Professional Documents
Culture Documents
Dynamic System Modeling and Contol Hugh Jack PDF
Dynamic System Modeling and Contol Hugh Jack PDF
page 1
by
Hugh Jack
page 1
PREFACE
How to use the book.
read the chapters and do drill problems as you read
examine the case studies - these pull together concepts from previous chapters
problems at the ends of chapters are provided for further practice
Tools that should be used include,
graphing calculator that can solve differential equations, such as a TI-85
computer algebra software that can solve differential equations, such as Scilab
Supplemental materials at the end of this book include,
a writing guide
a summary of math topics important for engineers
a table of generally useful engineering units
properties of common materials
Acknowledgement to,
Dr. Hal Larson for reviewing the calculus and numerical methods chapters
Dr. Wendy Reffeor for reviewing the translation chapter
Student background
a basic circuits course
a basic statics and mechanics of materials course
math up to differential equations
a general knowledge of physics
computer programming, preferably in C
Special notes
- despite all common wisdom, inertia is presented as a force, this makes it easier
for students attempting to learn, and keep sign conventions correct
TO BE DONE
small
italicize variables and important terms
fix equation numbering (auto-numbering?)
fix subscripts and superscripts
fix problem forms to include therefores, mark FBDs, etc.
check C programs for ANSI compliance
big
verify the phase angle relationships cos vs/ sin.
page 2
chapter rotation
replace the rotational case with IC motor
chapter non-linear systems
develop chapter
chapter magnetic
consider adding/writing this chapter
chapter fluids
consider adding/writing this chapter
chapter thermal
consider adding/writing this chapter
chapter c programming
review section
add problems
Buzzword topics:
- Distributed systems
- Intelligent manufacturing systems
- Adaptive control
- Architectures for signal processing and control algorithms
- Discrete event systems
- Hybrid systems
- Predictive control
- Robust control
page 1
1.
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1
1.1
1.2
1.3
1.4
2.
2.3
2.4
2.5
2.6
2.7
2.8
INTRODUCTION
MODELING
2.2.1
Free Body Diagrams
2.2.2
Mass and Inertia
2.2.3
Gravity and Other Fields
2.2.4
Springs
2.2.5
Damping and Drag
2.2.6
Cables And Pulleys
2.2.7
Friction
2.2.8
Contact Points And Joints
SYSTEM EXAMPLES
OTHER TOPICS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
2.1
2.3
2.4
2.4
2.8
2.10
2.18
2.21
2.23
2.25
2.25
2.35
2.36
2.36
2.41
2.45
3.4
3.5
3.6
3.7
3.8
3.9
3.10
4.
1.1
1.3
1.3
1.3
TRANSLATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1
2.1
2.2
3.
BASIC TERMINOLOGY
EXAMPLE SYSTEM
SUMMARY
PRACTICE PROBLEMS
INTRODUCTION
EXPLICIT SOLUTIONS
RESPONSES
3.3.1
First-order
3.3.2
Second-order
3.3.3
Other Responses
RESPONSE ANALYSIS
NON-LINEAR SYSTEMS
3.5.1
Non-Linear Differential Equations
3.5.2
Non-Linear Equation Terms
3.5.3
Changing Systems
CASE STUDY
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
3.1
3.2
3.16
3.17
3.23
3.28
3.31
3.33
3.34
3.38
3.41
3.47
3.51
3.51
3.56
3.61
page 2
4.1
4.2
4.3
4.4
DATA
4.5
4.31
4.6
4.7
4.8
4.9
4.10
4.11
5.
ADVANCED TOPICS
4.6.1
Switching Functions
4.6.2
Interpolating Tabular Data
4.6.3
Modeling Functions with Splines
4.6.4
Non-Linear Elements
CASE STUDY
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
4.33
4.33
4.36
4.37
4.39
4.39
4.46
4.47
4.50
4.60
ROTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1
5.1
5.2
5.3
5.4
5.5
5.6
5.7
5.8
6.
INTRODUCTION
4.1
THE GENERAL METHOD
4.1
4.2.1
State Variable Form
4.2
NUMERICAL INTEGRATION
4.10
4.3.1
Numerical Integration With Tools
4.10
4.3.2
Numerical Integration
4.15
4.3.3
Taylor Series
4.21
4.3.4
Runge-Kutta Integration
4.23
SYSTEM RESPONSE
4.29
4.4.1
Steady-State Response
4.30
DIFFERENTIATION AND INTEGRATION OF EXPERIMENTAL
INTRODUCTION
MODELING
5.2.1
Inertia
5.2.2
Springs
5.2.3
Damping
5.2.4
Levers
5.2.5
Gears and Belts
5.2.6
Friction
5.2.7
Permanent Magnet Electric Motors
OTHER TOPICS
DESIGN CASE
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
5.1
5.2
5.3
5.7
5.12
5.14
5.15
5.19
5.22
5.23
5.23
5.28
5.28
5.35
5.44
INTRODUCTION
THE DIFFERENTIAL OPERATOR
INPUT-OUTPUT EQUATIONS
6.1
6.1
6.4
page 3
6.4
6.5
6.6
6.7
6.8
6.9
7.
6.6
6.9
6.11
6.20
6.20
6.22
6.26
6.27
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
7.11
8.
6.3.1
Converting Input-Output Equations to State Equations
6.3.2
Integrating Input-Output Equations
DESIGN CASE
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSGINMENT PROBLEMS
REFERENCES
INTRODUCTION
MODELING
7.2.1
Resistors
7.2.2
Voltage and Current Sources
7.2.3
Capacitors
7.2.4
Inductors
7.2.5
Op-Amps
IMPEDANCE
EXAMPLE SYSTEMS
ELECTROMECHANICAL SYSTEMS - MOTORS
7.5.1
Permanent Magnet DC Motors
7.5.2
Induction Motors
7.5.3
Brushless Servo Motors
FILTERS
OTHER TOPICS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
7.1
7.1
7.2
7.4
7.8
7.10
7.11
7.16
7.18
7.26
7.26
7.28
7.29
7.32
7.33
7.33
7.34
7.38
7.43
8.4
8.5
INTRODUCTION
TRANSFER FUNCTIONS
CONTROL SYSTEMS
8.3.1
PID Control Systems
8.3.2
Manipulating Block Diagrams
8.3.3
A Motor Control System Example
8.3.4
System Error
8.3.5
Controller Transfer Functions
8.3.6
Feedforward Controllers
8.3.7
State Equation Based Systems
8.3.8
Cascade Controllers
SUMMARY
PRACTICE PROBLEMS
8.1
8.1
8.3
8.5
8.7
8.12
8.17
8.21
8.21
8.22
8.24
8.24
8.24
page 4
8.6
8.7
9.
