04/12/2023

Structural Dynamics

1
Structural Dynamics:
When there are change of inertia and the forces
acting on the structure with respect to time, then
we call it a Dynamic problem.

Degrees of Freedom:
The number of independent displacement
coordinates required to define the displaced
positions of all the masses relative to their original
position is called the number of Degrees of
Freedom of the system.
 Simple pendulum
O
y

Degree of freedom:
Coordinates x and y are related Equation of motion:
as the string length L is fixed. So
it has Single degree of freedom Taking moments about point O,
x y L
2 2 2

I  mgL sin   0

 f ( x, y , t )  0
When the constraint equation can
be expressed in this form, then
the system is called Holonomic
system.
 SDOF system
x(t ) x(t )
m f (t ) k f (t )

m
c

Assumptions
• Beam is rigid flexurally and having no ration but lateral
deformation only.
• Columns are massless.
• Mass of the system is ‘m’, stiffness is ‘k’, and damping coefficient is
‘c’, which is defined as the force required to induce unit velocity in
the system.
 SDOF system
x(t ) x(t )
m f (t ) k f (t )

m
c

D’ Alembert’s Principle:
Sum of the difference between forces acting on a system of particles
and the time derivative of the momentum projected onto a virtual
displacement consistent with the constraint is zero.

Which can be expressed as:

d
i ( F  dt mi vi ) ri  0
 SDOF system
x(t ) x(t )
m f (t ) k f (t )

m
c
Dynamic equilibrium equation:
Applying D’ Alembert’s principle on the system:

( F  kx  cx dtd mx ) ri  0
As,  ri cannot be zero,
 ( F  kx  cx dtd mx )  0
 kx  cx  F (t )
 mx
 Free Body Diagram Approach
x(t )
k f (t )
mx m
m kx f (t )
cx
c

Dynamic equilibrium equation:

Taking summation of all forces in x direction, we get
 kx  cx  F (t )
mx
For Undamped vibration the dynamic equilibrium equation
will be:
  kx  F (t )
mx
It is a second order linear differential equation of first degree.
 04/12/2023

Solution of homogenous DE
If a differential equation has the form,
x  ax  bx  0

 x  e t will be a trial solution.
 Auxiliary Eqn. will be
 2  a  b  0
1
or ,    a  a 2  4b 
2 
=1 , 2
Case 1, when 1 , 2 are real and distinct,
 x(t)=c1e1t  c 2 e2t
Case 2, when 1 , 2 are real and equal,
 x(t)=(c1  c 2t )e t
Case 3, when 1 , 2 are complex and in the form a  ib
 x(t)  e at [ A cos bt  B sin bt ]

8
 Undamped Free Vibration
x(t )
Dynamic equilibrium equation: k
  kx  0
mx m
c
k
Let , n  2

m
x  n 2 x  0
 
 Auxiliary Eqn. will be
  n  0
2 2

or ,   in
 x(t )  A cos nt  B sin nt
 Undamped Free Vibration
Imposing initial conditions
x(t )
x(0)  x0 ; k
x(0)  x0 ;
m
A  x0 ;
x 0 c
and B 
n
x 0
 x (t )  x0 cos n t  sin n t ;
n
x 0
Let, x0  C cos  ; and  C sin 
n
 x (t )  C cos(n t   )
x 0
where, C  x0  (
2
) 2  Amplitude of motion
n
x 0
  tan 1 ( )  Phase Angle
x0n
 04/12/2023

Undamped Free Vibration Response

Without damping the structure will keep vibrating indefinitely.

ωn is called natural frequency of the system, it is the frequency with
which a structure vibrates in free vibration.

11
 04/12/2023

Damped Free Vibration

Dynamic equilibrium equation:
k
 cx  kx  0
mx Let , n 2 
m
c
x  x  n 2 x  0
 
m
x(t )  Auxiliary Eqn. will be
k
c
m     n 2  0
2

m
c c c
 ( ) 2  4n 2
or ,   m m
2
c  (c) 2  4mk

2m
12
 04/12/2023

Damped Free Vibration

Case 1.
When, (c) 2  4mk  0
c  4mk
or , c  2mn  ccr

Ccr is called critical damping of the system.

ccr
 1  2 
2m
 Solution of the differential equation will be
ccr
( )t
x(t) =(c1  c2t )e 2m
 (c1  c2t )e nt
This type of system is called critically damped system.

