single degree of freedom (1
DOF) point of view, the concepts are equally applicable to the multi-degree of freedom
case, as well. For this, we need to use a nomenclature that reflects the fact that there is
more than one system pole (with its conjugate). We do this by using the subscript r(indicating resonance) with each of the modal parameters, as follows.
r r rj = + is the system pole for mode r, where r is the damping and r is the
damped natural frequency for mode r. From this, we define 2 2r r r = + as the
undampednatural frequency for mode r. Then, the damping ratio is simply defined as,
rr
r
=
. The general homogeneous solution then takes on the following form.
{ } { } { }*
*( ) trtr
r rr
t X e X e= +
The outgrowth of this approach is that it is no longer necessary to know the actual mass,stiffness and damping of the system in order to identify its modal characteristics. This is asignificant advantage when applied to experimental modal analysis because, as will be
shown later, the system poles ( r s) can be obtained directly from measured Frequency
Response Functions.
8/14/2019 Vibs1_review_notes.pdf
23/86
Mechanical Vibrations I
Lecture Notes -23- 06/16/06 12:25 PM
Homogeneous Example
Given a single degree-of-freedom mass/spring/damper system (MCK) with 5kgmass,
20N s m damping, and 1000Nm stiffness: identify the pole of the system ( r ), the
undamped natural frequency ( r ), and the damping ratio ( r ).
To calculate the system pole, evaluate the characteristic equation,
2 0ms cs k + + = 25 20 1000 0s s+ + =
which has the following roots. Note that the roots of the characteristic equation are the
system poles.
220 1000 20
2 5 5 2 5
N s N N sm m m
r jkg kg kg
=
( )2
2
22 200 2rad rad rad s sr sj =
2 14 radsr j =
The pole gives the damping ( r ) and the damped natural frequency ( r ) directly. Recall,
that to express the answer in units of Hertz, it is necessary to divide by 2 .
2rad
sr = -or-
1
0.318Hz Hz =
14 radsr = -or-7
2.228Hz Hz
=
For this problem, the undamped natural frequency ( r ) may be calculated from either
the magnitude of the system pole, or by evaluating the characteristic equation for 0c= .Both approaches yield the same solution.
2
22 14 200 14.14rad rad rad s sr r sj = = = = -or- 2.251Hz
221000 200 14.14
5
N
m rad rad sr skm kg = = = =
-or- 2.251Hz
Again, for this problem, the damping ratio may be calculated by either of twoapproaches, either directly from the system pole or from the systems mechanical
properties (MCK).
8/14/2019 Vibs1_review_notes.pdf
24/86
Mechanical Vibrations I
Lecture Notes -24- 06/16/06 12:25 PM
20.1414
14.14
radsr
rrad
sr
= = =
-or- 14.14%
200.1414
2 2 5 1000
N sm
rN
c m
c c
c mk kg
= = = =
-or- 14.14%
Recognize one important point in the solution method. The direct (MCK) solutions are
only valid for a single degree-of-freedom (SDOF) system while the solution methods
based upon the system pole are valid for a system with any number of degrees-of-freedom (MDOF).
8/14/2019 Vibs1_review_notes.pdf
25/86
Mechanical Vibrations I
Lecture Notes -25- 06/16/06 12:25 PM
#5 Estimating Modal Parameters Time Domain
There are many techniques, both simple and advanced, for estimating modal parameters
(frequency, damping, mode shape, and scaling) from measured experimental data. This
section will focus upon the single degree-of-freedom time domain technique known aslog decrement.
Log Decrement
The log decrement method is based entirely upon the characteristics of a single degree-
of-freedom free decay response.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15-4
-3
-2
-1
0
1
2
3
4
Time (sec)
Amplitude
Recalling the form of solution, ( ) rtx t Xe= , it is possible to estimate the system pole
( r r rj = + ) directly from the observed response.
Estimating the damped natural frequency ( r ) is easy. First estimate the period of one
oscillation ( ), the reciprocal of the period, 1
, is the damped natural frequency ( r ) in
Hz.
2 radsr
= -or- 1 Hz
Estimating the damping ( r ) requires a little more effort. The development proceeds as
follows.
Begin by evaluating the position, ( )t , for two time instants one cycle apart.
8/14/2019 Vibs1_review_notes.pdf
26/86
Mechanical Vibrations I
Lecture Notes -26- 06/16/06 12:25 PM
1 1 1
1( ) r r rt t j t x t Xe Xe e
= = 2 2 2
2( ) r r rt t j t x t Xe Xe e = =
Next evaluate the log decrement ( ) as:
1 1
2 2
1
2
( )ln ln
( )
r r
r r
t j t
t j t
x t Xe e
x t Xe e
= =
However, since the points are defined one cycle apart, 2 1t t = + . Substituting into the
above expression yields,
( ) ( )
1 1
1 1ln
r r
r r
t j t
t j t
Xe e
Xe e
+ +
=
Next, recall that cos sinje j = + . Hence,
( ) ( ) ( )11 11 12 2 2rr rr r rj tj t j tj t j t je e e e e e
++ += = = =
So the expression for the log decrement can be reduced to
( )( )( ) ( )
1
11
1
ln ln lnr
rr r
r
ttt
rt
ee e e
e
+
+
= = = =
Recalling that r r r = ,21r r r = and 2r
= . Substituting into the above
expression yields,
2 2
2 22
1 1
r r rr r r r r
rr r r
= = = = =
Therefore the log decrement equals,
2
2
1
r
r
=
If the damping ratio is small (eg.
0.1< -or- 10%), then the above
expression can be simplified to:
2 r
(Error less than 1% when 0.1< )
Multiple Cycle Solution
If time points multiple (n) periods apart
are used, the expression for the log
decrement is modified as follows:0
2
1 ( ) 2ln
( ) 1
r
n r
x t
n x t
= =
where: nis the number of cycles.
Which simplifies to:
01 ( )ln 2( )
r
n
x t
n x t
=
8/14/2019 Vibs1_review_notes.pdf
27/86
Mechanical Vibrations I
Lecture Notes -27- 06/16/06 12:25 PM
Log Decrement Example
Using the figure given above, estimate both the damped natural frequency and the
damping ratio.
First, select two convenient time points to estimate the period ( ). Between 0t= seconds and 9t= seconds, there are 11 cycles.
