CE 809 - Structural Dynamics

Lecture 2: Free Vibration Response of SDF Systems

Semester – Fall 2020

Dr. Fawad A. Najam Prof. Dr. Pennung Warnitchai

Department of Structural Engineering Head, Department of Civil and Infrastructure Engineering
NUST Institute of Civil Engineering (NICE) School of Engineering and Technology (SET)
National University of Sciences and Technology (NUST) Asian Institute of Technology (AIT)
H-12 Islamabad, Pakistan Bangkok, Thailand
Cell: 92-334-5192533, Email: [email protected]
 Equation of Motion of One-story Building

𝑑 2 𝒖 (𝑡)
𝑡 𝑑𝒖 (𝑡)
𝑡 𝑑 2 𝒖𝒈(𝑡)
𝑡 𝑑 2 𝒖(𝑡)
(𝑡) 𝑑𝒖(𝑡)
(𝑡)
𝑚 +𝑐 (𝑡)
+ 𝑘 𝒖 𝑡 = −𝑚 𝑚 2
+ 𝑐 + 𝑘 𝒖(𝑡) 𝑷 𝒆𝒇𝒇 𝑡
(𝑡) = 𝑝(𝑡)
(𝑡)
𝑑𝑡 2 𝑑𝑡 𝑑𝑡 2 𝑑𝑡 𝑑𝑡
𝑚 𝑚
𝑷𝒆𝒇𝒇 𝑡

(𝑡)
𝒖𝒈 (𝑡) Fixed Base

The deformation 𝒖(𝑡)

(𝑡) of the structure due to ground acceleration 𝒖𝒈 (𝑡)
(𝑡) is identical to the deformation
𝒖(𝑡)
(𝑡) of the structure if its base were stationary and if it were subjected to an external force 𝑷𝒆𝒇𝒇 (𝑡)
𝑡 =
− 𝑚𝒖𝒈 (𝑡). 2
 Free Vibration Response of SDF Systems

Free vibration response: the motion of an SDF system with the applied force set equal to zero.

Free vibration response in mathematical terms is the mathematical solution of the following
homogeneous differential equation:

𝑚 𝑢 𝑡𝑡 + 𝑐 𝑢 𝑡𝑡 + 𝑘 𝑢 𝑡𝑡 = 0 Equation (1)

3
 A Quick Review of Basic Mathematical Concepts

Solution form:

Consider a first-order differential equation By applying exponential operation,

𝑑𝑢(𝑡)
𝑡
+𝑘𝑢 𝑡 =0 𝑒 𝑙𝑛(𝑢 𝑡 ) 𝑡 = 𝑒 (−𝑘 𝑡+𝑐) = 𝑒 −𝑘 𝑡 𝑒 𝑐 = 𝑐0 𝑒 −𝑘 𝑡
= 𝑢(𝑡)
𝑑𝑡
𝑑𝑢(𝑡)
𝑡 The solution:
= −𝑘 𝑢(𝑡)
𝑡
𝑑𝑡 𝑢 𝑡 = 𝑐0 𝑒 −𝑘 𝑡
By separation of variables,

𝑑𝑢(𝑡)
𝑡 where 𝑐0 is an arbitrary constant.
= −𝑘 𝑑𝑡
𝑢(𝑡)
𝑡
Integrate both sides, It can be shown that the solutions of higher order
differential equations are also in this exponential
𝑙𝑛(𝑢 𝑡 ) = −𝑘 𝑡 + 𝑐 form.

Where 𝑐 is an arbitrary constant.