13
 04/12/2023

Critically Damped systems

Imposing initial conditions
x(0)  x0 ;
x(0)  x0 ;
c1  x0 ;
and c2  x 0  n x0
 x (t )  e n t [ x0 (1  n t )  x 0t ]

14
 04/12/2023

Damped Free Vibration

Case 2. When, (c) 2  4mk >0
c  (c) 2  4mk

2m
c  (c) 2  4m 2 2 n

2m
c  2mn  2  1 c
 [ Let,   ]
2m ccr
c  2md/
 [ Let, d/  n  2  1]
2m
=  n  d/
This type of system is called overdamped system. Where  >1

15
 04/12/2023

Overdamped systems
 Solution of the differential equation will be
n t d/ t d/ t
x(t) =e [c1e  c 2e ]

Imposing initial conditions

x(0)  x0 ;
x(0)  x0 ;

x 0  (n  d/ ) x0
c1 
2d/
 x 0  ( n  d/ ) x0
and , c2 
2d/

16
 04/12/2023

Damped Free Vibration

Case 3. When, (c) 2  4mk <0

c  1(4mk  c 2 )

2m
c  i 2mn 1   2 c
 [ Let,   ]
2m ccr
c  i 2md
 [ Let, d  n 1   2 ]
2m
=  n  id

ωd is called damped natural frequency of the system.

This type of system is called underdamped system. Where  <1

17
 04/12/2023

Underdamped systems
 Solution of the differential equation will be
x(t) =e nt [ A cos d t  B sin d t ]

Imposing initial conditions

x(0)  x0 ;
x(0)  x0 ;

A  x0 ;
x 0  n x0
B
d

18
 04/12/2023

Logarithmic decrement
x(t ) Let,
k x1 (t1 )  e nt1 [ A cos d t1  B sin d t1 ]
m
and , x2 (t2 )  e nt2 [ A cos d t2  B sin d t 2 ]
c if, t2  t1  Td
x1
  enTd
x2
2
Td  and d  n 1   2
d
fo ace

x1 2
Pl

 
h

 ln
gr r
ap

x2 1  2

 is called logarithmic decrement.

For very low damping,   2
19
 04/12/2023

Forced vibration
x(t )
k f (t ) Equation of motion,
m  kx  cx  F (t )
mx
c Let , F (t )  P sin t

C.F  xc (t )  e nt [ A cos d t  B sin d t ]

If the system is excited by a sinusoid, the response will also be a
sinusoid, so for finding P.I let the trial solution be,
x p (t )  C1 cos t  C2 sin t
 x p (t )  C1 sin  t  C2  cos  t
x p (t )   2 (C1 cos t  C2 sin t )
 

20
 04/12/2023

Forced vibration
Putting in the equation of motion,
P
x  2n x  n 2 x 
 sin  t
m
We get,
[-C1 2 +C 2  .2n +C1n 2 ]cos t
P
+ [-C 2  -C1 .2n +C 2n  ]sin t=0
2 2

m
[-C1 2 +C 2  .2n +C1n 2 ]  0.......................(1)
P
[-C 2  2 -C1 .2n +C 2n 2  ]=0..................(2)
m

Let, Frequency ratio, r 
n

21
 04/12/2023

We get,
C1 (1  r 2 )  C2 (2 r )  0...............(i)
P
C1 (2 r )  C2 (1  r 2 )  .....(ii )
mn 2

P (2 r )
 C1 
k (1  r 2 ) 2  (2 r ) 2
P (1  r 2 )
 C2 
k (1  r 2 ) 2  (2 r ) 2
 x p (t )  C1 cos t  C2 sin t
P 1
 [(1  r 2
) sin  t  (2 r ) cos  t ]
k (1  r )  (2 r )
2 2 2