90.8182
11sec
cycle
sec
cycles= =
Therefore the damped natural frequency is
111.222
9r
cyclesHz
sec = = -or- 7.679 rads
Second, select two convenient time points to estimate the log decrement ( ). Between
1.8t seconds and 6.7t seconds, there are 6 cycles with ( 1.8) 3x t and
( 6.7) 2x t .
0 01 ( ) 1 ( 1.8) 1 3ln ln ln 0.0676( ) 6 ( 6.7) 6 2n n
x t x t
n x t x t
= = = =
Since the log decrement ( ) is so small, first try the simplified approximation for
damping ratio ( ).
0.0676 0.01082 2
= = -or- 1.08%
(Error check: 0.1 < so the approximation is adequate.)
Just for completeness, since 0.1< , 21 1 , therefore
27.679
1
r radsr r
r
= =
-or- 1.222Hz
0.0108 7.679 0.0829rad rad s sr r r = = = -or- 0.0132Hz
8/14/2019 Vibs1_review_notes.pdf
28/86
Mechanical Vibrations I
Lecture Notes -28- 06/16/06 12:25 PM
#6 - Lagranges Method
Lagranges Equation
i
i i i i
d T T U DF
dt q q q q
+ + =
Where:
T= Kinetic Energy U= Potential Energy D = Dissipative Energy iF= Externally applied force/moment
iq = Generalized coordinate
Generalized Coordinates
, , , , ,x y zq x y z =
Kinetic Energy
2 2
1 1
1 1
2 2
jm NN
i i
i i
T M x J
= =
= +
Potential Energy
( )21 1
1
2
k mN N
i i
i i
U K x M g x= =
= +
Dissipative Energy
2
1
1
2
cN
i
i
D C x=
=
8/14/2019 Vibs1_review_notes.pdf
29/86
Mechanical Vibrations I
Lecture Notes -29- 06/16/06 12:25 PM
Lagranges Method
Typical problems with Lagrange formulations:
Be sure to establish number of degrees of freedom first and formulate all energy
terms in only those variables. Clearly identify which degrees of freedom arerelative coordinates versus absolute coordinates. Watch out for
rotational/translational problems.
For kinetic energy terms, be sure to formulate absolutevelocities before takingderivatives. Watch out for 2-D and 3-D vector motions.
For potential energy terms, be sure that the actual deflection, described by relativeand/or absolute coordinates, in spring elements is described. Watch out for 2-D
and 3-D vector motions.
There should be only one total kinetic energy equation, one total potential energyand one total dissipative energy equation for the system. The kinetic, potential anddissipative energy equations should involve only the N generalized coordinates
and the constants (mass, damping, stiffness) of the system. Apply the Lagrange Equation once for each generalized coordinate. For N degrees
of freedom, N generalized coordinates will yield N equations of motion.
If necessary, linearize the equations of motion by neglecting nonlinear terms inthe equations of motion. Note that the linear equations of motion may not
adequately describe the original equations of motion.
NOTE: For the single degree-of-freedom (only one coordinate), conservative (no
damping), un-forced problem (no external forces), there is a simplified method called the
Energy Method.
( ) 0d
T Udt
+ =
8/14/2019 Vibs1_review_notes.pdf
30/86
Mechanical Vibrations I
Lecture Notes -30- 06/16/06 12:25 PM
Lagrange Example A
In order to solve the Lagrange Equation,
i
i i i i
d T T U D Fdt q q q q
+ + =
,
for the given single degree-of-freedom
system, it is first necessary to formulatethe energy equations: kinetic, potential and
dissipative.
First, formulating the kinetic energy yields, 212T mx=
Second, formulating the potential energy yields, 212U kx mgx= +
Finally, formulating the dissipative energy yields, 212
D cx=
Next, it is necessary to evaluate each of the Lagrange terms,
( ) ( )212d T d d
mx mx mxdt x dt x dt
= = =
( )212 0T
mxx x
= =
( )212U
kx mgx kx mg x x
= + = +
( )212D
cx cxx x
= =
Which are combined to form the actual equation of motion.
0mx kx mg cx f + + + =
mx cx kx f mg + + =
Since the mgterm ends up on the right hand side with the external force, the coordinatex can be redefined to be from thestaticequilibrium position. As a result, the mgterm is
eliminated as shown.
mx cx kx f + + =
m
k c
x(t) f(t)
8/14/2019 Vibs1_review_notes.pdf
31/86
Mechanical Vibrations I
Lecture Notes -31- 06/16/06 12:25 PM
Lagrange Example B
Again, in order to solve the LagrangeEquation,
i
i i i i
d T T U DF
dt q q q q
+ + =
,
for the given single degree-of-freedomsystem, it is necessary to formulate the energy
equations: kinetic, potential and dissipative.
First, formulating the kinetic energy yields, ( )2
21 12 2cgT J m L = +
Second, formulating the potential energy yields, ( )1 cosU mgL =
Finally, formulating the dissipative energy yields, 0D=
Next, it is necessary to evaluate each of the Lagrange terms,
( )( ) ( )22 2 21 12 2cg cg cg d T d d J m L J mL J mLdt dt dt
= + = + = +
( )( )221 12 2 0cgT J m L
= + =
( )( )1 cos sinU
mgL mgL
= =
( )0 0D
= =
Which are combined to form the actual equation of motion.
2 0 sin 0 0cgJ mL mgL + + + =
This time, since the mgterm ends up on the left hand side (multiplied by the coordinate),
it is not possible to eliminate the mgterm from the solution. So the exactequation of
motion remains,
( )2 sin 0cgJ mL mgL + + =
, cgm JL
8/14/2019 Vibs1_review_notes.pdf
32/86
Mechanical Vibrations I
Lecture Notes -32- 06/16/06 12:25 PM
However, because the sinmgL term is non-linear in the desired coordinate, it is
possible to reduce the solution to a linearapproximation by using sin as follows.
( )2 0cgJ mL mgL + + =
8/14/2019 Vibs1_review_notes.pdf
33/86
Mechanical Vibrations I
Lecture Notes -33- 06/16/06 12:25 PM
Energy Example A
In order to solve the Energy Equation,
( ) 0d T Udt
+ = ,
for the given single degree-of-freedom system, it
is first necessary to formulate the energy
equations: kinetic and potential.
Formulating the kinetic energy yields, 212
T mx= ,
and formulating the potential energy yields, 212U kx mgx= + .