9.1
9.1
9.8
9.10
9.10
9.12
9.14
INTRODUCTION
BODE PLOTS
SIGNAL SPECTRUMS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
LOG SCALE GRAPH PAPER
10.1
10.5
10.21
10.22
10.22
10.25
10.35
10.36
12.
INTRODUCTION
PHASORS FOR STEADY-STATE ANALYSIS
VIBRATIONS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
11.
8.31
8.37
10.
INTRODUCTION
ROOT-LOCUS ANALYSIS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
11.1
11.1
11.10
11.11
11.14
11.25
12.4
12.5
INTRODUCTION
SOURCES OF NONLINEARITY
12.2.1
Non-Linear Relationships
NON-LINEAR ELEMENTS
12.3.1
Time Variant
12.3.2
Switching
12.3.3
Deadband
12.3.4
Saturation and Clipping
12.3.5
Hysteresis and Slip
12.3.6
Delays and Lags
SUMMARY
PRACTICE PROBLEMS
12.1
12.1
12.1
12.2
12.3
12.3
12.4
12.7
12.7
12.8
12.9
12.9
page 5
12.6
12.7
13.
12.9
12.9
13.5
13.6
13.7
13.8
13.9
14.
INTRODUCTION
ANALOG INPUTS
ANALOG OUTPUTS
NOISE REDUCTION
13.4.1
Shielding
13.4.2
Grounding
CASE STUDY
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
13.1
13.3
13.10
13.12
13.12
13.14
13.15
13.15
13.15
13.15
13.16
INTRODUCTION
14.1
INDUSTRIAL SENSORS
14.2
14.2.1
Angular Displacement
14.3
Potentiometers
14.3
14.2.2
Encoders
14.4
Tachometers
14.8
14.2.3
Linear Position
14.8
Potentiometers
14.8
Linear Variable Differential Transformers (LVDT)14.9
Moire Fringes
14.11
Accelerometers
14.12
14.2.4
Forces and Moments
14.15
Strain Gages
14.15
Piezoelectric
14.18
14.2.5
Liquids and Gases
14.20
Pressure
14.21
Venturi Valves
14.22
Coriolis Flow Meter
14.23
Magnetic Flow Meter
14.24
Ultrasonic Flow Meter
14.24
Vortex Flow Meter
14.24
Positive Displacement Meters
14.25
Pitot Tubes
14.25
14.2.6
Temperature
14.25
Resistive Temperature Detectors (RTDs)
14.26
Thermocouples
14.26
Thermistors
14.28
Other Sensors
14.30
page 6
14.2.7
14.3
14.4
14.5
14.6
14.7
14.8
14.9
15.
14.30
14.30
14.31
14.31
14.31
14.32
14.32
14.37
14.38
14.39
14.39
14.40
14.42
15.3
15.4
15.5
15.6
15.7
15.8
16.
Light
INTRODUCTION
ELECTRIC MOTORS
15.2.1
Basic Brushed DC Motors
15.2.2
AC Motors
15.2.3
Brushless DC Motors
15.2.4
Stepper Motors
15.2.5
Wound Field Motors
HYDRAULICS
OTHER SYSTEMS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
15.1
15.1
15.3
15.7
15.15
15.17
15.19
15.23
15.24
15.25
15.25
15.26
15.26
16.3
16.4
16.5
16.6
16.7
16.8
16.9
INTRODUCTION
MOTION PROFILES
16.2.1
Velocity Profiles
16.2.2
Position Profiles
MULTI AXIS MOTION
16.3.1
Slew Motion
Interpolated Motion
16.3.2
Motion Scheduling
PATH PLANNING
CASE STUDIES
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
16.1
16.2
16.2
16.11
16.14
16.15
16.16
16.17
16.19
16.21
16.23
16.23
16.24
16.25
page 7
17.
17.7
17.8
17.9
17.10
17.11
17.12
17.13
18.
INTRODUCTION
APPLYING LAPLACE TRANSFORMS
17.2.1
A Few Transform Tables
MODELING TRANSFER FUNCTIONS IN THE s-DOMAIN
FINDING OUTPUT EQUATIONS
INVERSE TRANSFORMS AND PARTIAL FRACTIONS
EXAMPLES
17.6.1
Mass-Spring-Damper Vibration
17.6.2
Circuits
ADVANCED TOPICS
17.7.1
Input Functions
17.7.2
Initial and Final Value Theorems
A MAP OF TECHNIQUES FOR LAPLACE ANALYSIS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
REFERENCES
17.1
17.3
17.4
17.9
17.11
17.14
17.21
17.21
17.23
17.25
17.25
17.26
17.27
17.28
17.28
17.31
17.35
17.37
INTRODUCTION
18.1
CONTROL SYSTEMS
18.1
18.2.1
PID Control Systems
18.3
18.2.2
Analysis of PID Controlled Systems With Laplace Transforms
18.5
18.3
18.4
18.5
18.6
18.7
18.8
19.
18.2.3
Finding The System Response To An Input
18.2.4
Controller Transfer Functions
ROOT-LOCUS PLOTS
18.3.1
Approximate Plotting Techniques
DESIGN OF CONTINUOUS CONTROLLERS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
18.8
18.13
18.13
18.17
18.21
18.21
18.22
18.27
18.27
CONVOLUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.1
19.1
19.2
19.3
19.4
19.5
19.6
INTRODUCTION
UNIT IMPULSE FUNCTIONS
IMPULSE RESPONSE
CONVOLUTION
NUMERICAL CONVOLUTION
LAPLACE IMPULSE FUNCTIONS
19.1
19.1
19.3
19.5
19.6
19.9
page 8
19.7
19.8
19.9
19.10
20.
21.9
21.10
21.11
21.12
21.13
21.14
20.1
20.13
20.15
20.18
20.18
20.19
20.19
20.19
20.19
INTRODUCTION
21.1
FULL STATE FEEDBACK
21.2
OBSERVERS
21.5
SUPPLEMENTAL OBSERVERS
21.11
REGULATED CONTROL WITH OBSERVERS
21.11
LQR
21.22
LINEAR QUADRATIC GAUSSIAN (LQG) COMPENSATORS 21.24
VERIFYING CONTROL SYSTEM STABILITY
21.24
21.8.1
Stability
21.25
21.8.2
Bounded Gain
21.26
ADAPTIVE CONTROLLERS
21.28
OTHER METHODS
21.31
21.10.1
Kalman Filtering
21.32
SUMMARY
21.32
PRACTICE PROBLEMS
21.33
PRACTICE PROBLEM SOLUTIONS
21.33
ASSIGNMENT PROBLEMS
21.33
23.
