Dynamics of Structures

A Brief Introduction

Jag Humar, PhD, P.Eng., FCAE

Distinguished Research Professor
Carleton University
 Structural dynamics
⚫ Structural dynamics deals with the response of
structures to time varying forces such as
– Wind loads
– Earthquakes
– Forces caused by oscillating machinery
– Forces induced by ocean waves
– Forces caused by travelling loads
– Blast loads
⚫ Analysis of response to dynamic forces includes the
determination of deformations and stresses in the
structure
 Degrees of freedom
⚫ The number of degrees of freedom is the minimum number of
independent coordinates that must be specified to determine the
position of a body in space.
⚫ A point moving on a plane has two degrees of freedom: for
example, the translations along x and y coordinates
⚫ A finite size object moving in a plane has three degrees of
freedom: translations along x and y coordinates and rotation
about the z axis. The same object moving in space has six
degrees of freedom: three translations and three rotations
⚫ Real structures have an infinite number of degrees of freedom,
but many can be idealized as single degree of freedom systems
Examples of single-degree-of-
freedom systems
EI No mass p(t)
m
A=∞ u
c u

m p(t) m rigid p(t)

u
k
A=∞
No mass
H c No
E EI
I mass
 Examples of systems with
multiple degrees of freedom
EI m
ഥ u(x, t)
(a)

u1 u2 u3 u4 u5 u6 u7
(b)

The beam shown in the Figure (a) has an infinite number of degrees of
freedom but can be idealized as a system with finite number of degrees of
freedom consisting of lumped masses connected by mass-less deformable
elements as shown in Figure (b)
Single degree of freedom

Rigid beams
Axially rigid, light
columns

Two degrees of freedom

mu + cu + ku = p(t )

p1 (t ) = f S = ku

p 2 (t ) = f D = cu

p 3 (t ) = f I = mu
 Inertia force, D'Alembert's principle
⚫ Newton’s second law of motion
states: the rate of change of
momentum is proportional to the
net force acting on the system
⚫ The term mu has the unit of force
and can be viewed as a force that
acts in a direction opposite to the
direction of motion, is fI
⚫ proportional to the acceleration,
and is balanced by part of the m
external force, P1(t). Force mu is
known as the inertia force. This
conceptualization is referred to as
the d'Alembert's principle. 𝑢ሷ
 Damping force
⚫ Dynamic motion usually encounters resistance arising from dry friction,
air resistance, and structural straining. It is customary to lump all of the
resisting forces and represent them as viscous damping force that
opposes motion and is proportional to the velocity, similar to the
resistance encountered by a piston moving in a cylinder full of oil. Thus

𝑓𝐷

c
▪ where c is known as the
c
damping constant and has the
units of force per unit velocity
𝑢ሶ
 Spring force
⚫ An elastic body undergoing
deformation under an applied
force sets up internal forces
that resist the deformation. f S = ku
The force of elastic constraint
or spring force is given by

⚫ where k is known as the 𝑓𝑆

spring constant; it has the
units of force per unit k

displacement.
u
k has the units of force c has the units of
per unit displacement or force per unit velocity
N/m or N.s/m
 f I = mu t
f D = cu
f S = ku

mu t + cu + ku = 0
u = u + u g
t
mu + cu + ku = −mu g
 Undamped free vibration
 + ku = 0
mu
⚫ Solution: uc (t ) = A cos t + B sin t

⚫ Where  = m k
is known as the natural circular
frequency and has the unit 1/sec
2 m
⚫ T=

= 2
k is the natural period with the unit of second
1
⚫ f =
T
is the natural frequency in cycles per second or Hertz
⚫ Constants A and B are determined from the given initial
displacement and the initial velocity
v0
u (t ) = u 0 cos t + sin  t

Prescribed initial conditions

60
50
40
30
Displacement, mm

20
10
0
-10 0 2 4 6 8 10 12 14 16 18 20
-20
-30
-40
-50
-60
Time, s

u(0)=50 mm v(0)=20 m/s

Undamped free vibrations mu + ku = 0,  = k m
T = 2π/ω in seconds is called period
Undamped free vibration

u =  sin (t + )
k 2
= T=
m 
u0 
2
 v0 
 = u + 
2
tan  =

0
v0
Damped free vibration system
mu + cu + ku = 0
⚫Try a solution of the form
u = Ge t , u = Ge t , u = G2 e t
⚫ Substitution in the equation of motion gives
m 2 + c + k = 0
2
c  c  k
1 =− +   −
2m  2 m  m
2
c  c  k
2 =− −   −
2m  2m  m
Damped free vibration system
⚫ Critically-damped system
c = c crit = 2 km = 2m
c
1 =  2 = − = −
2m

u (t ) = (C1 + C 2 t )e −t

⚫ On using initial conditions

u( t ) = u 0 + (u 0 + u 0 ) t e −t
Critically-damped system
Damped free vibration system
⚫ Over-damped system
c  2 k m = c crit = 2m  1
1 = − +   2 − 1
 2 = − −   2 − 1
(
u (t ) = C1e t t + C 2 e 2t = e − t C1e  t + C 2 e − t )
⚫ where  =   2 − 1