Evaluating the Energy equation yields,
( ) ( )2 21 12 2 0d d
T U mx kx mgx mxx kxx mgxdt dt
+ = + + = + + =
Observe the oddform of the equation. There is an extra velocity term, x , that must be
canceled out to yield the actual equation of motion.
0mx kx mg + + =
mx kx mg + =
Since the mgterm ends up on the right hand side with the external force, the coordinate
x can be redefined to be from thestaticequilibrium position. As a result, the mgterm is
eliminated as shown.
0mx kx+ =
NOTE: Comparing the solution generated by Lagrange Method to that of the Energy
Method*shows that the Energy Method is not simply a single degree-of-freedom
Lagrange Solution, but rather a related, by not equivalent technique. Recognize also thatthe application of the Energy Method is extremelylimited, i.e. single degree-of-freedom,
conservative and un-forced. The Lagrange Method has no such limitations and is in fact acompletely general solution method, comparable in usefulness to Newtons Method.
*Note that the Energy Method is presented for completeness. You may encounter it in your studies of other
reference material, but because it is not a general technique, it will not be used in this course.
m
k
x(t)
8/14/2019 Vibs1_review_notes.pdf
34/86
Mechanical Vibrations I
Lecture Notes -34- 06/16/06 12:25 PM
#7 - Steady-State Solution of Equation of Motion
The general solution to the equation of motion involves both the homogeneous and the
particular solution. While the form of the homogeneous solution is dictated by the nature
of the governing second order constant matrix coefficient differential equation, the formof the particular solution is controlled by the form of the forcing function (excitation).
Mathematically, there are infinite possibilities, but for practical applications, the form can
be restricted to harmonic functions. By limiting the form of the excitation to harmonicfunctions, the concept of steady-state solution can be defined. For vibration purposes, the
steady-state solution is the resulting response after all the initial condition transients have
decayed. When solving the response for the steady-state, the solution is most easily
formulated in either the Laplace Domain (s) or the Frequency Domain (). (NOTE: it
will be shown later that the Frequency Domain solution can be developed from the
Laplace Domain by letting s j= .)
Working with a representative single
degree-of freedom system (SDOF), theequation of motion is:
( ) ( ) ( ) ( )mx t cx t k x t f t + + =
For a harmonic excitation, the forcing
function can be represented as:
( ) stf t Fe=
For the assumed solution, the response will
have the compatible form,
( ) stx t Xe= , ( ) stx t sXe= and 2( ) stt s Xe=
where bothF&Xare complex scalars.
Substituting, the above solution forms into the differential equation of motion produces a
solution that must be valid for every value ofs. This results in the following LaplaceDomain solution.
2 st st st st ms Xe csXe kXe Fe+ + =
By collecting common terms, the expression reduces to:
( )2 st st ms cs k Xe Fe+ + =
m
k c
x(t) f(t)
8/14/2019 Vibs1_review_notes.pdf
35/86
Mechanical Vibrations I
Lecture Notes -35- 06/16/06 12:25 PM
At this point the response/excitation ratio
can be calculated. This ratio is called theTransfer Function. It is normally written
as ( )H s .
21( ) ( )XH s s
F ms cs k=
+ +
While the Transfer Function is a
convenient mathematical model for
defining the input-output relationship,well suited to controls applications, it is
less appealing for vibrations purposes
primarily because it a continuousmathematical expression. However,
because the complex surface defined by
the ( )H s expression is analytic, any slicethrough the surface contains all the
information necessary to reconstruct theentire surface. This recognition leads to
the use of the Fourier Transform, that while its mathematical background is different, it is
effectively the Laplace Surface evaluated at s j= . (Another argument for the
equivalence of the Laplace Transform and the Fourier Transform lies in the fact that both
transforms start with the same Time Domain function, i.e. 1 1( ) ( ( )) ( ( ))g t L G s F G = = .)
Evaluating the Transfer Function at s j= yields another input-output form known as
the Frequency Response Function (FRF).
2
1( ) ( )
XH
F m j c k
=
+ +
By contrast with the Transfer Function,the Frequency Response Function is well
suited for vibration applications,
particularly experimental applications.While there is no discrete equivalent to the
Laplace Transform, there is a discrete
Fourier Transform that yields essentiallythe same information as the continuousintegral transform. This is particularly
important when working with discrete,
sampled time data.
By plotting both the Transfer Function and the Fourier Transform (over a limited range),
the equivalence of both transforms can be shown.
Multi-Degree of FreedomTransfer Function
The same basic development is
applicable to the multi-degree of freedom
system, as well.
Starting from the equation
[ ]{ } [ ]{ } [ ]{ } { }( ) ( ) ( ) ( )x t C x t K x t f t+ + =
and combining with
{ } { }( ) stt X e= & { } { }( ) stf t F e=
yields the matrix equation
[ ] [ ] [ ] { } { }2 st st s C s K X e F e + + =
By recognizing the basic definition of the
Transfer Function, the expression
becomes,
[ ] [ ] [ ]1
2( )H s M s C s K
= + +
Multi-Degree of FreedomFrequency Response Function
Just as the Transfer Function has a multi-degree of freedom form, so does the
Frequency Response Function.
By evaluating the multi-degree of
freedom (matrix) Transfer Function at
s j= , the multi-degree of freedom
Frequency Response Function results.
[ ] [ ] [ ]1
2( )H M j C K
= + +
8/14/2019 Vibs1_review_notes.pdf
36/86
Mechanical Vibrations I
Lecture Notes -36- 06/16/06 12:25 PM
One interesting aspect of the difference, however, is that while the Transfer Functionrequires the plotting of two surfaces, the Frequency Response Function requires only two
curves, clearly a much more convenient representation.
Note that in addition to displacement per unit force input-output relationships, other
steady-state input-output relationships can be developed. Some of these alternative
relationships will be explored later as steady-state applications.