INTRODUCTION
OBSERVABILITY
CONTROLLABILITY
OBSERVERS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
BIBLIOGRAPHY
22.
19.10
19.10
19.10
19.10
21.
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
INTRODUCTION
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
22.1
22.10
22.10
22.10
22.10
page 9
23.1
23.2
23.3
23.4
23.5
23.6
23.7
24.
24.3
24.4
24.5
24.6
24.7
SUMMARY
MATHEMATICAL PROPERTIES
24.2.1
Resistance
24.2.2
Capacitance
24.2.3
Power Sources
EXAMPLE SYSTEMS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEMS SOLUTIONS
ASSIGNMENT PROBLEMS
24.1
24.1
24.2
24.4
24.6
24.8
24.10
24.10
24.10
24.10
25.3
25.4
25.5
25.6
25.7
26.
23.1
23.1
23.1
23.9
23.16
23.16
23.16
23.16
25.
INTRODUCTION
MATHEMATICAL PROPERTIES
23.2.1
Induction
EXAMPLE SYSTEMS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
INTRODUCTION
MATHEMATICAL PROPERTIES
25.2.1
Resistance
25.2.2
Capacitance
25.2.3
Sources
EXAMPLE SYSTEMS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
25.1
25.1
25.1
25.3
25.4
25.4
25.7
25.7
25.7
25.7
OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26.1
26.1
26.2
26.3
26.4
26.5
26.6
26.7
INTRODUCTION
OBJECTIVES AND CONSTRAINTS
SEARCHING FOR THE OPTIMUM
OPTIMIZATION ALGORITHMS
26.4.1
Random Walk
26.4.2
Gradient Decent
26.4.3
Simplex
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
26.1
26.2
26.6
26.9
26.9
26.10
26.10
26.10
26.10
26.10
page 10
26.8
27.
INTRODUCTION
COMMERCIAL CONTROLLERS
REFERENCES
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
28.1
28.7
28.7
28.7
28.8
28.8
28.8
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
REFERENCES
29.9
29.10
29.10
29.10
29.10
31.
27.1
27.2
27.4
27.12
27.13
27.13
27.13
27.13
30.
INTRODUCTION
FINITE ELEMENT MODELS
FINITE ELEMENT MODELS
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
BIBLIOGRAPHY
29.
26.10
28.
ASSIGNMENT PROBLEMS
INTRODUCTION
CASE STUDY
SUMMARY
PRACTICE PROBLEMS
PRACTICE PROBLEM SOLUTIONS
ASSIGNMENT PROBLEMS
30.1
30.3
30.3
30.3
30.3
30.3
WRITING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.1
31.1
31.2
31.3
31.4
31.5
31.1
31.2
31.3
31.3
31.3
31.5
page 11
31.6
31.7
31.8
31.9
31.10
31.11
31.12
31.13
31.14
31.15
31.16
32.
31.11
31.11
31.11
31.12
31.13
31.13
31.13
31.13
31.14
31.14
31.14
31.15
31.15
31.16
31.16
31.17
31.18
31.19
31.19
31.20
31.21
31.21
31.21
31.22
31.22
31.23
31.23
31.23
31.24
31.24
31.24
31.25
31.25
31.25
31.26
31.26
31.29
31.31
PROJECTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1
32.1
32.2
32.3
OVERVIEW
32.2.1
The Objectives and Constraints
MANAGEMENT
32.3.1
Timeline - Tentative
32.1
32.1
32.2
32.3
32.3
page 12
32.4
32.5
32.6
33.
32.4
32.5
32.5
32.5
32.6
32.6
32.7
32.8
32.8
32.9
32.9
32.10
32.10
32.10
32.18
32.19
33.4
33.5
34.
32.3.2
Teams
DELIVERABLES
32.4.1
Conceptual Design
32.4.2
EGR 345/101 Contract
32.4.3
Progress Reports
32.4.4
Design Proposal
32.4.5
The Final Report
REPORT ELEMENTS
32.5.1
Gantt Charts
32.5.2
Drawings
32.5.3
Budgets and Bills of Material
32.5.4
Calculations
APPENDICES
32.6.1
Appendix A - Sample System
32.6.2
Appendix B - EGR 345/101 Contract
32.6.3
Appendix C - Forms
33.1
33.1
33.2
33.2
33.2
33.2
33.3
34.2
34.3
INTRODUCTION
34.1.1
Constants and Other Stuff
34.1.2
Basic Operations
Factorial
34.1.3
Exponents and Logarithms
34.1.4
Polynomial Expansions
34.1.5
Practice Problems
FUNCTIONS
34.2.1
Discrete and Continuous Probability Distributions
34.2.2
Basic Polynomials
34.2.3
Partial Fractions
34.2.4
Summation and Series
34.2.5
Practice Problems
SPATIAL RELATIONSHIPS
34.3.1
Trigonometry
34.3.2
Hyperbolic Functions
Practice Problems
34.3.3
Geometry
34.1
34.2
34.3
34.4
34.4
34.5
34.6
34.9
34.9
34.9
34.11
34.14
34.16
34.17
34.17
34.22
34.23
34.24
page 13
34.4
34.5
34.6
34.7
34.3.4
Planes, Lines, etc.
