Dynamics of Motion Systems

Mechatronic system design
Mechatronic system design wb2414‐2013/2014
Course part 4
Dynamics of motion
systems
Prof.ir. R.H.Munnig Schmidt

Mechatronic System Design
1
WB2414-Mechatronic System Design 2013-2014
Delft
University of
Technology
Contents
• Stiffness in Precision Engineering
• Passive and active stiffness
• Compliance of (a combination of) dynamic elements
• Dynamic modelling of damped mass‐spring systems.
• Transmissibility
• Coupled mass‐spring systems
• Eigenmodes, eigenfrequencies and modeshapes
• Standard mechanical frequency responses
2
Delft
University of
Technology
1
Stiffness of objects
Well known objects

2
10 N/m Soft pillow
Car suspension
104 N/m
Soft couch
Table
105 N/m
Bicycle
107 N/m Office building
8
10 N/m Concrete pillar
9
10 N/m Steel train wheel on steel rail track
3
Delft
University of
Technology
What is stiffness?
Hooke’s law for force from spring:
Fs  kdx
Hooke‐Newton law for
external force:
Fr  F  kdx
F
dx 
k
Where should you place the stiffness if possible?
 Take the shortest force loop
4
Delft
University of
Technology
2
Natural frequency of the resonance of
a mass-spring system
• At resonance the forces are in
balance
• Deformation force (stiffness) plus m 
acceleration force (mass) is zero.
x0
d2 x d2 x k
Fa  Fd  m 2  kx  0  m 2  kx
dt dt
x  xˆ sin (0 t) 
Stationary reference
 mxˆ02 sin(0t )   kxˆ sin(0t ) 
k
0 
m
5
Delft
University of
Technology
The first natural frequency determines

the sensitivity to harmonic vibrations!
•The maximum force needed to follow the acceleration:
Fˆ  m aˆ  m xˆ f  2
•The maximum error due to this force:
Fˆ m.xˆf . 2
xê  
k k
•The natural frequency:
k
 02
m
•Which results in:
2 0 f 0 xˆ xˆf
xê  xˆf    f  f0  f
02  f xê xê
6
Delft
University of
Technology
3
Active stiffness in a CD player, bandwidth
•200 μm radial vibrations at 25 Hz
xˆr 200 106
f0  f  25  800 Hz •Mass lens: 10 ·10‐3 kg
r 0, 2 106
1 kr •Max radial error: 0,2 μm
f0   kr  4 2 mf 02 =2.5·105 N/m
2 m
7
Delft
University of
Technology
Virtual stiffness
• Measure position
• Actuate with force proportional and opposite to the deviation (feedback!)
F  G t  r   G m Ga G p Gc  r
F
• Gives virtual spring stiffness k r  Gt 
r
8
Delft
University of
Technology
4
Contents
9
Delft
University of
Technology
Stiffness and compliance
• Stiffness it the ability of a system to withstand a force by
minimising the resulting motion/deformation
• Compliance is the opposite
• Both can be real, in phase with a periodic force, or complex,
dynamic, frequency dependent, 90o out of phase with a periodic
force.
• A spring has a real stiffness/compliance:
x 1
Cs  
F k
10
10
Delft
University of
Technology
5
Compliance of (a combination of)
dynamic elements
m x
x 1
x0 Cs  
c F k F k
• k = stiffness of the spring

• c = damping coefficient of the damper
• m = mass of the body
11
11
Delft
University of
Technology
Stiffness and compliance of a damper
m x
x0
c F k
dx
F (t )  c , F ( s )  {F (t )}  scx
dt
F ( )   {F (t )}  jc x
1 x 1
Cd ( )   
kd ( ) F jc
12
Delft
University of
Technology
6
Stiffness and compliance of a body
m x
x0
c F k
d2 x
F (t )  m 2
, F ( s )  {F (t )}  ms 2 x
dt
F ( )   {F (t )}   m 2 x
1 x 1
Cm ( )   
km ( ) F m 2
13
Delft
University of
Technology
Combined Compliance of body,

spring and damper
 1 1 1  x
Ft ( )  Fs  Fd ( )  Fm ( )  x    
 Cs Cd Cm  Ct
x 1
Ct ( )  ( ) 
Ft 1 1 1
 