⚫ Constants C1 and C2 are determined from initial

conditions
Over-damped free vibration system
 Damped free vibration
⚫ Under-damped system
c  2 k m = c crit = 2m  1
1 = − + i 1 −  2
 2 = − − i 1 −  2
(
u (t ) = C1e t t + C 2 e 2t = e −t C1e d t + C 2 e −d t )
⚫ Where d =  1 −  2
 Damped free vibration
⚫ Under-damped system
⚫ Use de Movire's theorem

e ix = cos( x ) + i sin ( x )
e −ix = cos( x ) − i sin ( x )
u (t ) = e − t ( Acos  d t + Bsin d t )
⚫ Substitution of initial conditions gives
− t  v0 +  u 0 
u(t ) = e  u 0 cos d t + sin d t 
 d 
Damped free vibration response

25

20

15
Dispmacent, mm

10

0
0 50 100 150 200 250
-5

-10

-15

-20
Time, s

critically damped overdamped underdamped

 c
=
2 km

 = 1 critically damped
 > 1 overdamped
 < 1 underdamped

mu + cu + ku = 0
Figure:: Free vibrations of a damped system with different levels of damping
Damped free vibrations mu + cu + ku = 0
Damped free vibration

u = e −  t  sin (d t + )
2
 v0 + u0 
 = u + 
2

d
0
 
d =  1 −  2
Td = 2 d
u 0 d
tan  =
v0 + u0
Undamped free vibrations, m = 100 kg, k = 3,950 N/m, T = 1s
Undamped free vibrations, m = 100 kg, k = 987 N/m, T = 2s
Undamped free vibrations, m = 100 kg, k = 15,800 N/m, T = 0.5s
Undamped free vibrations, m = 100 kg, k = 35,500 N/m, T = 1/3s
Damped free vibrations, m = 100 kg, k = 3,950 N/m, damping 5%
 m = 91000 kg
m
u 0 = 30 mm
u1 = 20 mm
Td = 0.5 s

Find k, c, 
u ( t1 ) = e −  t1
sin (d t1 + )
t 2 = t1 + 2 d
u (t 2 ) = e −  ( t1 + 2  d )
sin (d t1 + )
u (t1 ) 2  d 2   1−  2 2 
=r=e =e e
u (t 2 )
 = ln (r ) = ln (30 20 ) = 2
 = 0.0645
Td = 0.5 s
d = 2 Td = 12.566
 = d 1 −  = 12.593
2

k = m = 1.443 10 N / m
2 7

c =   c crit = 2 km = 14790 N.s/m

Harmonic forcing function
mu + cu + ku = p 0 sin t

Frequency ratio =

Transient Solution

ut = e −  t
A cos d t + B sin d t
Harmonic forcing function

Steady state solution

u s =  sin (t − )
p0 1
=
k (1 −  ) + (2)
2 2 2

2
tan  =
1−  2
Transient solution

Steady state solution

Total solution

For  = 1
 1
Peak 
2
Variation of phase angle with frequency ratio and damping ratio
General dynamic loading

mu + cu + ku = p(t )

Response to ground motion

mu + cu + ku = − mug

 Response to impulsive load
An impulsive load is a load of short duration but a large magnitude. The
area under the force time graph is called impulse of the load. In
accordance with Newton's second law, the action of an impulse causes
change in the momentum. Thus if I =Fdt denotes the impulse, we have
mdv = I ,
A system at rest at time zero imparted an impulse will end up having a
velocity v 0 = I m
The displacement response of an un-damped
system with initial velocity v0 is given by

v0
u(t ) =
I Fdt
sin  t = sin  t = sin  t
 m m
For a damped system

u (t ) =
I
e − t sin  d t
m d
Response to Impulsive load,
Duhamel's Integral
F ( )d
du (t ) = sin  (t −  )
m

⚫ For damped system

t
1
 F(  )e
− ( t − )
u( t ) = sin  d ( t −  )d
m d 0

⚫ Duhamel’s integral for

undamped System
t
1
u (t ) = 
m 0
F ( ) sin  (t −  )d
 Solution to equation of motion
The solution to the equation of motion is given as the sum of
complementary and particular solutions:

For undamped SDOF system

t
u 01
u (t ) = u 0 cost + sin t +
 
m 0
F ( ) sin  (t −  )d

For damped SDOF systems

−t  u 0 + u 0 
t
1
 F(  )e
− ( t − )
u( t ) = e  u
 0 cos  t + sin  t
d 
 + sin  d ( t −  )d
d  m d
d
 0
Numerical integration