34.41
34.3.5
Practice Problems
34.43
COORDINATE SYSTEMS
34.45
34.4.1
Complex Numbers
34.45
34.4.2
Cylindrical Coordinates
34.48
34.4.3
Spherical Coordinates
34.49
34.4.4
Practice Problems
34.50
MATRICES AND VECTORS
34.51
34.5.1
Vectors
34.51
34.5.2
Dot (Scalar) Product
34.52
34.5.3
Cross Product
34.57
34.5.4
Triple Product
34.59
34.5.5
Matrices
34.59
34.5.6
Solving Linear Equations with Matrices
34.64
34.5.7
Practice Problems
34.65
CALCULUS
34.70
34.6.1
Single Variable Functions
34.70
Differentiation
34.70
Integration
34.73
34.6.2
Vector Calculus
34.77
34.6.3
Differential Equations
34.79
First-order Differential Equations
34.80
Guessing
34.81
Separable Equations
34.81
Homogeneous Equations and Substitution
34.82
Second-order Differential Equations
34.83
Linear Homogeneous
34.83
Nonhomogeneous Linear Equations
34.84
Higher Order Differential Equations
34.86
Partial Differential Equations
34.86
34.6.4
Other Calculus Stuff
34.87
34.6.5
Practice Problems
34.87
NUMERICAL METHODS
34.93
34.7.1
Approximation of Integrals and Derivatives from Sampled Data
34.93
34.8
34.9
34.10
34.11
34.12
34.7.2
Euler First-order Integration
34.7.3
Taylor Series Integration
34.7.4
Runge-Kutta Integration
34.7.5
Newton-Raphson to Find Roots
LAPLACE TRANSFORMS
34.8.1
Laplace Transform Tables
z-TRANSFORMS
FOURIER SERIES
TOPICS NOT COVERED (YET)
REFERENCES/BIBLIOGRAPHY
34.94
34.94
34.95
34.95
34.96
34.96
34.99
34.102
34.102
34.103
page 14
35.
35.7
35.8
35.9
36.
WHY USE C?
BACKGROUND
PROGRAM PARTS
HOW A C COMPILER WORKS
STRUCTURED C CODE
ARCHITECTURE OF C PROGRAMS (TOP-DOWN)
35.6.1
How?
35.6.2
Why?
CREATING TOP DOWN PROGRAMS
HOW THE BEAMCAD PROGRAM WAS DESIGNED
35.8.1
Objectives:
35.8.2
Problem Definition:
35.8.3
User Interface:
Screen Layout (also see figure):
Input:
Output:
Help:
Error Checking:
Miscellaneous:
35.8.4
Flow Program:
35.8.5
Expand Program:
35.8.6
Testing and Debugging:
35.8.7
Documentation
Users Manual:
Programmers Manual:
35.8.8
Listing of BeamCAD Program.
PRACTICE PROBLEMS
35.1
35.2
35.2
35.11
35.13
35.14
35.14
35.15
35.16
35.17
35.18
35.18
35.18
35.18
35.19
35.20
35.20
35.20
35.21
35.22
35.22
35.24
35.25
35.25
35.26
35.26
35.26
36.1
36.2
36.2
36.6
36.8
37.
37.
38.
FORMULA SHEET
37.4
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.1
38.1
TEXTBOOKS
38.1.1
Slotine and Li
38.1
38.1
page 15
38.1.2
39.
VandeVegte
38.1
39.1
39.4
39.8
39.9
39.9
39.9
39.10
39.10
39.12
39.14
39.15
39.16
39.16
39.17
39.18
39.18
39.18
39.18
39.18
39.18
39.18
39.18
39.19
39.19
39.19
39.19
introduction - 1.1
1. INTRODUCTION
Topics:
Objectives:
introduction - 1.2
Static
Dynamic
Stochastic
Deterministic
Figure 1.1
Distributed
Lumped
Model Classifications
Non-linear
Linear
Continuous
Discrete
introduction - 1.3
1.3 SUMMARY
translation - 2.1
2. TRANSLATION
Topics:
Basic laws of motion
Gravity, inertia, springs, dampers, cables and pulleys, drag, friction, FBDs
System analysis techniques
Design case
Objectives:
To be able to develop differential equations that describe translating systems.
2.1 INTRODUCTION
If the velocity and acceleration of a body are both zero then the body will be static.
If the applied forces are balanced, and cancel each other out, the body will not accelerate.
If the forces are unbalanced then the body will accelerate. If all of the forces act through
the center of mass then the body will only translate. Forces that do not act through the center of mass will also cause rotation to occur. This chapter will focus only on translational
systems.
The equations of motion for translating bodies are shown in Figure 2.1. These state
simply that velocity is the first derivative of position, and velocity is the first derivative of
acceleration. Conversely the acceleration can be integrated to find velocity, and the velocity can be integrated to find position. Therefore, if we know the acceleration of a body, we
can determine the velocity and position. Finally, when a force is applied to a mass, an
acceleration can be found by dividing the net force by the mass.
translation - 2.2
x,v,a
equations of motion
d
v ( t ) = ----- x ( t )
dt
d 2
d
a ( t ) = ----- x ( t ) = ----- v ( t )
dt
dt
OR
x ( t ) = v ( t ) dt = a ( t ) dt
v(t) =
a ( t ) dt
(1)
(2)
(3)
(4)
( t )(5)
a(t) = F
--------M
where,
x, v, a = position, velocity and acceleration
M = mass of the body
F = an applied force
Figure 2.1
An example application of these fundamental laws is shown in Figure 2.2. The initial conditions of the system are supplied (and are normally required to solve this type of
problem). These are then used to find the state of the system after a period of time. The
solution begins by integrating the acceleration, and using the initial velocity value for the
integration constant. So at t=0 the velocity will be equal to the initial velocity. This is then
integrated once more to provide the position of the object. As before, the initial position is
used for the integration constant. This equation is then used to calculate the position after
a period of time. Notice that the units are used throughout the calculations, this is a good
practice for any engineer.
translation - 2.3
Given an initial (t=0) state of x=5m, v=2m/s, a=3ms-2, find the system state 5 seconds
later assuming constant acceleration.
The initial conditions for the system at time t=0 are,
x 0 = 5m
Note: units are very important and should nor1
v 0 = 2ms
mally be carried through all calculations.
2
a 0 = 3ms
The constant acceleration can be integrated to find the velocity as a function of time.
Note:
v ( t ) = a 0 dt = a 0 t + C = a 0 t + v 0
(6)
v ( t ) = a0 t + C
v0 = a0 ( 0 ) + C
v0 = C
Next, the velocity can be integrated to find the position as a function of time.
a 2
(7)
x ( t ) = v ( t ) dt = ( a 0 t + v 0 ) dt = -----0- t + v 0 t + x 0
2
This can then be used to calculate the position of the mass after 5 seconds.
a 2
x ( 5 ) = -----0- t + v 0 t + x 0
2
2
2
1
3ms
= ---------------- ( 5s ) + 2ms ( 5s ) + 5m
2
= 37.5m + 10m + 5m = 52.5m
Figure 2.2
2.2 MODELING
When modeling translational systems it is common to break the system into parts.
