Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why...

21
MIT OpenCourseWare http://ocw.mit.edu 2.72 Elements of Mechanical Design Spring 2009 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms .

Transcript of Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why...

Page 1: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

MIT OpenCourseWare http://ocw.mit.edu

2.72 Elements of Mechanical Design Spring 2009

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Page 2: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

2.72

Elements of

Mechanical Design

Lecture 16: Dynamics

and damping

Page 3: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Schedule and reading assignment Quiz

� None

Topics � Vibration physics

� Connection to real world

� Activity

Reading assignment � Skim last gear reading assignment… (gear selection)

© Martin Culpepper, All rights reserved 2

Page 4: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Resonance Basic Physics

� Exchange potential-kinetic energy

� Energy transfer with loss

Modeling

� 2nd order system model

� Spring mass damper

Transforms

Why Do We Care

� Critical to understanding motion of

structures- desired and undesired

� Generally not steady state

� Location error

� Large forces, high fatigue ω

Gain

ωn

mk

c x

F(t)

m

k n =ω

KE PE

Loss

Differential equations, Laplace

© Martin Culpepper, All rights reserved 3

Page 5: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Vibrations - Input Oscillation of System

A Why Categorize

� Different causes

� Different solutions

Input Form

� Forced – Steady State

• Command Signal

• Electrical (60 Hz) A � Free – Transient

• Impacts

t

Forced

t

Free

© Martin Culpepper, All rights reserved 4

Page 6: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Vibrations - Source Source

10­9 Undriven � Disturbance

10­13 • Driven Disturbance

10­17 Device Motors Spectrum Rotating Components

10­2 1 102 104

• Undriven

Electrical (60 Hz)

People (2 Hz)

Cars (10 Hz)

System

Driven Disturbance Nearby Equipment

� Command

• Step Response

A

t Vibration Command Response

© Martin Culpepper, All rights reserved 5

Page 7: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Vibrations: Command vs. Disturbance

Underdamped response

Noise

Free vibration

Golda, D. S., “Design of High-Speed, Meso-Scale Nanopositioners

Driven by Electromagnetic Actuators,” Ph.D. Thesis, Massachusetts

Institute of Technology, 2008.

© Martin Culpepper, All rights reserved 6

Page 8: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Vibrations - Example Example

A Forced A � Chinook

� Identify:

• Mode

t• Form

Free

t

• Source

• Response

10­9

10­13 Spectrum

10­17

10410­2 1021

Undriven Disturbance

t Response

System

A

Driven Disturbance

© Martin Culpepper, All rights reserved 7

Page 9: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Attenuating Vibrations Change System

� Mass, stiffness, damping

� Adjust mode shapes

Change Inputs

� Command: Input Modulation, Feedback

� Disturbance, reduce:

• undriven vibrations, e.g. optical table

• driven vibrations – alter device structure (damping on motors, etc.)

1

ms2+bs+k

ZF

t

A

© Martin Culpepper, All rights reserved 8

Page 10: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Change Input shaping

m, k, c

9© Martin Culpepper, All rights reserved

Modulating Command Vibrations

Page 11: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Regimes � (1) Low Frequency (ω<ω n)

� (2) Resonance (ω≈ω n)

� (3) High Frequency (ω>ω n)

� Example – Spring/Mass demo

Behavior

x 1 1= =

F ms 2 + bs + k k(ω)

© Martin Culpepper, All rights reserved 10

ω/ωn

-1-1

1

1

10

10

Am

plifi

catio

n fa

ctor

(out

put/i

nput

)

1 2 3

1st order system

ζ = 0.05

ζ = 0.1

ζ = 0.2

ζ = 0.3

ζ = 0.4

ζ = 0.5ζ = 0.6ζ = 0.7

ζ = 0.8ζ = 0.9

ζ = 1.0

ζ = 2.0

Figure by MIT OpenCourseWare.

Page 12: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Behavior – Low Frequency Frequency Response Regimes

� (1) Low Frequency (ω<ωn)

� (2) Resonance (ω≈ωn)

� (3) High Frequency (ω>ωn)

� Example – Spring/Mass demo

Low Frequency (ω<ω n) � System tracks commands

� Ideal operating range

� High disturbance rejection

1x

F k ≈

Resonance Frequency (ω≈ω n) � System Response >> command

� keff ↓ � Disturbances will cause very large response

� Quality factor = magnitude of peak x 1 � Damping ↑ = Q ↓

F =

2 ms bs k+ +

High Frequency (ω≈ω n) � System Response << command

� High disturbance rejection

© Martin Culpepper, All rights reserved 11

ω/ωn

-1-1

1

1

10

10

Am

plifi

catio

n fa

ctor

(out

put/i

nput

)

1 2 3

1st order system

ζ = 0.05

ζ = 0.1

ζ = 0.2

ζ = 0.3

ζ = 0.4

ζ = 0.5ζ = 0.6ζ = 0.7

ζ = 0.8ζ = 0.9

ζ = 1.0

ζ = 2.0

Figure by MIT OpenCourseWare.

