Dynamics of Vibrating Systems Unit 4

Modelling damped systems

 What did we do so far?
Unit 2: Looked at undamped systems
– Undamped systems are easier to analyse
– Many real systems are lightly damped: important aspects of their behaviour can be
deduced from undamped analysis
Unit 3: Looked at different damping types and equivalent damping

4
 Damping
Damping force always acts in the opposite direction to velocity
Different dependencies on velocity:
– constant - Coulomb (friction)
– proportional to velocity (viscous)
Solid structural
– proportional to v2 (drag) Displacement squared
– Proportional to |displacement| Hysteretic (f(T))
– Proportional to displacement2

5
 Damping equivalence
𝑥ሷ + 𝑏𝑥ሶ + 𝜔2 𝑥 = 0
Damping model application Damping force Equivalent viscous
damping beq
Linear viscous Slow fluid 𝑏𝑥ሶ 𝑏
Air damping Fast fluid ሶ 𝑥ሶ 2
𝑎 𝑠𝑔𝑛(𝑥) 8𝑎𝜔𝑋
3𝜋
Coulomb damping Sliding friction 𝛽 sgn 𝑥ሶ 4𝛽
𝛽 = 𝜇𝑚𝑔 𝜋𝜔𝑋
Displacement squared Material damping ሶ 2
d sgn(𝑥)𝑥 4𝑑𝑋
damping 3𝜋𝜔
Solid or structural damping Internal damping 𝑐 sgn 𝑥ሶ |𝑥| 2𝑐
𝜋𝜔

JRK and OT:6

 What’s next?
Adding viscous damping to equations of motion: free 1 DoF
vibrations with viscous damping
2 DoF, viscous damping – matrix methods

Later:
– Lagrange equation and multiple degree of freedom systems
– Forced vibration and analysis
– More general tools to analyse n-DoF systems

7
Unit 4.1
Adding viscous damping to
equations of motion
Damped vibration (1DoF) Spring-mass-damper system
Viscous damping
– keeps system linear and
straightforward to solve
– Good approximation.

Viscous damping is referred as linear damping such as drag

on cars and boats. It increases linearly with velocity.

From Newton’s 2nd law: 𝐹 𝑡 = −𝑘𝑥 − 𝑏𝑥ሶ = 𝑚𝑥ሷ

9
Damped vibration (1DoF)
The equation of motion of the mass is
𝑚𝑥ሷ + 𝑏𝑥ሶ + 𝑘𝑥 = 0
𝑥 = 𝐴𝑒 𝑝𝑡 𝑏 𝑘
𝑥ሷ + 2𝜁𝜔𝑛 𝑥ሶ + 𝜔𝑛2 𝑥 =0 = 2𝜁𝜔𝑛 𝜔𝑛 =
𝑥ሶ = 𝑝𝐴𝑒 𝑝𝑡 𝑚 𝑚
𝑥ሷ = 𝑝2 𝐴𝑒 𝑝𝑡 𝑝2 𝐴𝑒 𝑝𝑡 + 2𝜁𝜔𝑛 𝑝𝐴𝑒 𝑝𝑡 + 𝜔𝑛2 𝐴𝑒 𝑝𝑡 = 0 𝜁 = Damping ratio
𝑥 = 𝐴1 𝑒 𝑝1 𝑡 + 𝐴2 𝑒 𝑝2 𝑡 𝑏 𝑏
𝑝2 + 2𝜁𝜔𝑛 𝑝 + 𝜔𝑛2 = 0 𝜁= =
2𝑚𝜔𝑛 2 𝑘𝑚

−𝑏 ± 𝑏 2 − 4𝑎𝑐
𝑥= 𝑝1,2 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1
2𝑎

Rao 2.6
11
Damped vibration (1DoF)
𝑥ሷ + 2𝜁𝜔𝑛 𝑥ሶ + 𝜔𝑛2 𝑥 = 0
General solution
𝑥 = 𝐴1 𝑒 𝑝1 𝑡 + 𝐴2 𝑒 𝑝2 𝑡
𝑝1,2 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1
𝜁 = Damping ratio
𝑏 𝑏
𝜁<1 𝜁= =
2𝑚𝜔𝑛 2 𝑘𝑚