These parts are then described with Free Body Diagrams (FBDs). Common components
that must be considered when constructing FBDs are listed below, and are discussed in
following sections.
gravity and other fields - apply non-contact forces
inertia - opposes acceleration and deceleration
springs - resist deflection
dampers and drag - resist motion
friction - opposes relative motion between bodies in contact
translation - 2.4
M = 10 kg
FBD Spring:
FBD Mass:
K = 20 N/m
FR1
FR1
Mg
FR2
Figure 2.3
translation - 2.5
forces, which will equal zero. The second method is known as DAlemberts equation
where all of the forces are summed and set equal to the inertial force, as shown in Figure
2.4. The acceleration is proportional to the inertial force and inversely proportional to the
mass.
F = Ma
F Ma =
Figure 2.4
(Newtons)
(11)
(DAlemberts)
(12)
An application of Newtons equation to FBDs can be seen in Figure 2.5. In the first
case an inertial force is added to the FBD. This force should be in an opposite direction
(left here) to the positive direction of the mass (right). When the sum of forces equation is
used then the force is added in as a normal force component. In the second case Newtons
equation is used so the force is left off the FBD, but added to the final equation. In this
case the sign of the inertial force is positive if the assumed positive direction of the mass
matches the positive direction for the summation.
translation - 2.6
DAlembertss form:
d 2
Ma = M ----- x
dt
d- 2
---F
=
F
M
x
dt x = 0
or
d- 2
---+
F
=
F
+
M
x
dt x = 0
+
Note: If using an inertial force then the direction of the force should be opposite to
the positive motion direction for the mass.
x
Newtons form:
d- 2
---F
=
F
=
M
+ x
dt x
F
or
d- 2
---F
=
F
=
M
x
dt x
Note: If using Newtons form the sign of the inertial force should be positive if the
positive direction for the summation and the mass are the same, otherwise if they
are opposite then the sign should be negative.
Figure 2.5
translation - 2.7
5
F 1 = 4 N
0
If both forces shown act through
the center of mass, what is the
acceleration of the ball?
7
F2 = 3 N
0
M=10kg
= F 1 + F 2 = Ma
5
7
N
+
4
3 N = ( 10Kg )a
0
0
2
0.2
0.2
1
Kgm
1
m
10Kg
Kg
s
s
0
0
0
Figure 2.6
Figure 2.7
translation - 2.8
translation - 2.9
0
0
N- =
g =
-----0 Kg
0
9.81
9.81
m---2
s
x(t)
X( t) = y(t)
z( t)
M = 5Kg
Fg
fx
F = f
y
fz
F g = Mg
Next, sum the forces and set them equal to inertial resistance.
d- 2
---F
=
Mg
+
F
=
M
dt X ( t )
fx
0
x(t)
d 2
m5Kg 0 ---+ f = 5Kg ----- y ( t )
y
dt
2
9.81 s
z(t)
fz
fx
0
x( t)
1
d- 2
m
---fy = y ( t )
0 ----2- + 0.2Kg
dt
z(t)
9.81 s
fz
Integrate twice to find the position components.
Note: When an engineer
solves a problem they
vx
fx
0
will always be looking
0
x(t)
1
d
m
at the equations and
f y t + v y = ----- y ( t )
0 ----2- + 0.2Kg
dt
unknowns. In this case
0
9.81 s
z
(
t
)
there are three equafz
vz
tions, and there are 9
0
vx
constants/givens fx, fy,
fx
x0
0
0
x
(
t
)
fz, vx0, vy0, xz0, x0, y0
2
1--m- + 0.2Kg 1
---fy t + vy t + y0 = y ( t )
0
and z0. There are 4
2
2
0
s
variables/unknowns x,
9.81
z(t)
fz
z0
vz
y, z and t. Therefore
2
0.1f x t + v x t + x 0 0
with 3 equations and 4
0
unknowns only one
x(t)
2
value (4-3) is required
m
0.1f y t + v y t + y 0
y(t) =
to find all of the
0
z(t )
unknown values.
9.81- + 0.1f t 2 + v t + z
-----------z
z0
0
2
Figure 2.8
translation - 2.10
Like gravity, magnetic and electrostatic fields can also apply forces to objects.
Magnetic forces are commonly found in motors and other electrical actuators. Electrostatic forces are less common, but may need to be considered for highly charged systems.
Given,
3
F1 = 4 N
0
0
Ng = 9.81 -----Kg
0
FBD:
M = 2Kg
F1
ans.
1.5
a = 7.81 m
---s
0
Figure 2.9
2.2.4 Springs
Springs are typically constructed with elastic materials such as metals and plastics,
that will provide an opposing force when deformed. The properties of the spring are determined by the Youngs modulus (E) of the material and the geometry of the spring. A primitive spring is shown in Figure 2.10. In this case the spring is a solid member. The
relationship between force and displacement is determined by the basic mechanics of
materials relationship. In practice springs are more complex, but the parameters (E, A and
L) are combined into a more convenient form. This form is known as Hookes Law.
translation - 2.11
L
= ------- F
EA
EA
F = -------
L
F = K
Figure 2.10
(Hookes law)
Hookes law does have some limitations that engineers must consider. The basic
equation is linear, but as a spring is deformed the material approaches plastic deformation,
and the modulus of elasticity will change. In addition the geometry of the object changes,
also changing the effective stiffness. Springs are normally assumed to be massless. This
allows the inertial effects to be ignored, such as a force propagation delay. In applications
with fast rates of change the spring mass may become significant, and they will no longer
act as an ideal device.
The cases for tension and compression are shown in Figure 2.11. In the case of
compression the spring length has been made shorter than its normal length. This requires
that a compression force be applied. For tension, both the displacement from neutral and
the required force reverse direction. It is advisable when solving problems to assume a
spring is either in tension or compression, and then select the displacement and force
directions accordingly.
translation - 2.12
x = deformed length
compression as positive:
Fc
Ks
Fc
Ks
F c = K s x
Fc
Ks
tension as positive:
Fc = Ks x
F c = K s x
Ks
Ft
Ft = Ks x
x
Ks
Ft
Ft = Ks x
x
Ks
Ft
Ft = Ks x
NOTE: the symbols for springs, resistors and inductors are quite often the same or similar. You will need to remember this when dealing with complex systems - and especially in this book where we deal with both types of components.