-1
-0.5
0
0.5
1
-10
-5
0
5
10
0
0.2
0.4
0.6
0.8
1
Sigma [rad/sec]Omega [rad/sec]
Magnitude
-1
-0.5
0
0.5
1
-10
-5
0
5
10
-4
-2
0
2
4
Sigma [rad/sec]Omega [rad/sec]
Phase[rad]
-10 -8 -6 -4 -2 0 2 4 6 8 100
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Omega [rad/sec]
M
agnitude
-10 -8 -6 -4 -2 0 2 4 6 8 10-4
-3
-2
-1
0
1
2
3
4
Omega [rad/sec]
Phase[rad]
8/14/2019 Vibs1_review_notes.pdf
37/86
Mechanical Vibrations I
Lecture Notes -37- 06/16/06 12:25 PM
#8 - The Frequency Response Function (FRF)
In general, for most vibrations problems, the input/output relationship known as the
Frequency Response Function (FRF) will be used. This function describes the
input/output relationship on a frequency-by-frequency basis.
Although not generally referred to as a Frequency Response Function, the effect of agraphic equalizer on the sound from a stereo system can be described by thefrequency
response function. In this case, the input is the raw audio and the output is the modifiedaudio. By adjusting the amplification or attenuation (and relative phasing) of the various
bands, the equalizer can give the resulting audio a variety of sound characteristics, from
flat responseto enhancing male or female voices to concert hall sound effects.
Recall that the Frequency Response Function can be expressed by evaluating the Transfer
Function at s j= .
2
1( ) ( )
XH
F m j c k
=
+ +
By contrast with the Transfer Function,
the Frequency Response Function is well
suited for vibration applications,particularly experimental applications.
While there is no discrete equivalent to the
Laplace Transform, there is a discrete
Fourier Transform that yields essentiallythe same information as the continuous
integral transform. This is particularlyimportant when working with discrete,
sampled time data.
Input-Output Model
For the frequency response function, the following pictorial model is helpful. In fact,
when used for determining the steady state
response of a system, its application isextremely simple. It is merely a
multiplication in the frequency domain.
( ) ( ) ( )X H F =
Multi-Degree of FreedomFrequency Response Function
Just as the Transfer Function has a multi-degree of freedom form, so does the
Frequency Response Function.
By evaluating the multi-degree of
freedom (matrix) Transfer Function at
s j= , the multi-degree of freedom
Frequency Response Function results.
[ ] [ ] [ ]1
2( )H M j C K
= + +
( )H ( )F ( )X
8/14/2019 Vibs1_review_notes.pdf
38/86
Mechanical Vibrations I
Lecture Notes -38- 06/16/06 12:25 PM
Complex Exponential vs. Sine/Cosine Relationship
The relationship between the complex exponential form and the sine/cosine form of the
can be developed as follows:
Starting with the complex exponential form for the forcing function,
*( ) j t j t f t Fe F e = +
and recalling the Euler identity,
cos sinie i = +
the relationship is easily developed. (Remember that Fis complex.) Expanding the
expression for ( )f t ,
( ) ( )(cos sin ) ( )(cos sin )R I R If t F jF t j t F jF t j t = + + +
Collecting real and imaginary terms yields,
( ) ( cos sin ) ( cos sin ) ( cos sin ) ( cos sin )R I I R R I I Rf t F t F t j F t F t F t F t j F t F t = + + + +
Observing that the imaginary terms cancel and that the real terms are equivalent yields,
( ) 2 cos 2 sinR If t F t F t =
Comparing with a traditional differential equations solution of ( ) cos sint A t B t = + shows that 2 RA F= and 2 IB F= . Despite the use of complex exponentials in the
development, clearly the resulting time domain waveform is real.
FRF Characteristics
This figure presents a typical
single degree-of-freedomfrequency response function.
The FRF is present two sided
(both positive and negativefrequency) to clearly show that
the FRF for positive and
negative frequencies arecomplex conjugates.
*( ) ( )H H =
-20 -15 -10 -5 0 5 10 15 2010
-5
10-4
10-3
10-2
Frequency [Hz]
Magnitude[m/
N]
-90
0
90
180
m = 5 kg ; c = 20 Ns/m ; k = 1000 N/m
Phase[deg]
8/14/2019 Vibs1_review_notes.pdf
39/86
Mechanical Vibrations I
Lecture Notes -39- 06/16/06 12:25 PM
Because of this complex conjugate relationship, usually only the positive frequency part
is plotted.
Predicted Response*
It is important to remember that although the FRF is defined for allfrequencies, the
steady-state response (output) for a linear system will always be at thesamefrequency as
the excitation (input). This prediction comes by manipulating the definition of thefrequency response function, as shown.
( ) ( ) ( )X H F =
For a single degree-of-freedom system, the predicted steady-state response has thefollowing form.
2( )( ) FX
m j c k
=
+ +
Peak Response
Another important aspect of the frequency response function involves the frequency of
maximum response. It will be designated by p .
Starting with the single degree-of-freedom frequency response function,
2
1( )H
m j c k
=
+ +
it is possible to calculate the frequency of maximum (peak) response. Formulating the
magnitude of the frequency response yields,
( ) ( )2 22
1( )H
k m c
= +
In order to identify the frequency of maximum response, it is necessary to evaluate thederivative of the frequency response magnitude with respect to frequency as follows.
*TODO: Show example F(w), show resulting X(w). (FRF continuous, F&X discrete) Express F(w) =
CONST . Show units of FRF = 1/k, compliance. Work example. Show MDOF FRF for comparison.
8/14/2019 Vibs1_review_notes.pdf
40/86
Mechanical Vibrations I
Lecture Notes -40- 06/16/06 12:25 PM
( )( ) ( )
2 22
1( )
d dH
d dk m c
= +
( ) ( ) ( )( )1
2 2 22( )d d
H k m cd d
= +
( ) ( ) ( )( ) ( ) ( )( )3
2 22 222 21( )2
d dH k m c k m c
d d
= + +
( ) ( ) ( )( ) ( ) ( ) ( ) ( )3
2 2 22 2 21( ) 2 22
d d dH k m c k m k m c c
d d d
= + +
( ) ( ) ( )( ) ( )( ) ( )( )3
2 222 21( ) 2 2 2
2d H k m c k m m c c
d
= + +
In order to satisfy the maximum (or minimum), the derivative of the magnitude of the
frequency response function, ( )( ) 0pd
Hd
= , must equal zero. Since only the second
term in the above relation can be equal to zero (except at = ), the solution can bereduced to solving the following equation.
( )( ) ( )( )22 2 2 0p p pk m m c c + =
( )( )2 2 22 0p pkm m c + =
By dividing by the mass, 2m , the expression can be simplified by recognizing that
2 k
m = , 2
c
m = and 0p = cannot be the peak response for all possible MCK
combinations.