4
 A Quick Review of Basic Mathematical Concepts

Superposition:
Substituting 𝑐1 𝜙1 (𝑡 info the left-hand side of equation
If a solution of a homogeneous linear differential of motion (Eq (1)), we get
equation is the multiplied by a constant, the resulting
function is also a solution. 𝑚 (𝑐1 𝜙1 𝑡(𝑡) + 𝑐 (𝑐1 𝜙1 𝑡𝑡 ) + 𝑘 (𝑐1 𝜙1 𝑡𝑡 ) =

The sum of two solutions is also a solution. 𝑐1 [𝑚 𝜙1 (𝑡

𝑡 + 𝑐 𝜙1 𝑡𝑡 + 𝑘 𝜙1 𝑡𝑡 ] = 𝑐1 . 0 = 0

Proof:
Hence 𝑐1 𝜙1 (𝑡)
𝑡 is also a solution of the equation of
Let 𝜙1 (𝑡)
𝑡 and 𝜙2 (𝑡)𝑡 be independent solutions of motion (Eq (1)).
governing differential equation of an SDF system,
such that In similar manner, by a direct substitution of
𝑚 𝜙1 𝑡𝑡 + 𝑐 𝜙1 𝑡𝑡 + 𝑘 𝜙1 𝑡𝑡 = 0 𝑐1 𝜙1 𝑡𝑡 + 𝑐2 𝜙2 (𝑡)
𝑡 into the left-hand side of Eq (1),
it can be shown that 𝑐1 𝜙1 𝑡𝑡 + 𝑐2 𝜙2 (𝑡)
𝑡 is also a
𝑚 𝜙2 (𝑡)
𝑡 + 𝑐 𝜙2 𝑡𝑡 + 𝑘 𝜙2 𝑡𝑡 = 0 solution of the equation of motion.

5
 A Quick Review of Basic Mathematical Concepts

Initial Conditions
Two conditions are needed because there are two
Consider 𝑢 𝑡 = 𝑐1 𝜙 𝜙11 𝑡𝑡𝑡 +
+𝑐𝑐22𝜙𝜙22(𝑡)
(𝑡)
𝑡 as a general unknown arbitrary constants to be specified.
solution of the governing equation of motion. Since
the constants 𝑐1 and 𝑐2 can have any value, the 𝑢 00 = 𝑐𝑐11𝜙𝜙
1 100𝑡 ++𝑐2
𝑐2𝜙𝜙 2 (𝑡)
2 (0)0
general solution can represent ∞ different solutions. 𝑢 00 = 𝑐1𝑐1𝜙𝜙
1 100𝑡 ++𝑐2
𝑐2𝜙𝜙22(0)
(𝑡)
0

Usually initial conditions are known and we are

seeking for one specific solution that satisfies these 𝜙1 00 , 𝜙2 00 , 𝜙1 00 , 𝜙2 00 , 𝑢 00 and 𝑢 00 all are
initial conditions. known. Therefore 𝑐1 and 𝑐2 can be determined.

Example of initial conditions: [For more details, see Erwin Kreyszig’s Advanced
Engineering Mathematics, John Wiley & Sons.]
𝑢(0)
0 and 𝑢(0)
0 are the initial displacement and initial
velocity of the SDF system.

6
 Free Vibration Response of SDF Systems (continued)
Now consider the equation governing the free vibration of an SDF system:

𝑚 𝑢 𝑡 + 𝑐 𝑢 𝑡 + 𝑘 𝑢 𝑡𝑡 = 0 ……………. (1)

Assuming that the solution of Eq (1) is in the exponential form:

𝑢 𝑡𝑡 = 𝐺 𝑒 𝑠 𝑡 ……………. (2)

where 𝐺 and 𝑠 are constants.