Figure 2.11
Previous examples have shown springs with displacements at one end only. In Figure 2.12, springs are shown that have movement at both ends. In these cases the sign of the
force applied to the spring is selected with reference to the assumed compression or tension. The primary difference is that care is required to correctly construct the expressions
for the tension or compression forces. In all cases the forces on the springs must be
assumed and drawn as either tensile or compressive. In the first example the displacement
and forces are tensile. The displacement at the left is tensile, so it will be positive, but on
the right hand side the displacement is compressive so it is negative. In the second exam-
translation - 2.13
ple the force and both displacements are shown as tensile, so the terms are both positive.
In the third example the force is shown as compressive, while the displacements are both
shown as tensile, so both terms are negative.
x1
x1
x1
Ks
Ks
Ks
x2
x2
x2
F = Ks ( x1 x 2 )
F = Ks ( x 1 + x 2 )
F = Ks ( x1 x2 )
Figure 2.12
The example in Figure 2.13 shows two masses separated by a spring. In the first
example the spring is assumed to be in tension. When x1 becomes positive it will put the
spring in compression, so it is made negative, however a positive x2 will put the spring in
tension, so it remains positive. In the second case the spring is assumed to be in compression and this time x2 will put it in tension so it is made negative.
translation - 2.14
M2
x1
x2
The system can be reduced to free body diagrams assuming the spring is tension.
Ks ( x1 + x2 )
M1
M2
x1
x2
Note: In this example the spring is assumed to be in tension and the signs of the
magnitude are made negative for terms that result in compression for positive changes.
The system can be reduced to free body diagrams assuming the spring is in compression.
Ks ( x1 x2 )
M1
x1
Figure 2.13
M2
x2
Sometimes the true length of a spring is important, and the deformation alone is
insufficient. In these cases the deformation can be defined as a deformed and undeformed
length, as shown in Figure 2.14.
translation - 2.15
x = l1 l 0
where,
x = deformation
l 0 = the length when undeformed
l 1 = the length when deformed
Figure 2.14
K ( x )
E P = -----------------2
Figure 2.15
translation - 2.16
Given,
Ks
NK s = 10 --m
Ks
x = 0.1m
x
Find F assuming the springs are
normally 4m long when unloaded
10m
ans. F=2N
Figure 2.16
translation - 2.17
Draw the FBDs and the equations of motion for the masses
F
Ks
M1
M2
x1
x2
ans.
x1 ( M1 ) + x1 ( Ks ) + x2 ( Ks ) = F
x 1 ( K s ) + x2 ( M2 ) + x2 ( K s ) = F
Figure 2.17
translation - 2.18
orifice
motion
fluid
Figure 2.18
piston
fluid
A physical damper
The basic equation for an ideal damper in compression is shown in Figure 2.19. In
this case the force and displacement are both compressive. The force is calculated by multiplying the damping coefficient by the velocity, or first derivative of position. Aside from
the use of the first derivative of position, the analysis of dampers in systems is similar to
that of springs.
translation - 2.19
x
Kd
d
F = K d ----- x
dt
(15)
Aside: The symbol shown is typically used for dampers. It is based on an old damper
design called a dashpot. It was constructed using a small piston inside a larger
pot filled with oil.
Figure 2.19
An ideal damper
Damping can also occur when there is relative motion between two objects. If the
objects are lubricated with a viscous fluid (e.g., oil) then there will be a damping effect. In
the example in Figure 2.20 two objects are shown with viscous friction (damping)
between them. When the system is broken into free body diagrams the forces are shown to
be a function of the relative velocities between the blocks.
x1
Fd = Kd ( x1 x2 )
Kd
Fd = Kd ( x1 x2 )
x2
Aside: Fluids, such as oils, have a significant viscosity. When these materials are put
in shear they resist the motion. The higher the shear rate, the greater the resistance to flow. Normally these forces are small, except at high velocities.
Figure 2.20
A damping force is proportional to the first derivative of position (velocity). Aerodynamic drag is proportional to the velocity squared. The equation for drag is shown in
translation - 2.20
Figure 2.21 in vector and scalar forms. The drag force increases as the square of velocity.
Normally, the magnitude of the drag force coefficient D is approximated theoretically
and/or measured experimentally. The drag coefficient is a function of material type, surface properties, object size and object geometry.
v
F
Figure 2.21
F = D v v
Aerodynamic drag
The force is acting on the cylinder, resulting in the velocities given below. What is the
applied force?
d- x = 0.1 m
------dt 1
s
d- x = 0.3 m
------dt 2
s
k d = 0.1 Ns
-----m
ans. F = 0.02N
Figure 2.22
translation - 2.21
T1
for a massless frictionless pulley
T1 = T2
T2
Figure 2.23
If we have a pulley that is fixed and cannot rotate, the cable must slide over the
surface of the pulley. In this case we can use the coefficient of friction to determine the relative ratio of forces between the sides of the pulley, as shown in Figure 2.24.
T1
T2
Figure 2.24
T2
( )
----- = e k
T1
translation - 2.22
Given,
s = 0.35
k = 0.2
M = 1Kg
Find F to start the mass moving up and down, and then
the force required to maintain a low velocity motion.
M
F
ans.
F up = 9.81e
0.2 ---
2
9.81N
F down = --------------e
Figure 2.25
0.2 ---
2
Although the discussion in this section has focused on cables and pulleys, the theory also applies to belts over drums.
translation - 2.23
2.2.7 Friction
Viscous friction was discussed before, where a lubricant would provide a damping
effect between two moving objects. In cases where there is no lubricant, and the touching
surfaces are dry, dry coulomb friction may result. In this case the surfaces will stick in
place until a maximum force is overcome. After that the object will begin to slide and a
constant friction force will result.
Figure 2.26 shows the classic model for (dry Coulomb) friction. The force on the
horizontal axis is the force applied to the friction surfaces while the vertical axis is the
resulting friction force. Beneath the slip force the object will stay in place. When the slip
force is exceeded the object will begin to move, and the resulting kinetic friction force will
be relatively constant. (Note: If the object begins to travel much faster then the kinetic
friction force will decrease.) It is common to forget that friction forces are bidirectional,
but it always opposes the applied force or motion. The friction force is a function of the
coefficient of friction and the normal force across the contact surfaces. The coefficient of
friction is a function of the materials, surface texture and surface shape.
F result
Fs
Fk
Fapplied
Fs
Fg
N
Fk = k N
F k, F s
Fs s N
Note: When solving problems with friction remember that the friction force will always
equal the applied force (not the maximum force) until slip occurs. After that the friction is approximately constant. In addition, the friction forces direction opposes
applied forces, and motion.