Cs Cd Cm
14
Delft
University of
Technology
7
Overview of the dynamic properties
15
Delft
University of
Technology
The separate element responses in a

bode plot
1 1 k
  0 
k m02 m
Log |x/F|
x 1 Mass line (‐2)
Spring line  Cs 
Fs k x 1
Damper line (‐1)  Cm 
x 1 Fm m 2
 Cd 
Fd c
0
Spring line  0 
Phase
Damper line  90 
Mass line  180 
180
ω0 Log ω
16
Delft
University of
Technology
8
Due to energy
No damping conservation response
becomes infinite
Log |x/F| (resonance)
0
Phase
180
ω0 Log ω
17
Delft
University of
Technology
Low damping
Log |x/F|
0
Phase
180
ω0 Log ω
18
Delft
University of
Technology
9
High damping
Log |x/F|
0
Phase
180
ω0 Log ω
19
Delft
University of
Technology
Extreme damping, two first order

systems
Log |x/F|
0
Phase
180
ω0 Log ω
20
Delft
University of
Technology
10
Contents
21
Delft
University of
Technology
Start with second law of Newton

F=m.a
d2 x dx
F (t )  m 2
 c  kx)
m dt dt
x
x0 Laplace gives:
c F k
F ( s )  {F (t )}  x(ms 2  cs  k )
1
xm 1 k With only positive
Ct ( s )  (s)  2 
F ms  cs  k m s 2  cs  1 imaginary terms (Fourier):
k k
x Cs
1 k c Ct ( )   {F (t )}  ( )  2
 Cs 0    FRF F s s
k m 2 km  2 1
02 0
22
Delft
University of
Technology
11
The magnitude and the phase
Start with s=jω
Cs
Ct ( ) 
2 2
 2   
1  2    2 
 0   0 
And phase angle φ is – arctan (imaginary/real of the denominator)
23
Delft
University of
Technology
Compliance Bode plot with damping
m x
x0
c F k
• k = 104 N/m
• m = 0.25 kg
24
Delft
University of
Technology
12
The damping ratio  is related to the
pole location in the Laplace plane
s    j
Poles are those values of s where denominator of Ct is zero
1
x k Cs
Ct   
F m s 2  cs  1 s 2 s Im
 2 1
k k 0
2
0
p1    jd and p2    jd
Re
If c  0 then   0 no damping!
Cs
Ct  and p1   j0 and p2   j0
m 2
s 1
k
25
Delft
University of
Technology
Critical damping ratio   1
1
x k Cs
Ct   
F m s 2  cs  1 s 2 s Im
 2 1
k k 02 0 < 1
p1    jd and p2    jd

= 1
> 1
If c  2 km then   1 and
> 1 Re
1
k Cs Cs Cs
Ct    
m 2 cs  s  1
2 2
s  1 m 2 m  m 
s 2 s 1  s  1 < 1
k k k k
 k 
26
Delft
University of
Technology
13
Above critical damping the mass-spring
system reduces to two first order systems
1 1 k
  0 
k m02 m
Log |x/F|
x 1 Mass line (‐2)
Spring line  Cs 
Fs k x 1
Damper line (‐1)  Cm 
x 1 Fm m 2
 Cd 
Fd c
0
Spring line  0 
Phase
Damper line  90 
Mass line  180 
180
ω0 Log ω
27
Delft
University of
Technology
Effect of the damping ratio  on

the stepresponse
At ζ = 1 no overshoot,
Critical damping
time domain related term.
28
Delft
University of
Technology
14
Energy at natural Frequency
Energy will be “trapped” in the system.

When excited in this frequency the amplitude will
continue to rise.
For electrical engineering (frequency domain) the quality

factor Q is defined for this property as resonators are
also useful. But in mechanical engineering the time
domain related term “damping ratio” is more commonly
used.
Let’s bridge the gap

29
Delft
University of
Technology
Q=100 means peak level at 100 times

spring-line level
m x • k = 104 N/m
x0
c F k
• m = 0.25 kg
At Q = 1 no resonance peak
frequency domain related term
30
Delft
University of
Technology
15
Relation Q and 
1 1
Q  
2 2Q
Electrical vs mechanical, Time vs frequency domain
1 1
EE  ME 
2 ME 2 EE
31
Delft
University of
Technology
Velocity response, kinetic energy
m x
x0
c F k
• k = 104 N/m
• m = 0.25 kg
dx
v sx
 dt  
F F F
s  Cs
 2
s s
 2 1
02
0
32
Delft
University of
Technology
16
Resonance is energy storage,
damping is velocity related
• Driving force is in phase with velocity at resonance
• Power is Force times speed
• Max energy transfer at resonance
• Damper is just the opposite
• Q tells something about the height of the resonance peak
Maximum energy stored