Substituting this solution into the equation of motion (Eq (1)),

𝑚 (𝑠 2 𝐺 𝑒 𝑠 𝑡 ) + 𝑐 (𝑠 𝐺 𝑒 𝑠 𝑡 ) + 𝑘 (𝐺 𝑒 𝑠 𝑡 ) = 0

𝑚 𝑠2 + 𝑐 𝑠 + 𝑘 𝐺 𝑒 𝑠 𝑡 = 0 ……………. (3)

2
To have a non-zero solution of 𝑢(𝑡)
𝑡 , the term (𝑚 𝑠 + 𝑐 𝑠 + 𝑘) must be zero,
2 𝑐 𝑐 𝑘
𝑘
2 𝑠 +𝑚𝑠 +𝑚 = 0
𝑠 + 𝑠+ =0 ……………. (4)
𝑚 𝑚 7
 Case 1: Undamped Free Vibration Response
In this case, 𝑐 = 0.
Introducing the notation

𝑘
𝜔=
𝑚
The equation (4) becomes,
𝑠 2 + 𝜔2 = 0 ……………. (5)

Which has two solutions,

𝑠 =±𝑖𝜔 ……………. (6)
Where 𝑖 = −1

Hence the general solution of 𝑢(𝑡)

𝑡 is

𝑢 𝑡 = 𝐺1 𝑒 𝑖 𝜔 𝑡 + 𝐺2 𝑒 − 𝑖 𝜔 𝑡 ……………. (7)

Where 𝐺1 and 𝐺2 are arbitrary constants.

8
 Case 1: Undamped Free Vibration Response (continued)

𝑢 𝑡 = 𝐺1 𝑒 𝑖 𝜔 𝑡 + 𝐺2 𝑒 − 𝑖 𝜔 𝑡 ……………. (7)

Since there are two arbitrary constants, two initial conditions need to specified, i.e. 𝑢(0)
0 and 𝑢(0)
0 .

𝑢 00 = 𝐺1 𝑒 0 + 𝐺2 𝑒 0 = 𝐺1 + 𝐺2

𝑢 00 = 𝑖 𝜔 𝐺1 𝑒 0 − 𝑖 𝜔 𝐺2 𝑒 0 = 𝑖 𝜔 𝐺1 − 𝑖 𝜔 𝐺2
Therefore,

1 𝑢 00
𝐺1 = 𝑢 00 +
2 𝑖𝜔
……………. (8)
1 𝑢 00
𝐺2 = 𝑢 00 −
2 𝑖𝜔

9
 A Quick Review of Basic Mathematical Concepts

Taylor Series of 𝒆𝒙 (expand around x = 0): Taylor series of cos 𝜔𝑡 is

𝑥 𝑥 𝑥2 𝑥3 (𝜔 𝑡)2
𝑒 =1+ + + + ⋯ for −∞ < 𝑥 < ∞ 1− +⋯
1! 2! 3!
2!
(𝑖 𝜔 𝑡)2 (𝑖 𝜔 𝑡)3 Similarly, the Taylor series of sin 𝜔𝑡 is
𝑒𝑖 𝜔 𝑡 =1+𝑖𝜔𝑡+ + +⋯
2! 3!
(𝜔 𝑡)3
(𝜔 𝑡)2 𝑖 (𝜔 𝑡)3 𝜔𝑡− +⋯
𝑒𝑖 𝜔 𝑡 = 1 + 𝑖 𝜔 𝑡 + (−1) + (−1) +⋯ 3!
2! 3!
Therefore,
(𝜔 𝑡)2 (𝜔 𝑡)3 𝑒 𝑖 𝜔 𝑡 = cos 𝜔𝑡 + 𝑖 sin(𝜔𝑡)
𝑒𝑖 𝜔 𝑡 = 1− +⋯ +𝑖 𝜔𝑡− +⋯
2! 3!

This is called Euler’s equation.

10
 Case 1: Undamped Free Vibration Response (continued)

Introducing the Euler's equations:

𝑒 ± 𝑖 𝜔 𝑡 = cos 𝜔𝑡 ± 𝑖 sin(𝜔𝑡) ……………. (9)

And the expressions for 𝐺1 and 𝐺2 (Eq (8)) into the solution (Eq (7)) , we obtain

𝑢 00 ……………. (10)
𝑢 𝑡 = 𝑢 0 cos 𝜔𝑡 + sin(𝜔𝑡)
𝜔

It is easy to verify that this equation is the solution of governing equation of motion by direct
substitution.