Figure 2.26
Dry friction
translation - 2.24
Many systems use kinetic friction to dissipate energy from a system as heat, sound
and vibration.
s = 0.3
k = 0.2
10 kg
ans.
Figure 2.27
1 = 5
2 = 35
translation - 2.25
F Ay
F By
F Ax
F Bx
M1
B
M1 g
M2 g
M2
F Bx
F By
Figure 2.28
Note: Dont forget that forces on connected FBDs should have equal
magnitudes, but opposite directions.
translation - 2.26
1. Assign letters/numbers to designate force components (if not already done) this will allow you to refer to components in your calculations.
2. Define positions and directions for any moving masses. This should include the
selection of reference points.
3. Draw free body diagrams for each component, and add forces (inertia is
optional).
4. Write equations for each component by summing forces.
5.(next chapter) Combine the equations by eliminating unwanted variables.
6.(next chapter) Develop a final equation that relates input (forcing functions) to
outputs (results).
Note: When deriving differential equations, the final value can be checked for
errors using unit analysis. This method
involves replacing variables with their
unit equivalents. All the units should
match.
e.g., x 2 ( M 2 ) + x 2 ( B ) + x 2 ( K s2 ) + x 1 ( K s2 ) = F
m
m Ns
N
N
----2 ( Kg ) + ---- ------ + m ---- + m ---- = N
s
m
m
m
s
N + ( N ) + ( N ) + ( N ) = N
coefficient
units
KgmN = ----------2
s
Ks
N--m
Kd
Ns
-----m
Kg
Consider the cart in Figure 2.29. On the left is a force, it is opposed by a spring and
damper on the right. The basic problem definition already contains all of the needed definitions, so no others are required. The FBD for the mass shows the applied force and the
reaction forces from the spring and damper. When the forces are summed the inertia is on
the right side of the equation in Newtons form. This equation is then rearranged to a second-order non-homogeneous differential equation.
translation - 2.27
Kd
F
M1
K d x
F
M 1 x
Ks
M1
Ksx
The forces for the cart are in a single direction and can be summed as,
+
Fx = F Kd x Ks x = M1 x
This equation can be rearranged to a second-order non-homogeneous diff. eqn.
Kd
K
Fx + ------- x + -------s- x = ------M1
M1
M1
Aside: later on we will solve the differential equations, or use other methods to
determine how the system will behave. It is useful to have all of the output
variables for the system on the left hand side, and everything else on the other.
Figure 2.29
translation - 2.28
Develop the equation relating the input force to the motion (in terms of x) of the lefthand
cart for the problem below.
x1
x2
K d1
K d2
F
M1
Figure 2.30
K s1
M2
K s2
A simplified model of an elevator (M1) and a passenger (M2) are shown in Figure
2.31. In this example many of the required variables need to be defined. These are added
to the FBDs. Care is also taken to ensure that all forces between bodies are equal in magnitude, but opposite in direction. The wall forces are ignored because they are statically
indeterminate, and x-axis force components are irrelevant to the forces in the y-axis.
translation - 2.29
M1
M2
Kd
K S2
K S1
FD = Kd ( y2 y1 )
M1 y1
y2
M1 g
F S2 = K s2 ( y 1 y 2 )
FD
M2 y2
F S2
F S1 = K s1 y 1
y1
Figure 2.31
The forces on the FBDs are summed and the equations are expanded in Figure
2.32.
translation - 2.30
FM
FM
= F S1 + F S2 + M 1 g + F D = M 1 a 1
= F S2 + M 2 g F D = M 2 a 2
translation - 2.31
M2
k1
M1
K S1
k2
ans.
Figure 2.33
Consider the springs shown in Figure 2.34. When two springs are combined in this
manner they can be replaced with a single equivalent spring. In the parallel spring combination the overall stiffness of the spring would increase. In the series spring combination
the overall stiffness would decrease.
translation - 2.32
Parallel
K S1
Series
K S2
K S1
K S2
Figure 2.34
Figure 2.35 shows the calculations required to find a spring coefficient equivalent
to the two springs in series. The first step is to draw a FBD for the mass at the bottom, and
for a point between the two springs, P. The forces for both of these are then summed. The
next process is to combine the two equations to eliminate the height variable created for
point P. After this, the equation is rearranged into Hookes law, and the equivalent spring
coefficient is found.
translation - 2.33
First, draw FBDs for P and M and sum the forces assuming the system is static.
K S1 y 2
+ F y = K S1 y 2 K S2 ( y 1 y 2 ) = 0 (1)
K S2 ( y 1 y 2 )
K S1
K S2 ( y 1 y 2 )
K S2
M
+
Fy
= K S2 ( y 1 y 2 ) F g = 0 (2)
Fg
M
y1
K S2 ( y 1 y 2 ) Fg = 0
Fg
y 1 y 2 = --------K S2
Fg
y 2 = y 1 ---------K S2
(2) becomes
(3)
K S1 y 2 K S2 ( y 1 y 2 ) = 0
(4)
y 2 ( K S1 + K S2 ) = y 1 K S2
Fg
F g = y 1 1 ---------------------------- K S2
K S1 + K S2
K S1 + K S2 K S2
F g = y 1 -------------------------------------------- K S2
K S1 + K S2
K S1 K S2
F g = y 1 ----------------------------
K S1 + K S2
Finally, consider the basic spring equation to find the equivalent spring coefficient.
K S1 K S2
K equiv = ---------------------------K S1 + K S2
Figure 2.35
y2
translation - 2.34
K S1
K S2
ans.
Figure 2.36
Consider the drill problem. When an object has no mass, the force applied to one
side of the spring will also be applied to the other. The only factor that changes is displacement.
translation - 2.35
Show that a force applied to one side of a massless spring is the reaction force at the other
side.
ans.
Figure 2.37
translation - 2.36
E = EP + EK
2
Mv
E K = ----------2
(7)
E P = Fd = Mgd
(9)
d- E
P = Fv = ---dt
(10)
Figure 2.38
(8)
2.5 SUMMARY
FBDs are useful for reducing complex systems to simpler parts.
Equations for translation and rotation can be written for FBDs.