Q  2
Energy lost per cycle
33
Delft
University of
Technology
Q and energy
34
Delft
University of
Technology
17
Contents
35
Delft
University of
Technology
Transmissibility, transfer of motion through

the support of a dynamic system
The force acting on the body equals
d 2 xm d(x  xm ) m xm
Ft,m (t )  m 2
c f  k ( xf  xm )
dt dt
c k
Table xf
xm (ms  cs  k )  xf (cs  k )
2
cs
1
xm cs  k k
 2 
x f ms  cs  k m s 2  cs  1
k k
s
2 1
k c xm 0
With: 0     2
m 2 km xf s s
 2 1
 2
0 0
36
Delft
University of
Technology
18
Bode plot of transmissibility
37
Delft
University of
Technology
Contents
38
Delft
University of
Technology
19
Coupled mass-spring systems
x1 x2
k
F m1 m2
Elastically coupled Multi-body or finite-element non-rigid body dynamics
39
Delft
University of
Technology
Which statement is true? There is a

frequency where:
x1 x2
k
F m1 m2
1. Only m1 will resonate with k
2. Only m2 will resonate with k
3. Only m1 will stop moving
4. Only m2 will stop moving
40
Delft
University of
Technology
20
Equations of Motion
• First body
x1 x2 d 2 x1 (t )
m1  F (t )  k ( x1  x2 )
dt 2
k
F m1 m2 m1s 2 x1 ( s)  F  k ( x1  x2 )
Stationary reference • Second body
d 2 x2 (t )
m2  k ( x1  x2 )
dt 2
m2 s 2 x2 ( s )  k ( x1  x2 )
note: The domain (s or t) is only mentioned once!

41
Delft
University of
Technology
Resulting equations in the Laplace

domain
x1 m2 s 2  k
(s) 
F m1m2 s 4  k  m1  m2  s 2
Fourth order system
x2 k
( s) 
F m1m2 s 4  k  m1  m2  s 2
Low values of s at High values of s at

low frequencies: high frequencies:
x1 x 1 x1 1 x2 k
(s)  2 (s)  ( s)  , (s) 
F F  m1  m2  s 2 F m1s 2 F m1m2 s 4
42
Delft
University of
Technology
21
At mid frequency the driven body
shows a strange effect
x1 m2 s 2  k
 Combination
F m1m2 s 4  k  m1  m2  s 2 of poles and zeros
x2 k
(s) 
F m1m2 s 4  k  m1  m2  s 2
when
k
s 2   2  
m2
then the compliance of x1 shows a dip
43
Delft
University of
Technology
m1 = 0.1 m2 , an actuated large mass

by a lighter actuator.
x1 x 1
LF: ( s)  2 (s) 
F F  m1  m2  s 2
x1 m2 s 2  k
(s) 
F m1m2 s  k  m1  m2  s 2
4
x2 k
(s) 
F m1m2 s  k  m1  m2  s 2
4
x1 1 x2 k
HF: (s)  , ( s) 
F m1s 2 F m1m2 s 4
m1  2.5 103 kg
m2  25 103 kg
k  10 4 N/m
44
Delft
University of
Technology
22
m1 = m2 , motor mass is optimised
to the driven mass.
x1 m2 s 2  k
(s) 
F m1m2 s 4  k  m1  m2  s 2
x2 k
(s) 
F m1m2 s 4  k  m1  m2  s 2
m1  2.5 103 kg
m2  2.5 103 kg
k  10 4
N/m
45
Delft
University of
Technology
m1 = 10 m2 , parasitic resonances
x1 m2 s 2  k
(s) 
F m1m2 s  k  m1  m2  s 2
4
x2 k
(s) 
F m1m2 s  k  m1  m2  s 2
4
m1  0.25 kg
3
m2  25 10 kg
k  10 4
N/m
46
Delft
University of
Technology
23
Contents
• Dynamic modelling of damped mass spring systems.
• Coupled mass spring systems
47
Delft
University of
Technology
Eigenmodes
When excited, non‐rigid bodies and complex mass‐spring systems