2
22 0pk c
m m
+ =
( ) ( )22 22 2 0p + =
2 2 2 22 0p =
( )2 2 21 2p =
8/14/2019 Vibs1_review_notes.pdf
41/86
Mechanical Vibrations I
Lecture Notes -41- 06/16/06 12:25 PM
Therefore, the frequency of maximum (peak) response is as follows.
21 2p =
For comparison, recognize that the damped natural frequencyis 21r r = .
Resonant Response
Starting with the definition of resonance, = , (i.e. the response of the system to anexcitation at the undampednatural frequency of the system.), the amplitude of the system
response can be calculated.
Starting with the single degree-of-freedom frequency response function,
2
1( )H
m j c k
=
+ +
from the definition of resonance, = , (and 2k m= ) the resonant response, ( )H ,
equals
1 1 1 1 1( )
2 22cH
kj c j c j kj km j
= = = = =
If the damping is small (i.e. 0.1< ), then the peak response can be approximated as
1( ) ( )
2pH H
j k
=
Half Power Method
After defining the resonant frequency, it is possible to estimate the damping, , from the
shape of the frequency response function in the vicinity of the resonance. Again,assuming that the damping is small (i.e. 0.1< ), the half-power bandwidth points can
be calculated and then damping can be estimated from the bandwidth and the resonance
location.
Starting with the definition of the half-
power bandwidth points,
L
H
p
( )2
pH ( )pH
8/14/2019 Vibs1_review_notes.pdf
42/86
Mechanical Vibrations I
Lecture Notes -42- 06/16/06 12:25 PM
1
2( ) ( ) ( )p L HH H H = =
and recognizing that for a lightly damped system, p , then expanding the single
degree-of-freedom frequency response yields,
2
1 1 1
22 j k k m j c =
+
Manipulating the expression yields a form from which the half-power bandwidth points
may be calculated.
2 2
2 2
1 1 1 1
221 1 2
cj j
k
= =
+ +
2
22 2 1 2j
= +
Squaring both sides yields
22 2 2 4 2
2 2 2
2 2 2 4 28 1 4 1 2 4
= + = + +
( ) ( )4 2
2 24 2 1 8 0
+ + =
Solving for
2
yields,
( ) ( ) ( )2
2 2 22 4 2 4 2 4 1 8
2
=
2
2 21 2 2 1
= +
Assuming is small (i.e. 1 ), then 2 0 , and the half-power bandwidth points are
approximated by,
2
1 2
8/14/2019 Vibs1_review_notes.pdf
43/86
Mechanical Vibrations I
Lecture Notes -43- 06/16/06 12:25 PM
Taking the difference between the points yields,
2 2
24H L H L H L
+
Using the approximation, 2H L + , allows reducing the expression to,
2H L
Therefore the damping factor can be approximated as,
2
H L
This approximation is also often written as2 c
f
f
or
2 c
BW
f .
While this method appears particularly simple, in practice, there are several difficulties
utilizing it effectively. Specifically, when using real sampled data, the ability to
determine the actual magnitude and location of ( )pH is limited by the measured data.
The actual sampled data will not in general include the exact frequency ( p ) and so both
the location and magnitude will be in error. Since the magnitude of ( )pH is used to
estimate the half-power bandwidth points, the location of L and H will be in error. Ingeneral, the errors tend to be the following: under estimating the magnitude of ( )pH ,
under estimating the magnitude of the half-power points, estimating L too low, and
estimating H too high. The compound effect of these errors is that generally the
damping factor ( ) is estimated too high.
8/14/2019 Vibs1_review_notes.pdf
44/86
Mechanical Vibrations I
Lecture Notes -44- 06/16/06 12:25 PM
FRF Example A
Given a single degree-of-freedom (SDOF) system with 5m kg= , 20N s mc = &
1000Nmk= , assuming a harmonic excitation of magnitude 20N@ 10Hz, determine the
response of the system to the given force.
Solution
First, identify the actual steady-state operating frequency.
10Hz= -or- 20 62.832rad rad s s =
Then, evaluate the SDOF frequency response function.
2 1( )H m j c k
= + +
( )2
1(20 )
5 20 20 20 1000
rads
rad rad N s N s s m m
Hkg j
= + +
( )1
(20 )18, 739 1, 257
rads
Nm
Hj
= +
( ) 6(20 ) 53.125 3.563 10rad ms NH j =
Finally, determine the response in the frequency domain.
( ) ( ) ( )X H F =
( ) 6(20 ) 53.125 3.563 10 20rad ms NX j N =
( ) 3(20 ) 1.063 0.071 10radsX j m = -or- 1.06 179.16 mm -or-
1.06 3.075rad mm
Therefore, the time domain response of the system to the given forcing function, ( )f t , is
expressed by*:
( ) 20cos(20 )f t t N=
( ) 1.06cos(20 3.075)t t mm=
*TODO: Need to show equivalence of two sided FRF solution.
8/14/2019 Vibs1_review_notes.pdf
45/86
Mechanical Vibrations I
Lecture Notes -45- 06/16/06 12:25 PM
#9 - Rotating Unbalance
For many situations, the source of excitation is an unbalance in a rotating machine part.
This internal (self) excitation is a function of the operation of the mechanism. However,
the techniques developed are applicable to this condition, as well.
By defining the position of the unbalance mass ( em ) as ( ) ( )t e t+ and the mass of the
system moving in translation only as em m , we
can write the equation of motion for this
representative single degree-of-freedom systemas:
( ( ) ( )) ( ) ( ) ( ) ( ) 0e em x t e t m m x t cx t k x t + + + + =
Rearranging terms yields:
( ) ( ) ( ) ( )emx t cx t k x t m e t + + =
Comparing this solution with the previously
developed equation of motion shows that the expression ( )em e t is equivalent in form to
the original external forcing function ( )f t . As such we can use the same harmonic
function assumed solution form to convert this to a steady-state solution.
By using both,
( ) st
x t Xe= , ( ) st
x t sXe= &2
( ) st
t s Xe= and
( ) ste t Ee= , ( ) ste t sEe= & 2( ) ste t s Ee=
where bothX&Eare complex scalars.
Substituting, the above solution forms into the differential equation of motion produces asolution that must be valid for every value ofs. This results in the following Laplace
Domain solution.