P 1
 sin( t   )
k (1  r )  (2 r )
2 2 2

1
=xst sin(t   )
(1  r )  (2 r )
2 2 2

22
 04/12/2023

Dynamic amplification factor

DMF:
DMF is the ratio of the dynamic response to static response of a
structure, and is given by

1
D
(1  r 2 ) 2  (2 r ) 2

fo ace
Pl

h
gr r
ap
Dynamic Magnification Factor (DMF)

23
 04/12/2023

Phase angle

fo ace
Pl

h
gr r
ap

24
 04/12/2023

Force transmitted to support

x(t )
k f (t )

m
FT  kx  cx c
 k .xst .D.sin(t   )  c.xst .D..cos(t   )
 kP sin( t   )  c..P.cos( t   )
 P k 2  (c ) 2 sin( t     )
c
 tan    2 r
k

25
 04/12/2023

Force transmitted to support

P k 2  (c ) 2
AFT 
k (1  r 2 ) 2  (2 r ) 2
1  (2 r ) 2
P
(1  r 2 ) 2  (2 r ) 2
AFT 1  (2 r ) 2
   Tr
P (1  r )  (2 r )
2 2 2

Tr is called transmissibility of the system

26
 04/12/2023

Force transmitted to support

fo ace
Pl

h
gr r
ap

27
 04/12/2023

Forced vibration
x(t )
Equation of motion, k f (t )

 kx  cx  F (t )
mx m
Let , F (t )  P sin t c

xc (t )  e nt [ A cos d t  B sin d t ]  Transient response

1
x p (t )  xst sin(t   )  Steady state response
(1  r )  (2 r )
2 2 2

28
 04/12/2023

Forced vibration
x(t )
k f (t )
Equation of motion,
m
 kx  cx  F (t )
mx
Let , F (t )  P
c

 xc (t )  e nt [ A cos d t  B sin d t ]

P
 x p (t ) 
k
n t P
 x(t )  e [ A cos d t  B sin d t ] 
k
Imposing initial conditions,
x(0)  x0 ;
x(0)  x0 ;
29
 04/12/2023

P
A  ( x0  )
k
P
x 0  n ( x0  )
and , B  k
d
P
 x (t )  e nt [( x0  ) cos d t
k
P
x0  n ( x0  )
k P
 sin d t ] 
d k

30
 04/12/2023

For a special case if, x(0)  x0  0;

x(0)  x0  0;
and, P=1
1
1
n ( )
 x (t )  e n t
[ cos d t  k sin  t ]  1
k d d
k
1  1
e n t
[ cos d t  sin d t ] 
k k 1  2 k
1 
 [1  e n t [cos d t  sin d t ]]
k 1  2

 G(t)

G(t) is called indicial response of a system.

31
 04/12/2023

Let ,
U (t )  0 when t < 0
=1 when t > 0
And let ,
f (t )  F .U (t )  F. U(t  t a )
F t a [U (t )  U (t  t a )] du
 lim f(t )  lim    (t )
ta  0
F ( t )0
ta  0
F ( t ) 0
ta dt
a a

 (t ) is Dirac Delta function.

Now the response will be,
 x (t )  F .G (t )  F .G (t  t a )
F t a [G(t )  G (t  t a )] dG
 lim x(t )  lim   h(t )
ta  0
F ( ta )  0
ta  0
F ( ta )  0
ta dt

h(t) is called impulse response function of a

system.
32
 04/12/2023

1 n t 
 h(t )   e {d sin d t  d cos d t}
k 1  2

1 
 .(n ).e nt {cos d t  sin d t}]
k 1  2

1 
 [(n ).e nt {cos d t  sin d t}
k 1  2


 e nt {d sin d t  d cos d t}]
1  2

1 nt 2
 e sin d t[ n  d ]
k 1  2

1
 e nt sin d t[ 2n  n (1   2 )]
mn 2 (1   2 )
1
 e nt sin d t
mn (1   2 )
1
 e n t sin d t
md
33
 04/12/2023

Duhamel Integral / Convolution Integral

f (t )

 T
d
t
t
x (t )   F ( ).h(t   )d
0
t
  F (   ).h(t )d
0

34
 04/12/2023

t
 x (t )   F ( ).h(t   )d
0
t
1
  sin d (t   ) d 
n ( t  )
 F ( ).e
md 0

e n t
t
 sin d t[  F ( ).en cos(d ) d ]
md 0

e n t
t
 cos d t[  F ( ).en cos(d ) d ]
md 0
t t
 A D (t )  ...d   BD (t )  ...d 
0 0

Trapizoidal or simpson's rule can be use

to evaluate the integrals.