vibrate in different eigenmodes with two main properties:
• Eigenfrequency, the related natural (resonance) frequency.
• Mode‐shape, the deformation that corresponds with the
eigenmode, described in a multiple degree of freedom “shape‐
function ”.
Modelling is done by discretisation of the system in multiple mass‐

spring systems.
• The shape function reduces to an eigenvector with one value for

each body for the relative motion magnitude and direction (sign).
48
Delft
University of
Technology
24
Mode-shape of complex non-rigid
body
49
Delft
University of
Technology
Eigenfrequencies of multiple
eigenmodes in 6-DOF
50
Delft
University of
Technology
25
Eigen modes of coupled bodies
1
1   
1
1
2   
 1
51
Delft
University of
Technology
Multiplicative expression shows two

resonating eigenfrequencies
Starting with:
x1 m2 s 2  k x2 k
 
F m1m2 s 4  k  m1  m2  s 2 F m1m2 s  k  m1  m2  s 2
4
x1 1 1

F  m1  m2  s 2
m2 s 2  k 
mc s 2  k
 0
1  0
m1m2 m1  m2
with: mc 
m1  m2 k
2 
x2 1 k M

F  m1  m2  s mc s 2  k
2
52
Delft
University of
Technology
26
Additive expression:
x1 1 1 m1m2

F  m1  m2  s 2
m2 s 2  k 
mc s 2  k
 with: mc 
m1  m2
x2 1 k

F  m1  m2  s 2 mc s 2  k
• Can be written as a combination of two eigenmodes with modal

i
mass i and stiffness i using eigenvector :
m22 m22
x1 1 m  m  1 m  m 
2 2 1
  1 2 2 =  1 2 2 1   
F (m1  m2 ) s 2
mc s  k 1s  1 2 s  2
2 1

m1m2 m1m2
 
x2 1  m1  m2 
2
1  m1  m2 
2
1
  =  2   
F (m1  m2 ) s 2 mc s 2  k 1s 2  1 2 s 2  2  1
53
Delft
University of
Technology
Modes 1 and 2 combined
m22
m  m 
2
x1 1
  1 2 2
F 1s  1 2 s  2
2
m1m2

 m1  m2 
2
x2 1
 
F 1s  1
2
2 s  2
2
54
Delft
University of
Technology
27
Theory on modal decomposition
• General equation of motion (vector/matrix) of finite element system:
Mx(t )  Kx (t )  F (t )
• General transfer function:
Ms2 x ( s)  Kx ( s)  F ( s)
• In absence of external force:
Ms2 x ( s)  Kx ( s)  0
i
• Decoupled by eigenvalue problem, where is the eigenvector:
 K  0,2 i M  i  0
55
Delft
University of
Technology
Modal mass and stiffness,

orthogonality of eigenmodes.
• Modal mass:
i M  j  0 (i  j )
i M i  i
• Modal Stiffness:
i K  j  0 (i  j )
i K i  0,2 i i  i
56
Delft
University of
Technology
28
Scaling
• Length of eigenvector is not defined (only the direction)
i M i  i
• Three scaling methods are often applied:
1: i  1
2: max i   1
3 : i  1
57
Delft
University of
Technology
Modal coordinates
• When qi equals the motion of eigenmode i then the total

displacement vector x(t) will be:
x (t )  q1 (t )1  q2 (t )2  ...qn (t )n

where qi is called the modal coordinate
• This gives the following displacement for DOF xk(t)
x k (t )  q1 (t )1,k  q2 (t )2,k  ...qn (t )n,k
58
Delft
University of
Technology
29
Full set of uncoupled equations
x (t )  q(t )
 q1 (t ) 
  
 
with :   1 2 ... n  and q(t )   qi (t ) 
 
  
qn (t ) 
2
applied to : Ms x ( s)  Kx ( s)  F ( s)
gives with pre-multiplication with   :
  M q(t )    K q(t )    F (t )
59
Delft
University of
Technology
This can be transformed into a