2 2st st st st
ems Xe csXe kXe m s Ee+ + =
By collecting common terms, the expression reduces to:
( )2 2st stems cs k Xe m Es e+ + =
Again, comparison with the Laplace Domain steady-state solution shows that2
eF m Es . Calculation of the effective Transfer Function yields:
m
k c
x(t)mee
8/14/2019 Vibs1_review_notes.pdf
46/86
Mechanical Vibrations I
Lecture Notes -46- 06/16/06 12:25 PM
2
2( ) ( )
e
X sH s s
m E ms cs k
= =
+ +
Or expressed in the Frequency Domain as:
2
2( ) ( )
e
XH
m E m j c k
= =
+ +
In the Frequency Domain, the effective forcing function is 2eF m E . As can be seen
from the form, the magnitude of the force is linearly proportional to both the unbalance
mass and eccentricity, however, it is proportional to the square of the rotational frequency
. Because of this, the unbalance is normally expressed in mass times length units ( em e )
and hence the above frequency response form (displacement / unbalance).
-1
-0.5
0
0.5
1
-10
-5
0
5
10
0
2
4
6
8
10
Sigma [rad/sec]Omega [rad/sec]
Magnitude
-1
-0.5
0
0.5
1
-10
-5
0
5
10
-4
-2
0
2
4
Sigma [rad/sec]Omega [rad/sec]
Phase[rad]
-10 -8 -6 -4 -2 0 2 4 6 8 100
1
2
3
4
5
6
7
8
9
O me a r ad /s ec
M
agnitude
-10 -8 -6 -4 -2 0 2 4 6 8 10-4
-3
-2
-1
0
1
2
3
4
Omega [rad/sec]
Phase[rad]
8/14/2019 Vibs1_review_notes.pdf
47/86
Mechanical Vibrations I
Lecture Notes -47- 06/16/06 12:25 PM
Rotating Example A
A machine of 100 kg mass has a 20 kg rotor with 5 mm eccentricity. The mounting
springs have 85 kNmk= and the damping is negligible. The operating speed is 600 rpm
and the unit is constrained to move vertically. (a)Determine the dynamic amplitude ofthe machine. (b)Redesign the mounting so that the dynamic amplitude is reduced to one
half of the original value, but maintaining the same natural frequency.
Solution
Given the following parameters:
100m kg= , 20em kg= , 5e mm= , 85 kNmk= ,
0N sm
c & 600rpm=
(a) Starting with the frequency response function
previously developed,
2
2( ) ( )
e
XH
m e m j c k
= =
+ +
Solve for the dynamic amplitude X by rearranging the expression and evaluating its
magnitude,
2
2( ) e
m eX
m j c k
=
+ +
The additional parameters needed are the unbalance and the rotational frequency in rads .
320 5 10 0.1em e kg m kg m= =
min
6002 20 62.83
60rad rad
s ss
rpm = = =
Evaluating the dynamic amplitude with 0c= yields,
2 2
2 2
(0.1 )(20 )
85,000 100 (20 )
radse
N radm s
m e kg mX
k m kg
= =
2
2
394.780.001274
309,784
kg m
s
kg
s
X m
= =
or 1.274mm
m
k c
x(t)mee
8/14/2019 Vibs1_review_notes.pdf
48/86
Mechanical Vibrations I
Lecture Notes -48- 06/16/06 12:25 PM
(b)There are two options for reducing the dynamic amplitude by half and yet keeping thesame natural frequency, increasing mass and increasing damping.
12
0.6372reduced original
X X mm= =
85,00029.15
100
Nm rad
s
k
m kg = = =
Option 1: increasing mass.
Manipulating the expression for dynamic amplitude
2 2 2
2 2 2 2( ) ( )
e e e
km
m e m e m eX
k m m m
= = =
yields an expression for the increased mass.
2 2
2 2 2 2
(0.1 )(20 )200
( ) 0.6372 ((29.15 ) (20 ) )
radse
rad rad s s
m e kg mm kg
X mm
= = =
However, keeping the same implies a corresponding increase in k. Therefore,
2 2200 (29.15 ) 170,000rad N s mk m kg = = = or 170 kNm
Notice how the solution makes sense intuitively. The operating speed is above the natural
frequency of the system. In that region, the frequency response function magnitude
(e
X
m e) approaches
1
m, so it would be expected that to reduce the dynamic amplitude by
half would require approximately doubling the mass.
Option 2: increasing damping.
Again, manipulating the expression for dynamic amplitude
2
2
em e
X m j c k
= + +
2 2
eX m j c k m e + + =
2
2 em e
m j c k X
+ + =
8/14/2019 Vibs1_review_notes.pdf
49/86
Mechanical Vibrations I
Lecture Notes -49- 06/16/06 12:25 PM
( )2
22
2 2
2( )
em ek m c
X
+ =
( )2
22
2 2
2( )
em ec k m
X
=
yields an expression for the necessary damping
( )2
22
2
2
2
em ek m
Xc
=
( )( )
22
22
2
2
0.1 (20 )85,000 100 (20 )
(0.006372 )
(20 )
rads
N radm s
rads
kg mkg
m
c
=
( )2 2
42 2
24 4
2 2
2 2
32
9 99 2
155.85 10309,784 383.87 10 95.966 10406.0 10 8,540
3,947.8 3,947.8
kg m
kgskg kg s
kgs ss
rad rad s s
mc
= = =
or more familiarly, 8,540N s mc = . This can be expressed as a fraction of critical damping
(c
c
c = ) by recalling the expression for the critical damping value, 2cc km= .
8,5401.465
2 2 85,000 100
N sm
Nm
c
km kg
= = =
or 146.5%
Notice that to reduce the dynamic response of this system utilizing only damping requires
an over-damped condition of nearly 150%.
For this problem, the mass addition is most likely to be the easiest and cheapest practical
solution.
8/14/2019 Vibs1_review_notes.pdf
50/86
Mechanical Vibrations I
Lecture Notes -50- 06/16/06 12:25 PM
#10 - Steady-State Applications
In this section, two different but related steady-state applications will be developed. The
first is the force transmissibility relationship. For transmissibility, the interest is in
determining the ratio of the force transmitted to the base ( ( )TF ) to the applied force
( ( )F ) as a function of frequency. The second application is vibration isolation. For
isolation, the interest is in determining the ratio of the response of the system ( ( )X ) to
the motion of the base ( ( )Y ) as a function of frequency. (This is sometimes referred to
as response ratio.) Just as in the earlier development of the Frequency Response Function(FRF), for both applications, the focus will be upon harmonic excitation.