35
 04/12/2023

Complete Solution:
 X 0  n X 0  t
X (t)  e n t
 X 0 cos d t  sin d t    h(t   ) f ( )d
 n 1   2
 0

36
 04/12/2023

Estimation of ‘C’
(Half power Bandwidth method)

Dmax Energy of a signal=  x 2 (t ).dt

Dmax
1
 Power   (t ).dt
2
2 x
T
1
r1 r2 D
(1  r 2 ) 2  (2 r ) 2
X st
X max 
2
 For half power,
X st X st

(1  r )  (2 r )
2 2 2
2 2
37
 04/12/2023

 (1  r 2 ) 2  (2 r ) 2  8 2
or , r 4  r 2 [4 2  2]  (1  8 2 )  0
2  4 2
 (4 2
 2) 2
 4(1  8 2
)
r 
2

2
 1  2 2  2 1   2
 r1  1   2  
and , r2  1   2  
r1  r2
 
2

This method is only applicable to SDOF systems.

38
 04/12/2023

Fourier Series and transform

If , f (t )

i 2 nt
x(t )    n exp( ) Periodic function
n  T
1 T /2 i 2 nt
 n   x(t ) exp( ) dt
T  T / 2 T

1 T /2 i 2 nt i 2 nt
 x(t )   [  x(t ) exp( ) dt]exp( )
n  T T T
 T / 2

Now,
2 n 1
 ; n  n ; f n  ; f n 1  f n  f n 
T T T


 [f 
T /2
 x(t )  x(s) exp( i 2 fs ) ds]exp(i 2 ft )
T / 2
n 

  X(f
n 
n ) exp(i 2 ft ) f
This is called discrete Fourier series.
39
 04/12/2023

Continuous representation…

lim x(t )   X ( f ) exp(i 2 ft )df
f  0 

Backward transform

1
  d
it
x(t )  X ( ) e
2 
Forward transform

1
X ( ) 
2 

x(t )e  it dt

40
 04/12/2023

Perseval’’s Identity

Energy of a signal
E  (t ).dt
2
x

 1
 

x (t )[
2 
 X ( )eit d  ]dt


 1
 

X( )[
2


X (t)eit dt ]d 


 
X( )[X* ( )]d 


2
X( ) d 


41
 04/12/2023

Frequency Response Function

x(t )  cx  kx  f (t )
mx
k f (t )
Let , f (t )  eit
m
and , x(t )  H ( ).eit
c Putiing in the equation of motion,
1
H ( ) 
k  m 2  ci

H ( ) is called Frequency response function.

t
 x(t )   F ( ).h(t   )d  Time Domain
0

 X( )=H ( ) F ( )  Frequency Domain

42
 04/12/2023


x(t )   h(t   ) f ( )d


 1  
  h(t   )   F ( )e d   d
i

  2  
1

  
 F ( )   h(t   )e d  d 
i

2    
1

  
 F ( )   h(u )e du  d 
i ( t  u )

2    

1

2 

F ( ) H ( )eit d 


1

2 

X ( )eit d 

43
 04/12/2023

Response due to Support Motion

ks
x t 
ms
cs
xg t 
Eqn. of motion,
 mx c( x  xg )  k ( x  xg )  0
Where, x(t ) is the absolute response.
if, xr (t )  x(t ) - xg (t )
Then,
xr  xg )  cxr  kxr  0
m( 
r  cxr  kxr   mxg
or , mx 
xr  2rr xr  r2 xr   xg
 