combined transfer function:
x
• The response of DOF by
Fk
eigenmode i on force equals:
 x  i ,ki ,
  ( s) 
 Fk i i s2  i
• The total response equals:
x n
x  n  
( s)        i ,k2 i ,
Fk i 1  Fk i i 1 i s  i
60
Delft
University of
Technology
30
Graphical representation
• The modal coordinate is
represented by the angle of the
lever or the displacement at
distance 1.
• Modal mass and stiffness are
connected to the modal
coordinate
 x  i ,ki ,
• The eigenvectors determine the   ( s) 
ratio of “controllability” of  Fk i i s2  i
actuator and “observability” of
sensor.
61
Delft
University of
Technology
Effective modal values
As perceived at the actuator
62
Delft
University of
Technology
31
Mass on leafspring, two eigenmodes
F F
x x
Stationary reference Stationary reference
a: Mode shape 1 b: Mode shape 2
63
Delft
University of
Technology
Modes 1 and 2 combined by measuring

non co-local with the force!
‐2 slope with ‐360o
phase is non‐minimum
phase system!
64
Delft
University of
Technology
32
Non-minimum phase step response
65
Delft
University of
Technology
Summary of terms
• Eigenvector:
• Vector with terms that represent the relative motion amplitude of each element in the
discretised structure. It is the discretised “shape function”.
• Scaling is at will, as long as the scaling is equal for all properties of that eigenmode.
• Modal mass and stiffness: i M i  i i K i  i
• relate to the mass and stiffness matrix by double multiplication with the eigenvector.
• Scaling of the eigenvector also scales the modal mass and stiffness (squared).
• Modal coordinates:
• Represent the motion of the eigenmode at the location of the modal mass and
stiffness.
• It is used to determine the motion of all elements with the eigenvector and as such it
also depends on the chosen scaling.
• Effective modal mass and stiffness:
• Is the modal mass and stiffness as perceived at the actuator position. It is calculated
from the modal mass and stiffness by dividing by the corresponding eigenvector term
squared, hence the chosen scaling no longer plays a role.
66
Delft
University of
Technology
33
Mode-shape of complex non-rigid
body
67
Delft
University of
Technology
Contents
68
Delft
University of
Technology
34
Typical combination of two
eigenmodes, one rigid and one non-
rigid.
•Actuated at xa.
•Sensed at different
locations.
69
Delft
University of
Technology
Different combinations of two

eigenmodes
70
Delft
University of
Technology
35
4 types of responses result
71
Delft
University of
Technology
Eigenfrequencies of multiple
eigenmodes in 6-DOF
72
Delft
University of
Technology
36

Dynamics of Motion Systems

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dynamics of Motion Systems

Uploaded by

Copyright:

Available Formats

Mechatronic system design

Prof.ir. R.H.Munnig Schmidt

Well known objects

The first natural frequency determines

Stiffness and compliance

• k = stiffness of the spring

Stiffness and compliance of a damper

Combined Compliance of body,

The separate element responses in a

Extreme damping, two first order

Start with second law of Newton

And phase angle φ is – arctan (imaginary/real of the denominator)

Compliance Bode plot with damping

p1    jd and p2    jd

Critical damping ratio   1

p1    jd and p2    jd

Effect of the damping ratio  on

Energy will be “trapped” in the system.

For electrical engineering (frequency domain) the quality

Let’s bridge the gap

Q=100 means peak level at 100 times

Velocity response, kinetic energy

Maximum energy stored

Transmissibility, transfer of motion through

Elastically coupled Multi-body or finite-element non-rigid body dynamics

Which statement is true? There is a

note: The domain (s or t) is only mentioned once!

Resulting equations in the Laplace

Low values of s at High values of s at

m1 = 0.1 m2 , an actuated large mass

When excited, non‐rigid bodies and complex mass‐spring systems

Modelling is done by discretisation of the system in multiple mass‐

• The shape function reduces to an eigenvector with one value for

Multiplicative expression shows two

• Can be written as a combination of two eigenmodes with modal

Modes 1 and 2 combined

Modal mass and stiffness,

• When qi equals the motion of eigenmode i then the total

x (t )  q1 (t )1  q2 (t )2  ...qn (t )n

x k (t )  q1 (t )1,k  q2 (t )2,k  ...qn (t )n,k

This can be transformed into a

Effective modal values

Modes 1 and 2 combined by measuring

Different combinations of two

You might also like