𝑘
𝜔𝑛 =
𝜁=1 𝑚

𝜁>1

Rao 2.6
12
Underdamping 𝜁<1
𝑥ሷ + 2𝜁𝜔𝑛 𝑥ሶ + 𝜔𝑛2 𝑥 = 0
General solution
𝑥 = 𝐴1 𝑒 𝑝1 𝑡 + 𝐴2 𝑒 𝑝2 𝑡 𝑝1,2 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1 𝜁 = Damping ratio
𝑏 𝑏
𝜁= =
𝜁<1 𝑝1,2 = −𝜁𝜔𝑛 ± 𝑖𝜔𝑛 1 − 𝜁 2
2𝑚𝜔𝑛 2 𝑘𝑚

𝑘
complex 𝜔𝑛 =
𝑚
1−𝜁 2 𝑡 1−𝜁 2 𝑡 )
𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 (𝐴1 𝑒 𝑖𝜔𝑛 + 𝐴2 𝑒 −𝑖𝜔𝑛

𝜔𝑑 = 𝜔𝑛 1 − 𝜁 2
𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 (𝐴1 𝑒 𝑖𝜔𝑑 𝑡 + 𝐴2 𝑒 −𝑖𝜔𝑑 𝑡 )
𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 ( 𝐴1 + 𝐴2 𝑐𝑜𝑠 𝜔𝑡 + 𝑖 𝐴1 − 𝐴2 sin(𝜔𝑡))

Rao 2.6 𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 (𝐴′1 cos(𝜔𝑑 𝑡) + 𝐴′2 𝑠𝑖𝑛(𝜔𝑑 𝑡))

13
Underdamping 𝜁<1
General solution 𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 (𝐴′1 cos(𝜔𝑑 𝑡) + 𝐴′2 𝑠𝑖𝑛(𝜔𝑑 𝑡))
𝜁 = Damping ratio
How do we find constants from initial conditions? 𝑏 𝑏
𝑥(0) = 𝐴′1 𝐴′1 = 𝑥(0) 𝜁= =
2𝑚𝜔𝑛 2 𝑘𝑚
Differentiate and find velocity:
𝑘
𝑥ሶ = −𝜁𝜔𝑛 𝑒 −𝜁𝜔𝑛 𝑡 (𝐴′1 cos(𝜔𝑑 𝑡) + 𝐴′ 2 𝑠𝑖𝑛(𝜔𝑑 𝑡)) 𝜔𝑛 =
𝑚
+𝑒 −𝜁𝜔𝑛 𝑡 (- 𝜔𝑑 𝐴′1 𝑠𝑖 𝑛 𝜔𝑑 𝑡 + 𝜔𝑑 𝐴′ 2 𝑐𝑜𝑠(𝜔𝑑 𝑡))

𝑥(0)
ሶ = −𝜁𝜔𝑛 𝐴′1 + 𝜔𝑑 𝐴′ 2 𝜔𝑑 = 𝜔𝑛 1 − 𝜁 2

1
𝐴′ 2 = (𝑥ሶ 0 + 𝜁𝜔𝑛 𝑥(0))
𝜔𝑑

Rao 2.6
14
Underdamped oscillations ζ < 1
-zwnt Exponential envelope of decay.
e
Frequency of damped
oscillation:
𝜔𝑑 = 𝜔𝑛 1 − 𝜁 2

Low damping factors: close to ωn

but still ωd ≠ ωn.
If ωd ≈ ωn: don’t use ωd ⁄ωn to
calculate ζ.

15
Critical damping 𝜁 = 1
𝑥ሷ + 2𝜁𝜔𝑛 𝑥ሶ + 𝜔𝑛2 𝑥 = 0
General solution
𝑝1,2 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1 𝜁 = Damping ratio
𝑝𝑡
𝑥 = (𝐴1 +𝐴2 𝑡)𝑒 𝑏 𝑏
𝜁= =
2𝑚𝜔 𝑛 2 𝑘𝑚
𝜁=1 Implications:
𝑝1,2 = −𝜔𝑛 𝑘
𝑥 = (𝐴1 +𝐴2 𝑡) 𝑒 −𝜁𝜔𝑛 𝑡 • Real exponents 𝜔𝑛 =
𝑚
• No oscillations.
From initial conditions:
𝐴1 = 𝑥 0 𝑏 = 2𝜁 𝑘𝑚 ⇒
𝐴2 = 𝑥ሶ 0 + 𝜔𝑛 𝑥 0 𝑏𝑐𝑟𝑖𝑡𝑖𝑐𝑎𝑙𝑙𝑦 𝑑𝑎𝑚𝑝𝑒𝑑 = 2 𝑘𝑚