11
Case 1: Undamped Free Vibration Response (continued)
𝑢 0
𝑢 𝑡 = 𝑢 0 cos 𝜔𝑡 + sin(𝜔𝑡)
𝜔

𝑢 00 2
2 𝑢 00
Amplitude 𝜌 = 𝑢 0 +
𝜔
Displacement, 𝑢 𝑡
𝑢 0

Time 𝑡

2𝜋
𝑇=
𝜔

Deformed position of structure corresponding to location 1, 2, 3, 4 and 5 on response-time plot 12

 Case 1: Undamped Free Vibration Response (continued)

The structure vibrates in simple harmonic motion (or oscillation).

The amplitude of oscillation depends upon 𝑢(0)

0 and 𝑢(0)
0 . The above equation may be transformed
into
𝑢 𝑡 = 𝜌 cos(𝜔𝑡 − 𝜃) ……………. (11)

Where

2
2 𝑢 00
𝜌= 𝑢 00 +
𝜔

−1
𝑢 00
𝜃 = tan ……………. (12)
𝜔 𝑢(0)
0

13
 Case 1: Undamped Free Vibration Response (continued)
• The oscillation does not decay because the structure is undamped. The period of oscillation 𝑇 is
the time required for one cycle of free oscillation. For undamped structure,

2𝜋 1
𝑇= = ……………. (13)
𝜔 𝑓

Where 𝜔 is the natural circular frequency,

𝑓 is the natural (cyclic) frequency (cycle/sec, Hz), and
𝑇 is the natural period (sec)

• This term "natural" is used to qualify each of the above quantities to emphasize the fact that these
are “natural properties” of the structure.

• These properties are independent of the initial conditions.

14
 Case 2: Damped Free Vibration Response
In this case 𝑐 ≠ 0; i.e. damping is present in the structure.

𝑐 𝑘
The solutions of 𝑠2 + 𝑠+ =0 for this case are
𝑚 𝑚

𝑐 𝑐 2
𝑠=− ± − 𝜔2 ……………. (14)
2𝑚 2𝑚

𝑐 2
The characteristics of “𝑠" depends upon the sign of the term − 𝜔2
2𝑚
𝑐 2
Case 2 (a): The equation will have distinct real roots, if − 𝜔2 > 0
2𝑚
𝑐 2
Case 2 (b): The equation will have complex conjugate root, if − 𝜔2 < 0
2𝑚
𝑐 2
Case 2 (c): The equation will have real double roots, if − 𝜔2 = 0
2𝑚
15
 Case 2 (b): Underdamped Systems (𝑐 < 2 𝑚 𝜔)

Let’s define 𝑐𝑐 : critical damping: 𝑐𝑐 ≡ 2 𝑚 𝜔

𝑐 𝑐 ……………. (15)
Let’s define 𝜉 : critical damping ratio; 𝜉 ≡ =
𝑐𝑐 2𝑚𝜔

Hence, in underdamped systems, 0 < 𝜉 < 1

Rewriting the solution in terms of 𝜉 , we get 𝑠 = −𝜉 𝜔 ± 𝜉𝜔 2 − 𝜔2

𝑠 = −𝜉 𝜔 ± 𝜔 2 1 − 𝜉 2 −1
𝑠 = −𝜉 𝜔 ± 𝑖 𝜔𝐷 ……………. (16)

Where 𝜔𝐷 = 𝜔 1 − 𝜉 2 ……………. (17)

16
 Case 2 (b): Underdamped Systems (𝑐 < 2 𝑚 𝜔) (continued)

Then the general solution of 𝑢(𝑡)

𝑡 is

𝑢 𝑡 = 𝐺1 𝑒 𝑠1 𝑡 + 𝐺2 𝑒 𝑠2 𝑡 = 𝐺1 𝑒 (− 𝜉 𝜔 𝑡 + 𝑖 𝜔𝐷 𝑡) + 𝐺2 𝑒 (− 𝜉 𝜔 𝑡 − 𝑖 𝜔𝐷 𝑡)