44
 04/12/2023

t
1
 g  sin d (t   ) d 
n ( t  )
 xr (t )  
x ( ).e
d 0

 f s ,max  k .[ xr (t )]max
 k .S d ( ,  )
S d ( ,  )  Spectral Displacement

Eqn. of motion,
r  cxr  kxr   mxg

mx
when, c=0,
k

xr  xr  n2 xr
m

S pa ( ,  )  n2 .S d ( ,  )
S pa ( ,  )  Pseudo-Spectral Acceleration

45
 04/12/2023

Now,
 f s ,max  k .S d ( ,  )  m n2 S d ( ,  )  m S pa ( ,  )

Response Spectrum is the maximum response quantity of a SDOF

system for a given earthquake and damping, for the entire range of
natural frequency.
if,

xg  eit
xr  eit , x
r  i e
it
, xr   2 eit
S pa
 S pv   n S d  2 f .S d
n
 log( S pv )  log( f )  log(2 f .S d )  Slope of 135
S pa
 log( S pv )   log( f )  log(2 f . )  Slope of 45
2

46
 04/12/2023

MDOF system
m2
x2 (t ), F2 (t)

k2 , c2
m1
x1 (t ), F1 (t)

k1 , c1

x1 (t ), F1 (t) x2 (t ), F2 (t)
k1 k2
m1 m2

c1 c2

47
 04/12/2023

Free Body Diagram

Approach
m1 x1 k2 ( x1  x2 ) m2 
x2
m1 m2
k1 x1 f1 (t ) k2 ( x2  x1 ) f 2 (t )
c1 x1 c2 ( x1  x2 ) c2 ( x2  x1 )

Dynamic equilibrium equation:

Taking summation of all forces in x direction, we get
x1  k1 x1  c1 x1  c2 ( x1  x2 )  k2 ( x1  x2 )  F1 (t )
m1
and,
x2  c2 ( x2  x1 )  k2 ( x2  x1 )  F2 (t )
m2 

48
 04/12/2023

Coupled Equation..
 m1 x1   c1  c2
0    c2   x1   k1  k2 k2   x1   F1 
            
0 x2   c2
m2    c2   x2    k2 k2   x2   F2 

or ,  M 
x  C x   K x  F 

49
 04/12/2023

Undamped Free Vibration

Dynamic equilibrium equation:
 M x   K x  0
Let ,
xi (t )  ai sin(t   ) , i  1, 2
x(t )  a sin(t   )

x(t )   a  2 sin( t   )
Putting in the equation of motion we get,
 2  M a sin( t   )   K a sin( t   )  0
or ,  K    2  M  a sin( t   )  0
For, non trivial solution,
 K    2 M   0
50
 04/12/2023

Numerical example
m2  M x   K x  0
x2 (t ), F2 (t)
10 0 
k2 , c2 Where,  M    
m1  0 15 
x1 (t ), F1 (t)
 220 120 
k1 , c1 and ,  K    
 120 120 
Now,
 K    2 M   0
m1=10 kg   2  2.958, 27.04
m2=15 kg    1.71, 5.20
k1=100 N/m
 1 1 
k2=120 N/m Eigen vectors, a   
1.58 0.42 
Mode shapes

51
 04/12/2023

Orthogonality of modes
Mode 1
Force  12 m1a11 12 m2 a21
Shape  a11 a21
Mode 2
Force  22 m1a12 22 m2 a22
Shape  a12 a22
Applying Betty's Theorem,
12 m1a11a12  12 m2 a21a22  22 m1a12 a11  22 m2 a22 a21
or ,  1
2
 22   m1a11a12  m2 a21a22   0
as,  1
2
 22   0
  m1a11a12  m2 a21a22   0  Orthogonality
Orthogonal Equations,
n

a
k 1
ki mk akj  0  if i  j

ak   M ak   0 and ak   K ak   0

T T

52
 04/12/2023

Orthogonality of modes
m2
x2 (t ), F2 (t)

k2 , c2
m1  M 
x  C x   K x  F (t )
x1 (t ), F1 (t)
Now,
k1 , c1
 K    2  M    0
 K   n2 M 
Applying Betty's Theorem for m th and n th mode
f mT n  f nT m
now, mT Kn  mT n2 M n
mT M n  0  when m  n
and , mT Kn  0  when m  n