Rao 2.6
16
Overdamping 𝜁>1
𝑥ሷ + 2𝜁𝜔𝑛 𝑥ሶ + 𝜔𝑛2 𝑥 = 0
General solution
𝑥 = 𝐴1 𝑒 𝑝1 𝑡 + 𝐴2 𝑒 𝑝2 𝑡 𝑝1,2 = −𝜁𝜔𝑛 ± 𝜔𝑛 𝜁 2 − 1 𝜁 = Damping ratio
𝑏 𝑏
𝜁= =
2𝑚𝜔 𝑛 2 𝑘𝑚
𝜁>1 𝜁 2 −1 𝜁 2 −1 )
𝑥 = (𝐴1 𝑒 −𝜁𝜔𝑛 𝑡+𝜔𝑛 𝑡 + 𝐴2 𝑒 −𝜁𝜔𝑛 𝑡−𝜔𝑛 𝑡 𝑘
Implications: 𝜔𝑛 =
𝑚
• Real exponents
• No oscillations
From initial conditions:
𝑥 0 = (𝐴1 + 𝐴2 )
𝑥ሶ 0 = (−𝜁𝜔𝑛 + 𝜔𝑛 𝜁 2 − 1)𝐴1 𝑒 −𝜁𝜔𝑛 𝑡+𝜔𝑛 𝑡 𝜁 2 −1 + (−𝜁𝜔𝑛 − 𝜔𝑛 𝜁 2 − 1)𝐴2 𝑒 −𝜁𝜔𝑛 𝑡−𝜔𝑛 𝑡 𝜁 2 −1

Rao 2.6
17
Damped free vibration (1DoF)
 < 1 underdamped
 =1 critically damped
 >1 overdamped

To calculate these I used

initial conditions of
x(0) = 1
and
ẋ(0) = 0

18
Unit 4.2
What system response can tell us:
Logarithmic decrement method
 Experimental determination of ,
when  < 0.7

𝑥 = 𝑒 −𝜁𝜔𝑛 𝑡 cos(𝜔𝑑 𝑡 + 𝜑)
𝑒 −𝜁𝜔𝑛 𝑡

20
 Logarithmic Decrement
How to determine the amount of
damping present in a system? 𝑥 = 𝐴𝑒 −𝜁𝜔𝑛 𝑡 cos(𝜔𝑑 𝑡 + 𝜑)
What else can we determine?

𝑥1 = 𝐴𝑒 −𝜁𝜔𝑛 𝑡1 cos(𝜔𝑑 𝑡1 + 𝜑)
2𝜋
𝑥2 = 𝐴𝑒 −𝜁𝜔𝑛 𝑡2 cos 𝜔𝑑 𝑡2 + 𝜑 = 𝐴𝑒 −𝜁𝜔𝑛 (𝑡1+𝜏𝑑 ) cos 𝜔𝑑 𝑡1 + +𝜑
𝜔𝑑
= 𝐴𝑒 −𝜁𝜔𝑛 (𝑡1 +𝜏𝑑 ) cos(𝜔𝑑 𝑡1 + 𝜑)

Rao 2.6
Thomson 2.7 Tongue 2.10
21
 Logarithmic Decrement
−𝜁𝜔𝑛 𝑡1 𝜔𝑑 = 𝜔𝑛 1 − 𝜁 2
𝑥1 𝐴𝑒 cos 𝜔𝑑 𝑡1 + 𝜑 𝜁𝜔𝑛 𝜏𝑑
= −𝜁𝜔 𝑡1 +𝜏𝑑 cos 𝜔 𝑡 + 𝜑
= 𝑒
𝑥2 𝐴𝑒 𝑛
𝑑 1

𝑥1 2𝜋 2𝜋 2𝜋
𝛿 = ln = 𝜁𝜔𝑛 𝜏𝑑 = 𝜁𝜔𝑛 = 𝜁𝜔𝑛 =𝜁
𝑥2 𝜔𝑑 𝜔𝑛 1 − 𝜁 2 1 − 𝜁2
.

𝛿
𝜁=
2𝜋 2 + 𝛿2
𝛿
For a small ζ, 𝜁≈
2𝜋

22
 Logarithmic Decrement
To get more reliable results, sometimes the time for n periods is
used:
1 𝑥1 2𝜋
𝛿 = ln =𝜁
n 𝑥𝑛 1 − 𝜁2

By using more than one time interval, you can check whether the
damping is linear or not.
Will work for half cycles!