Page 13: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Behavior - Resonance

Frequency Response Regimes � (1) Low Frequency (ω<ωn)

� (2) Resonance (ω≈ωn)

� (3) High Frequency (ω>ωn)

� Example – Spring/Mass demo

Low Frequency (ω<ω n) � System tracks commands

� Ideal operating range

� High disturbance rejection

1

x Q < ≤Resonance Frequency (ω≈ω n)

k F k� System Response >> command

� keff ↓ � Disturbances will cause very large response

� Quality factor = magnitude of peak x 1 = � Damping ↑ = Q ↓

2F ms + bs + k

High Frequency (ω≈ω n) � System Response << command

� High disturbance rejection

© Martin Culpepper, All rights reserved 12

ω/ωn

-1-1

1

1

10

10

Am

plifi

catio

n fa

ctor

(out

put/i

nput

)

1 2 3

1st order system

ζ = 0.05

ζ = 0.1

ζ = 0.2

ζ = 0.3

ζ = 0.4

ζ = 0.5ζ = 0.6ζ = 0.7

ζ = 0.8ζ = 0.9

ζ = 1.0

ζ = 2.0

Figure by MIT OpenCourseWare.

Page 14: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Behavior – High Frequency Frequency Response Regimes

� (1) Low Frequency (ω<ωn)

� (2) Resonance (ω≈ωn)

� (3) High Frequency (ω>ωn)

� Example – Spring/Mass demo

Low Frequency (ω<ω n) � System tracks commands

� Ideal operating range

� High disturbance rejection

Resonance Frequency (ω≈ω n) � System Response >> command

� keff ↓ � Disturbances will cause very large response

� Quality factor = magnitude of peak x 1 � Damping ↑ = Q ↓

F =

2 ms bs k+ +

x 1 High Frequency (ω≈ω n)

� System Response << command F ≈

2 mω

� Poor disturbance rejection

© Martin Culpepper, All rights reserved 13

ω/ωn

-1-1

1

1

10

10

Am

plifi

catio

n fa

ctor

(out

put/i

nput

)

1 2 3

1st order system

ζ = 0.05

ζ = 0.1

ζ = 0.2

ζ = 0.3

ζ = 0.4

ζ = 0.5ζ = 0.6ζ = 0.7

ζ = 0.8ζ = 0.9

ζ = 1.0

ζ = 2.0

Figure by MIT OpenCourseWare.

Page 15: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Constitutive Relations Relevant equations

c � Damping ratio ζ =

2 km

k 2 � ω n ω = ωd = ωn 1− ζ

n m

1 1 � Gain G = ζ <P

2ζ 1−ζ 2

Q = 2ζ (1−ζ 2 )� Quality factor

© Martin Culpepper, All rights reserved 14

1

2

Page 16: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Application of Theory Relate Variables to Actual Parameters

� Vibrational Mode

� Mass

� Stiffness mk

c x

F(t) � Damping

Transfer between Model and Reality Iteration � Iterative

� Start simple (1 mass, 1 spring)

Mode? � Add complexity

� Limits k?

Example – building (video) m?

c?

© Martin Culpepper, All rights reserved 15

Page 17: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Strategies for damping Material Pros and cons of each

� Grain boundary

� Internal lattice

� Viscoelastic (elastomers/goo)

Viscous � Air

� Fluid

Electromagnetic

Friction

Active

Combinations � Sponge

© Martin Culpepper, All rights reserved 16

Page 18: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Example: Couette flow relationships

Relevant equations

dx� τ = µ

dy

dx� x� (F , x�)F = τA = Aµ = Aµ

dy

F = cx� µ

Aµc =

h

© Martin Culpepper, All rights reserved 17

h

h

A

Page 19: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Exercise (see next page too)

Perform a frequency analysis of the part� Develop & prove (FEA) how to increase nat. freq. via geometry change

� Any constraints you might have? Geometry changes can’t be unbounded

� Explain effect of your change on vibration amplitude (relative to outer base) at given ω, via sketches & plots

Xtra credit, assume: � Flexure is contained between two parallel plates (on top and bottom)

� Viscous air damping in the gaps on both sides

� 1 micron gap between the flexure sides and plates

� Elaborate on how well flexure is damped (don’t just use intuition)

Useful equations (c = damping coefficient, k = stiffness, m = mass)

Gain at peak amplitude Frequency at peak with max gain Damping ratio

1c G =ζ = ω p = ω n 1− 2ζ 2 p 4ζ 2 − 4ζ 4

2 km © Martin Culpepper, All rights reserved 18

Page 20: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Flexure Flexure design

Constrained on 4 sides top

Bottom

See this diagram for extra credit:

19

Top plate

Flexure

Bottom plate © Martin Culpepper, All rights reserved

Page 21: Spring 2009 For information about citing these …...Vibrations - Input Oscillation of System A Why Categorize Different causes Different solutions Input Form Forced – Steady State

Multiple ResonancesP

has

e L

ag (

deg

) G

ain

Z-Axis2 Center Stage Modes 10

0 10

Measured

-2 Model10

Fit

1 2 3 10 10 10

Frequency (Hz)

0

-100

-200

-300

1 2 3 10 10 10

Frequency (Hz)© Martin Culpepper, All rights reserved 20