53
 04/12/2023

m2
Damped forced vibration
x2 (t ), F2 (t)

k2 , c2
m1
x1 (t ), F1 (t)

k1 , c1

 M 
x  C x   K x  F (t )
Let us use the trasformation,
x   z
  M  
z  C  z   K  z  F (t )
or ,    M  
z    C  z     K  z    F (t )
T T T T

54
 04/12/2023

Uncoupling of Equation of
motions
   M  
z    C  z     K  z    F (t )
T T T T

By, orthogonality property

   M     M d   Modal mass  Diagonal Matrix.
T

   K     K d   Modal stiffness  Diagonal Matrix.

T

If, The structure is classically damped,

i.e C     M     K 
Then,   C    Cd   Modal Damping  Diagonal Matrix.
T

Mass Normalisation of mode shapes,


n  n  M d 
1

55
 04/12/2023

m2 Damped forced vibration

x2 (t ), F2 (t)

k2 , c2
m1
x1 (t ), F1 (t)

k1 , c1

 M 
x  C x   K x  F (t )
Let us use the trasformation,
x   z
  M  
z  C  z   K  z  F (t )
or ,    M  
z    C  z     K  z    F (t )
T T T T

if ,   is mass normalised mode shapes, and C  is classical damping,

then,    M     I 
T

   K    n2  
T

and ,   C    2nn 

T

56
 04/12/2023

Solution of α and β
Let,
C     M     K 
   C        M        K  
T T T

if ,    Mass normalised mode shapes

 2 nn     I    n2 


Consideri ng , First two modes,

211    12
2 22    22
When, Two equations, and two unknowns are present,
exact solution can be found.
When, More than two equations, and two unknowns are present,
Least square solution can be found.

57
 04/12/2023

Damped vibration under m2

x2 (t )
support motion k2 , c2
 M 
x  C x   K x   M r x g (t) m1
x1 (t )
1
r  ..  influence vector k1 , c1
1
 
Let us use the trasformation, 
xg (t )
x   z
 M   z  C  z   K  z   M r  x g (t)
or ,    M  z    C  z     K  z     M r xg (t)
T T T T

    M r
T

 I 
z  2 nn  z  n  z  
2

xg (t)
   M  
 T

 I  z  2 nn  z   n2  z  Fd (t )



where, Fd (t )  Modal Force

   M r  Modal Mass Participation Factor
T

   M  
T

58
 04/12/2023

Time domain solution

    n n     n 
I 
z  2  z   2

 z  Fd (t )
Initial conditions,
 X t 0   Z t 0
 X 0    Z 0 
and ,  X0    Z 0 
Using complete time domain solution for SDOF systems,
t
z i (t)  e ini t  Ai cos d t  Bi sin d t    hni (t   ) Fdni ( )d
0

59
 04/12/2023

Frequency domain solution

 M 
x  C x   K x  F (t)
Using fourier transform,

1
 x( )e d
i t
x(t ) 
2 

1   1  
 
 M    2
x(  ) e i t
d    C   i x( ) e it
d  
2    2   
1   1  
 
 K    x( )e d   
i t
  F ( )e d  
it

2    2   


 {[K  M   iC ] X ( )  F( )}eit d   0

2
or ,


The dynamic equilibrium holds for every  ,

so using localisation theorem
[K  M  2  iC ] X ( )  F( )  0
or , X ( )  [K  M  2  iC ]1 F( )
 X ( )  H ( ) F( )
H ( )  [K  M  2  iC ]1

60
 04/12/2023

Frequency domain solution in

modal coordinates
    n n     n   z  Fd (t )
I 
z  2  z   2

r th equation will be
zr  2 rr zr  r2 zr  Fdr (t )

1
 zr ( )  ( 2 ) Fd ( )
r

r    2irr
2

 zr ( )  H r ( ) Fdr ( )
1
 H r ( )  2
r   2  2irr
Now,
X ( )   Z ( )
61