Rectangular
summation

Trapezoidal
method

Simpson’s method
 Duhamel’s integral

Steady state response

Transient component

− t  v0 + u 0 
ut = e  u 0 cos d t + sin  d t 
 d 
 Direct integration methods
⚫ Solutions based on Duhamel's Integral are valid only for
linear systems where superposition can be used
⚫ For nonlinear systems the methods of direct numerical
integration may be used
⚫ A class of direct numerical acceleration methods known by
the general term Newmark-β methods include
– Constant acceleration method
– Average acceleration method
– Linear acceleration method
 Average acceleration method
(t ) = (un + un +1 )
1
u
2
Integrating and substituting u = u n
at t = 0 and u = u n +1 at t = h
u (t ) = un +(un + un+1 )
t
2
un+1 = un + (un + un+1 )
h
(a)
2

Integrating u(t ) and substituting u = un

at t = 0 and u = u n+1 at t = h
h2
u n +1 = u n + hu n + (un + un +1 ) (b)
4
Average acceleration method
⚫ The equation of motion at n+1 is
(c) mu + cu + ku = p
n +1 n +1 n +1 n +1

⚫ Simultaneous solution of Equations (a) and (b)

from previous slide and Equation (c) above
gives u n +1
 4m 2c  4 4  2 
 2 + + k  un+1 = pn+1 + m  2 un + un + un  + c  un + un  (d)
 
h h  h h  h 
⚫ Next u n +1 and un +1 can be obtained from
Equations a and b of the previous slide
Direct Time integration

Linear acceleration method

 Linear acceleration method

 6 m 2c   6 6 
 2 + + k u n +1 = p n +1 + m 2 u n + u n + 2u n  +
 
h h  h h 
3 h 
c u n + 2u n + u n 
h 2 

u n +1 = 2 (u n +1 − u n − hu n ) − 2u n

6
h
u n +1 = u n + (u n + u n +1 )
h
2
 Response to ground motion

(Figure based on Chopra)

u + u + u = −u g (t )
c k
m m u max (,  ) = SD
u + 2u + 2 u = −u g (t ) = spectral displacement
 Displaceme nt SD

Velocity SV = SD

Accelerati on SA =  SD
2

(Figure based on Chopra)

k 2 (u1 − u 2 )
c 2 (u 1 − u 2 )
m1u1
 Equations of motion

m1 0 0   u1  c1 + c 2 − c2 0   u 1 

0 m       
 2 0  u 2  +  − c 2 c 2 + c 3 − c 3   2
u
 0 0 m 3  u 3   0 − c3 c 3  u 3 
k1 + k 2 − k2 0   u1   p1 
     
+  − k2 k 2 + k 3 − k 3  u 2  = p 2 
 0 − k3 k 3  u 3  p 3 

 + Cu + Ku = p
Mu
Mu + Cu + Ku = 0
 t

u1t  u1  ug 

 t    
u2  = u2  + ug 
ut  u  u 
 3  3  g
 = u
u t
 + u
 g = u
 + 1ug
 + Cu + Ku = −Mu
Mu  g
 Free vibration response

Mu + Ku = 0

u = A1q1 sin (1t + 1 ) + A 2q 2 sin (2 t +  2 ) +   

+ A n q n sin (n t +  n )

=  A jq j ( j t +  j )
N

j=1
Mode shapes and frequencies

An N story building has N frequencies

1 , 2 ,  , N
Corresponding to each frequency there is
a mode shape
q1 , q 2 ,    , q N
Properties of mode shapes

q Kq j = 0 i  j
T
i

q Mq j = 0 i  j
T
i

q iT Kq i
T
= i
2

q i Mqi

Modal matrix Q = q1 q 2   q N 

  2m 0 0  3k 0 − k0 
M= K= 
 0 m 0  − k 0 k0 
 k0 1  2k 0 1
1 = q1 =   2 = q2 =  
2m 0  2 m0 − 1
Vibrations in First Mode
Vibrations in Second Mode
Vibrations in Third Mode
Mode superposition analysis

 + Cu + Ku = −M1ug
Mu
Let u =  q n yn
Substitute in the equation of motion
Multiply both sides by q nT , we get
q Tn Mqn yn + q Tn Cq n y n + q nT Kq n yn = −q Tn M1ug
or
T
q n M1
yn + 2n n y n + n yn = − T
2
ug = n ug
q n Mqn
Mode superposition analysis
Solve the SDOF equations using say the
first M modes
yn + 2n n y n +  yn = nug
2
n

where the participation factor

q Tn M1
n = T
q n Mqn
Mode superposition analysis
Now obtain response in the physical
coordinates
M
u =  q n yn
n =1

The forces are given by

M M
f =  Kq n yn =  n2 Mqn yn
n =1 n =1
 References

1. Jagmohan L. Humar. 2002. Dynamics of

Structures, 2nd Edition, A.A. Balkema
Publishers, Lisse, The Netherlands
2. Anil K. Chopra. 2000. Dynamics of
Structures: Theory and Application to
Earthquake Engineering, 2nd Edition, Prentice
Hall, Upper Saddle River, New Jersey