𝑢 𝑡 = 𝑒 (− 𝜉 𝜔 𝑡) 𝐺1 𝑒 (− 𝑖 𝜔𝐷 𝑡) + 𝐺2 𝑒 (− 𝑖 𝜔𝐷 𝑡) ……………. (18)

When the initial conditions of 𝑢(0)0 and 𝑢 0 0 are introduced, the constants 𝐺1 and 𝐺2 can be
evaluated, and after using Euler’s equations we finally obtain,

(− 𝜉 𝜔 𝑡)
𝑢 00 + 𝑢 00 𝜉 𝜔 ……………. (19)
𝑢 𝑡 =𝑒 sin 𝜔𝐷 𝑡 + 𝑢 00 cos 𝜔𝐷 𝑡
𝜔𝐷
17
 Case 2 (b): Underdamped Systems (𝑐 < 2 𝑚 𝜔) (continued)

The response in above equation can also be presented as

𝑢 𝑡 = 𝑒 −𝜉 𝜔 𝑡 𝜌 cos(𝜔𝐷 𝑡 − 𝜃) ……………. (20)

Where

2
𝑢 00 + 𝑢 00 𝜉 𝜔 2
𝜌= + 𝑢 00
𝜔𝐷 ……………. (21 a, b)

−1
𝑢 00 + 𝑢 00 𝜉 𝜔
𝜃 = tan
𝜔𝐷 𝑢 00

The equation (20) says that the underdamped system in its free vibration stage will oscillate
with circular frequency 𝜔𝐷 and with exponentially decreasing amplitude.

18
Case 2 (b): Underdamped Systems (𝑐 < 2 𝑚 𝜔) (continued)

𝑢 0 Exponential Decay Undamped Structure

Damped
Structure
𝑢 𝑡
𝑢 0
Displacement

Time 𝑡

𝑇
𝑇𝐷

The effect of damping on free vibration 19

 Effect of Damping on Free Vibration

• In most structures, the critical damping

The effect of damping on
ratio 𝜉 is less than 0.2 and hence, 𝜔𝐷 = 𝜔 natural frequency of vibration
and 𝑇𝐷 = 𝑇.

Damping Ratio 𝜉
• The rate of amplitude decay depends on 𝜉 .

Range of damping
for most structures

Damped Natural Frequency 𝜔𝐷

=
Undamped Natural Frequency 𝜔 20
 Effect of Damping on Free Vibration

𝑢(0)
𝑢(𝑡)
=
Displacement at Time 𝑡
Initial Displacement

Time 𝑡
=
Natural Vibration Period 𝑇

The effect of damping on free vibration. Curves 1, 2, 3 and 4 are for

damping ratio 0, 1, 2 and 5 percent 21
 Damping in Structures

• In seismic design of most structures, 𝜉 = 0.05 Typical

is used. Type of Construction Damping
Ratios (𝝃)
• For tall buildings subjected to strong winds, we
Steel frame with welded connections and
generally assume 𝜉 = 0.005 − 0.02. 0.02
flexible walls
• For single cables, 𝜉 = 0.003 − 0.01. Steel frame with welded connections,
0.05
normal floors and exterior cladding
Steel frame with bolted connections,
0.1
normal floors and exterior cladding
Concrete frame with flexible internal walls 0.05
Concrete frame with flexible internal walls
0.07
and exterior cladding
Concrete frame with concrete or masonry
0.1
shear walls
Concrete or masonry shear wall 0.1
Wood frame and shear wall 0.15
22
 Case 2 (c): Critical Damped Systems (𝑐 = 𝑐𝑐 = 2 𝑚 𝜔)

In this case, 𝑐 = 𝑐𝑐 = 2 𝑚 𝜔 and 𝜉 = 1. This will yield,

𝑠 =−𝜔

The general solution of the governing equation of motion in this case will be of the form.