23
 Damped vibration maths sheet
1 𝑥1 2𝜋 𝑘
𝛿 = ln =𝜁 𝜔𝑛 =
n 𝑥𝑛 1 − 𝜁2 𝑚

𝛿 𝜔𝑑 = 𝜔𝑛 1 − 𝜁 2
𝜁=
2𝜋 2 + 𝛿2
𝑏 𝑏
𝜁= =
2𝑚𝜔 𝑛 2 𝑘𝑚

24
 Have you ever been
on such a ride?
Oscillation period approx. 6.5s:
2𝜋
angular frequency is 𝜔𝑑 = = 0.967 𝑟𝑎𝑑/𝑠
𝑇

Decaying amplitude
Ratio between two positive
amplitudes approx. 1.3
2𝜋𝜁
𝛿 = ln 1.3 =
1 − 𝜁2

𝛿
𝜁= 𝜁 = 0.042
2𝜋 2 + 𝛿2

25
 How does this depend on mass?
• If we knew the mass, we could also calculate k and b
𝑘 𝑏 𝑏
𝜁 = 0.042 𝜔𝑑 = 𝜔𝑛 1 − 𝜁2 𝜔𝑛 = 𝜁= =
𝑚 2𝑚𝜔𝑛 2 𝑘𝑚
• Changing mass will change natural frequency and damping ratio
1. 5 adults, 70kg average mass (350kg+mcage= 400kg)
2. 4 teens, 40kg average mass (160kg total+mcage= 210kg)

𝜁𝑛1 𝑚2 𝜔𝑛1 𝑚2 𝜔𝑑1 𝜔𝑛1 1 − 𝜁12

= = 0.724 = = 0.724 = = 0. 7246
𝜁𝑛2 𝑚1 𝜔𝑛2 𝑚1 𝜔𝑑2 𝜔𝑛2 1 − 𝜁 2
2

𝜔𝑑1 0.967 Oscillation period 1 6.5s

𝜔𝑑2 = = 𝑟𝑎𝑑/𝑠 = 1.336 𝑟𝑎𝑑/𝑠
26 0.7246 0.7246 Oscillation period 2 4.7s
 Summary: Logarithmic decrement
Can get parameters such as natural frequency, damping ratio,
damping coefficient, spring stiffness - from observing system response
Need to measure
– decay of amplitude
– damped frequency

By using more than one time interval, can check whether

the damping is linear or not

27
 Examples of real systems
When would you want:
underdamped
overdamped
or critically damped oscillations?

Car suspension

Door Closer Anti-Vibration table

28
 Choice of damping ratio
Real damping:
Not ideally viscous (linear).
Maybe even time-dependent.

Designing for ζ = 1 can lead to overdamping

Fast response and settling time with ζ ≈ 0.7 often chosen. No single
optimal value:
Comfortable car suspension: ζ ≈ 0.6..0.8
High-fidelity measurement: ζ ≈ 0.4..0.8

29
 Summary: Damped oscillator
Harmonic oscillator with damping can have one of three
damping cases:
– Underdamped ( < 1).
– Critically damped ( = 1) – rather academic.
– Overdamped ( > 1).
Critically damped quickest to return to rest

30
Unit 4.3
2 degrees of freedom
Matrix methods for damped systems
 Undamped systems – a recap
Lumped systems without damping
Equations of motion: Matlab can solve this by
code:
𝑴𝑿ሷ + 𝑲𝑿 = 𝟎 [v,e] = eig (K, M)
V is eigenvector and e is
Solution form: x(t) = x0exp(iωt) eigenvalue
You need to take sqrt from
Standard eigenvalue solution form e to get natural frequency

A x = ω2 x
(with A = M-1 K)

32
 Damped free vibration of MDOF
Equation of motion is given by

We could attempt to solve this by assuming x(t) = Peλt

(λ2 𝐌 + λ𝐂 + 𝐊)𝐱 = 𝟎

Matlab can solve standard eigenvalue problems (A-λB) but hard to get
eigenvalues and eigenvectors when C is present

Tongue 4.10

35
 For the next session
Videos
Quiz
Examples sheet
Further challenge tasks:
– derive equivalent damping for parallel and series dampers. Hint – consider
damping force being equal
– Rao book, problems 2.129 and 2.130 on page 283
Do not forget the DOF sketch competition!

36