Transmissibility
For transmissibility, we want to determine the ( )TFF
relationship. We begin with the
equation of motion for our representative single degree-of -freedom system. (NOTE: we
assume negligible motion of the base.)
( ) ( ) ( ) ( )mx t cx t k x t f t + + =
In addition, we need the restraining force at
the other end of the spring/damper.
( ) ( ) ( )T
cx t k x t f t + =
Applying the same transformationsdeveloped earlier, namely that the forcing
functions and responses can be express as
complex exponentials, and collecting terms,we get:
( )2 st st ms cs k Xe Fe+ + = and
( ) st st Tcs k Xe F e+ =
Finally, by taking the ratio of the second equation to the first, we get the desiredtransmissibility relationship.
2( )T
F cs ks
F ms cs k
+=
+ +
m
k c
x(t) f(t)
fT(t)
8/14/2019 Vibs1_review_notes.pdf
51/86
Mechanical Vibrations I
Lecture Notes -51- 06/16/06 12:25 PM
Or expressed in the Frequency Domain as:
2( )T
F j c k
F m j c k
+=
+ +
Isolation
For isolation, we want to determine the ( )X
Y relationship. We begin with the equation
of motion for our representative single degree-of freedom system. (NOTE: this time thebase motion is NOT negligible, however there is no explicitly applied external force.)
Because there is motion for both the structure
and the base, we need to use the difference inmotion across the spring/damper.
( ) ( )( ) ( ) ( ) ( ) ( ) 0mx t c x t y t k x t y t + + =
Transforming the equation as before, we get:
( ) ( )2 0st st st st stms Xe c sXe sYe k Xe Ye+ + =
Collecting terms, and moving the base motion to
the other side of the equation yields,
2
( ) ( )
st st
ms cs k Xe cs k Ye+ + = +
Finally, by cross dividing the relevant terms, we get the desired isolation relationship.
2( )
X cs ks
Y ms cs k
+=
+ +
Or expressed in the Frequency Domain as:
2( )
X j c k
Y m j c k
+=
+ +
m
k c
x(t)
y(t)
8/14/2019 Vibs1_review_notes.pdf
52/86
Mechanical Vibrations I
Lecture Notes -52- 06/16/06 12:25 PM
Conclusion
Comparing the expression for transmissibility to the expression for isolation, we find thatthey are identical. This interesting result implies that the process of determining the force
transmitted to the base is equivalent to determining the base motion transmitted to the
system.
-1
-0.5
0
0.5
1
-10
-5
0
5
10
0
2
4
6
8
10
Sigma [rad/sec]Omega [rad/sec]
Magnitude
-1
-0.50
0.5
1
-10
-5
0
5
10
-4
-2
0
2
4
Sigma [rad/sec]Omega [rad/sec]
Phase[rad]
-10 -8 -6 -4 -2 0 2 4 6 8 100
1
2
3
4
5
6
7
8
9
10
Omega [rad/sec]
Magnitude
-10 -8 -6 -4 -2 0 2 4 6 8 10-4
-3
-2
-1
0
1
2
3
4
Omega [rad/sec]
Phase[rad]
8/14/2019 Vibs1_review_notes.pdf
53/86
Mechanical Vibrations I
Lecture Notes -53- 06/16/06 12:25 PM
Transmissibility Example A
A vertical single-cylinder diesel engine of 500 kg mass is mounted on springs with
200 kNmk= and dampers with 0.2 = . The rotating parts are well balanced. The mass of
the equivalent reciprocating parts is 10 kg and the stroke is 200 mm. Find the dynamicamplitude of the vertical motion, the transmissibility, and the force transmitted to the
foundation, if the engine is operated at (a)200 rpm; (b)600 rpm.
Solution
Given the following parameters:
500m kg= , 10Rm kg= , 200stroke mm= ,
200 kNmk= & 0.2 =
(a)200 rpm
Starting with the single degree-of-freedom
frequency response function previously developed,
2
1( ) ( )
XH
F m j c k
= =
+ +
Solve for the dynamic amplitude X by rearranging the expression and evaluating its
magnitude,
2
( )( )
FX
m j c k
=
+ +
The additional parameters needed are the equivalent force, the damping and the operating
frequency expressed in rads . From the mechanical unbalance development, the equivalent
force can be expressed as 2eq eF m e= . Recognizing that the reciprocating mass is the
effective mass, e Rm m= , and that the effective eccentricity is half the stroke,
2strokee= , yields,
min
200 202 20.94
60 3rad rad
s ss
rpm = = =
2 210 0.1 (20.94 ) 438.7radseq eF m e kg m N= = =
2 2 0.2 200,000 500 4000N N sm mc km kg = = =
m
k c
x(t)
stroke
8/14/2019 Vibs1_review_notes.pdf
54/86
Mechanical Vibrations I
Lecture Notes -54- 06/16/06 12:25 PM
Evaluating the dynamic amplitude yields,
2 2
438.7
200,000 500 (20.94 ) (20.94 )(4000 )
eq
N rad rad N sm s s m
F NX
k m j c kg j = =
+ +
( ) 2438.7
0.005119, 324 83, 776 kg
s
NX m
j= =
+or 5.1mm
Evaluating the transmissibility yields,
2 2
200,000 (20.94 )(4000 )
200,000 500 (20.94 ) (20.94 )(4000 )
N rad N sm s m
N rad rad N sm s s m
k j c jTR
k m j c kg j
+ += =
+ +
( )( )
200,000 83,7762.52
19, 324 83, 776
Nm
Nm
jTR
j
+= =
+
From the definition of transmissibility, namely TF
TRF
= , the force transmitted to the
foundation can be determined.
2.52 438.7 1106T eqF TR F TR F N N= = = =
Notice that due to the characteristics of the system, the force transmitted to the
foundation is approximately 2.5 times the force applied to the system.