𝑢 𝑡 = 𝐺1 𝑒 𝑠 𝑡 + 𝑡 𝐺2 𝑒 𝑠 𝑡 = 𝐺1 + 𝑡 𝐺2 ) 𝑒 − 𝜔 𝑡

The second term must contain 𝑡 since the two roots of quadratic equation in 𝑠 are identical.

𝑢 𝑡 = −𝜔 𝐺1 + 𝑡 𝐺2 ) 𝑒 − 𝜔 𝑡 + 𝐺2 𝑒 − 𝜔 𝑡

23
 Case 2 (c): Critical Damped Systems (𝑐 = 𝑐𝑐 = 2 𝑚 𝜔)

Using initial conditions 𝑢(0)

0 and 𝑢(0)
0 , the constants 𝐺1 and 𝐺2 can be determined as follows.

𝐺1 = 𝑢 00
𝐺2 = 𝑢 00 + 𝜔 𝑢 00
The general solution will be,
𝑢 𝑡 = [𝑢 00 1 + 𝜔 𝑡 + 𝑢 00 𝑡] 𝑒 −𝜔 𝑡

No oscillations. Critical damping just eliminated them.

24
Case 2 (c): Critical Damped Systems (𝑐 = 𝑐𝑐 = 2 𝑚 𝜔)

Displacement
𝑢 𝑡

𝑢 00

No oscillations

𝑢 0

Free-vibration response with critical damping

(Clough and Penzien (2003) Dynamics of Structures, 3rd Edition).

25
 Case 2 (a): Overdamped Systems (𝑐 > 𝑐𝑐 )

• The response of an over-critically-damped system is similar to the motion of a critically-damped

system.
• Not encountered in practice
• No oscillations

26
 Summary

𝜉=1
𝑢(0)
𝑢(𝑡)

𝜉=2
=
Displacement at Time 𝑡
Initial Displacement

Time 𝑡
=
Natural Vibration Period 𝑇

𝜉 = 0.1

Free vibration of under-damped, critically damped, and over-damped systems

27
(Chopra (2012) Dynamics of Structures, 4th Edition)
 Decay of Free Vibration Response

𝑢 𝑡 𝑢0
𝑢1
𝑢2
𝑢3
Displacement

𝑢4 𝑢5 𝑢6
Time (𝑡)

Measured displacement response from a free-vibration test

28
 Free-vibration Tests

It can be shown that the ratio of any two successive peaks is

𝑢𝑖 𝜔
(−2 𝜋 𝜉 )
=𝑒 𝜔𝐷
𝑢𝑖+1
Taking the natural logarithm on both sides gives the logarithmic decrement 𝛿 , as follows.

𝑢𝑖 𝜔
𝛿 ≡ 𝑙𝑛 =2𝜋𝜉
𝑢𝑖+1 𝜔𝐷
Hence for structure with low 𝜉 ,
𝛿 ≈2𝜋𝜉
29
The above equation is very useful and can be used for the identification of 𝜉 in existing structures.
 Free-vibration Tests
𝑚50%

𝑢𝑢𝑖 𝑖
Sometimes it is more appropriate to consider the ratio
𝑢𝑖+𝑚
𝑢𝑖+𝑚
where 𝑚 > 1,

𝑢𝑖 𝜔
𝑙𝑛 =2𝑚𝜋𝜉
𝑢𝑖+𝑚 𝜔𝐷

1 𝑢𝑖
𝜉≈ 𝑙𝑛
2𝑚𝜋 𝑢𝑖+1

To determine the number of cycles elapsed for a 50% reduction

in displacement amplitude ( 𝑚50% ), we obtain the following
Damping Ratio 𝜉
relation from the above equation.
The number of cycles required to reduce the
0.11 free vibration amplitude by 50%
𝑚50% =
𝜉 (Chopra (2012) Dynamics of Structures, 4th Edition)
30
Thank you