(b)600 rpm
Again solving for the dynamic amplitude Xusing the rearranged expression andevaluating its magnitude,
2
( )( )
FX
m j c k
=
+ +
The additional parameters needed are the equivalent force, the damping and the operating
frequency expressed in rads . From the mechanical unbalance development, the equivalent
force can be expressed as 2eq eF m e= . Recognizing that the reciprocating mass is the
effective mass, e Rm m= , and that the effective eccentricity is half the stroke,
2strokee= , yields,
min
6002 20 62.83
60rad rad
s ss
rpm = = =
8/14/2019 Vibs1_review_notes.pdf
55/86
Mechanical Vibrations I
Lecture Notes -55- 06/16/06 12:25 PM
2 210 0.1 (62.83 ) 3,948radseq eF m e kg m N= = =
2 2 0.2 200,000 500 4000N N sm mc km kg = = =
Evaluating the dynamic amplitude yields,
2 2
3,948
200,000 500 (62.83 ) (62.83 )(4000 )
eq
N rad rad N sm s s m
F NX
k m j c kg j = =
+ +
( ) 26 33,948
0.00221.774 10 251.3 10 kg
s
NX m
j= =
+ or 2.2mm
Evaluating the transmissibility yields,
2 2
200,000 (20.94 )(4000 )
200,000 500 (62.83 ) (62.83 )(4000 )
N rad N sm s m
N rad rad N s
m s s m
k j c jTR
k m j c kg j
+ += =
+ +
( )( )
3
6 3
200,000 251.3 100.179
1.774 10 251.3 10
Nm
Nm
jTR
j
+ = =
+
From the definition of transmissibility, namely TF
TRF
= , the force transmitted to the
foundation can be determined.
0.179 3,948 708T eqF TR F TR F N N= = = =
Notice that although the equivalent applied force is 900% larger than for the first case,
the force actually transmitted to the foundation is about 30% smaller.
8/14/2019 Vibs1_review_notes.pdf
56/86
Mechanical Vibrations I
Lecture Notes -56- 06/16/06 12:25 PM
#11 - Multiple Degree of Freedom Systems
MDOF Example
Starting with the simple two degree-of-freedom
system shown, the objective will be to determine
the equations of motion and the naturalfrequency and mode shapes.
The equations of motion can be found by eitherof two techniques, Newtons Method or
Lagranges Method.
Equations of Motion NewtonsMethod
As has been shown, Newtons Method involves
identifying all forces, both internal and external,acting on each body and then expressing those forces in terms of the given coordinates.
Therefore, the first action must be the development of thefree body diagramsfor each
body.
The given two degree-of-freedom
system has the two free body
diagrams, show at right.
Evaluating Newtons equation,
i
i
f ma=
, for each body yields a set
of equations describing the motion ofthe system.
( )1 1 1 2 1 2 1 1 1f k x k x x m g m x + =
( )2 2 2 1 2 2 2f k x x m g m x + =
Collecting and rearranging the terms
into standard form yields,
( )1 1 1 2 1 2 2 1 1m x k k x k x f m g + + = +
2 2 2 2 2 1 2 2m x k x k x f m g + = +
1( )x t 1( )f t
2( )x t 2( )f t
1k
2k
1m
2m
( )2 2 1k x x
2m
1 1k x
1m
( )2 1 2k x x
=1m
1 1m x
=2m
2 2m x
body #1
body #2
2m g
1m g
2f
1f
8/14/2019 Vibs1_review_notes.pdf
57/86
Mechanical Vibrations I
Lecture Notes -57- 06/16/06 12:25 PM
Or expressed in matrix form as,
1 1 1 2 2 1 1 1
2 2 2 2 2 2 2
0
0
m x k k k x f m g
m x k k x f m g
+ + = +
Recall that when the gravitational term ( im g) ends up on the force side of the equation,
the associated coordinate ( ix ) can be redefined from the static equilibrium position
thereby eliminating the im gterm from the equation. In this case, both terms are
eliminated.
1 1 1 2 2 1 1
2 2 2 2 2 2
0
0
m x k k k x f
m x k k x f
+ + =
Equations of Motion Lagranges Method
As has been shown, Lagranges Method involves identifying the kinetic,potentialand
dissipativeenergies for the entire system, then evaluating the Lagrange Equation for each
degree of freedom.
i
i i i i
d T T U DF
dt q q q q
+ + =
First, identifying the kinetic energy term,
2 21 11 1 2 22 2
T m x m x= +
And the potential energy term,
( )221 1
1 1 2 2 1 1 1 2 22 2U k x k x x m gx m gx= +
Provides all the information necessary to determine the equations of motion. Note that
there are no dissipative terms for this problem, so 0D= .
Second, evaluating the various partial derivatives for each degree of freedom, yields,
(First degree of freedom)
( ) ( )2 21 11 1 2 2 1 1 1 12 21 1
d T d d m x m x m x m x
dt x dt x dt
= + = =
8/14/2019 Vibs1_review_notes.pdf
58/86
Mechanical Vibrations I
Lecture Notes -58- 06/16/06 12:25 PM
( )( ) ( )( )221 11 1 2 2 1 1 1 2 2 1 1 2 2 1 12 21 1
1U
k x k x x m gx m gx k x k x x m g x x
= + = +
(Second degree of freedom)
( ) ( )2 2
1 11 1 2 2 2 2 2 22 2
2 2
d T d d m x m x m x m xdt x dt x dt
= + = =
( )( ) ( )221 11 1 2 2 1 1 1 2 2 2 2 1 22 22 2
Uk x k x x m gx m gx k x x m g
x x
= + =
Finally, collecting the various terms for each degree of freedom yields the equations of
motion.
( )1 1 1 2 1 2 2 1 1m x k k x k x f m g + + = +
2 2 2 2 2 1 2 2m x k x k x f m g + = +
Or expressed in matrix form as,
1 1 1 2 2 1 1 1
2 2 2 2 2 2 2
0
0
m x k k k x f m g
m x k k x f m g
+ + = +
Again recall, when the gravitational term ( im g) ends up on the force side of the equation,
then the associated coordinate ( ix ) can be redefined from the static equilibrium position
thereby eliminating the im gterm from the equation. In this case, both terms are
eliminated.
1 1 1 2 2 1 1
2 2 2 2 2 2
0
0
m x k k k x f
m x k k x f
+ + =
Notice that the actual equations of motion identified are identical, regardless of whether
Newtons Method or Lagranges Method was used.
Natural Frequency and Mode Shape
In order to identify the natural frequency and mode shape for a set of simultaneous
differential equations of motion, it is necessary to solve an eigenvalue problem. By
transformi