The equations can be integrated for dynamic cases, or solved algebraically for
static cases.
translation - 2.37
2. Derive the effective damping coefficients for the pairs below from basic principles,
a)
Kd1
Kd2
b)
Kd1
Kd2
Ks
M1
M2
translation - 2.38
x2
K d1
K s1
M1
K s2
M2
B
Ks
M1
M2
B1
B2
x2
K d1
K s1
M1
K s2
M2
B
x2
K d1
K s1
K d2
M1
K s2
K d3
M2
K s3
translation - 2.39
K s2
M2
K d1
F
B
x1
K s1
M1
K s2
M2
F
s, k
K d1
K s1
M1
x1
M1
K s1
K d1
x2
M2
K s2
K d2
translation - 2.40
12. Write the differential equations for the system below. Assume that the pulley is massless and
frictionless and that the system begins undeflected.
x
K s1
M
B
K s2
13. Write the differential equations for the system below. In this system the upper mass, M1, is
between a spring and a cable and there is viscous damping between the mass and the floor. The
suspended mass, M2, is between the cable and a damper. The cable runs over a massless, frictionless pulley.
x1
Ks
M1
B
x2
M2
Kd
translation - 2.41
K s1
M1
s, k
x2
M2
K s2
K eq = K d1 + K d2
3.
F
B
x + x ----- = ----M
M
4.
Ks
Ks
F
x 1 + x 1 ------- + x 2 --------- = ------M1
M1
M1
Ks
Ks
x 2 + x 2 ------- + x 1 --------- = 0
M 2
M2
b)
K d1 K d2
K eq = -----------------------K d1 + K d2
translation - 2.42
5.
K s1 + K s2
K s2
K d1
x 1 + x 1 --------- + x 1 ----------------------- + x 2 ----------- = 0
M1
M1
M1
K s2
K s2
B
F
x 2 + x 2 ------- + x 2 -------- + x 1 ----------- = ------M2
M2
M2
M2
6.
Ks
K s
Kd + B1
x 1 + x 1 ------------------- + x 1 ------- + x 2 --------- = 0
M1
M 1
M1
Ks
K s
B2
F
x 2 + x 2 ------- + x 2 ------- + x 1 --------- = ------M2
M2
M2
M2
7.
K s1
K s2
K d1
x 1 + x 1 --------- + x 1 -------- + x 2 ----------- = 0
M1
M1
M1
Ks
F
B
x 1 + x 1 ------- + x 2 + x 2 ------- + x 2 ------- = ------M2
M2
M2
M2
translation - 2.43
8.
M 1 x1
FBDs:
K d1 x 1
M1
K s1 x 1
For M1:
K d2 ( x 1 x 2 )
K s2 ( x 1 x 2 )
K d3 x 2
M2
K s3 x 2
M2 x2
x 1 ( M 1 ) + x 1 ( K d1 + K d2 ) + x 1 ( K s1 + K s2 ) + x 2 ( K d2 ) + x 2 ( K s2 ) = 0
K s1 + K s2
K s2
K d1 + K d2
K d2
x 1 + x 1 ------------------------ + x 1 ----------------------- + x 2 ------------ + x 2 ----------- = 0
M1
M1
M1
M1
For M2:
+
= K d2 ( x 1 x 2 ) + K s2 ( x 1 x 2 ) + F K d3 x 2 K s3 x 2 = M 2 x 2
x 2 ( M 2 ) + x 2 ( K d2 + K d3 ) + x 2 ( K s2 + K s3 ) + x 1 ( K d2 ) + x 1 ( K s2 ) = F
K s2 + K s3
K s2
K d2 + K d3
K d2
x 2 + x 2 ------------------------ + x 2 ----------------------- + x 1 ------------ + x 1 ----------- = F
M2
M2
M2
M2
9.
x 1 ( M 1 ) + x 1 ( K d1 + B ) + x 1 ( K s1 ) + x 2 ( B ) = 0
K s1 B
K d1 + B-
- + x 2 ------- = 0
+ x 1 ------x 1 + x 1 ----------------- M1
M1
M 1
x 2 ( M 2 ) + x 2 ( B ) + x 2 ( K s2 ) + x 1 ( B ) = F
K s2
B
B
F
x 2 + x 2 ------- + x 2 -------- + x 1 ------- = ------ M 2
M2
M 2
M2
translation - 2.44
10.
x 1 ( M 1 ) + x 1 ( K d1 ) + x 1 ( K s1 ) = FF
K s1
FF
K d1
x 1 + x 1 --------- + x 1 -------- = ------ M1
M1
M1
x 2 ( M 2 ) + x 2 ( K s2 ) = F F F
K s2
F FF
x 2 + x 2 -------- = ---------------M2
M2
where,
FF s M2 g
if
x1 = x2
x1 x2
F F = k M 2 g -----------------
-
x1 x2
if
x1 x2
11.
(assuming no gravity
x 1 ( M 1 ) + x 1 ( K d1 ) + x 1 ( K s1 ) + x 2 ( K d1 ) + x 2 ( K s1 ) = F
K s1
K s1
K d1
K d1
F
x 1 + x 1 --------- + x 1 -------- + x 2 ------------ + x 2 ----------- = ------ M1
M1
M1
M1
M1
x 2 ( M 2 ) + x 2 ( K d1 + K d2 ) + x 2 ( K s1 + K s2 ) + x 1 ( K d1 ) + x 1 ( K s1 ) = 0
K s1 + K s2
K s1
K d1 + K d2
K d1
x 2 + x 2 ------------------------ + x 2 ----------------------- + x 1 ------------ + x 1 ----------- = 0
M2
M2
M2
M2
12.
if ( x 0 )
T = 0
if ( x < 0 )
T = K s2 x
K s1
T
B
x + x ------- + x -------- = ------M1
M1
M1
translation - 2.45
13.
Ks x 1
FBDs:
M1
Bx 1
For M1:
For M2:
+
For T:
M1 x1
F = Ks x1 Bx 1 + T = M1 x1
x1 ( M1 ) + x1 ( B ) + x1 ( Ks ) = T
Ks
B
T
x 1 + x 1 ------- + x 1 ------- = ------ M 1
M 1
M1
= T + Kd x2 M2 g = M2 x2
x2 ( M2 ) + x2 ( K d ) = T M 2 g
T M2 g
Kd
x 2 + x 2 ------- = ------------------- M 2
M2
if T <= 0 then T = 0N
if T > 0 x1 = x2
14.
if ( ( x 2 x 1 ) < 0 )
T = 0
if ( x 1 = 0 )
F F M 1 g s
if ( x 1 0 )
x1
F F = M 1 g k ------x1
T FF
K s1
x 1 + x 1 -------- = --------------M1
M1
K s2
T
x 2 + x 2 -------- = ------- g
M2
M2
M2
M2 g
M2 x2
Kd x2
Topics:
First and second-order homogeneous differential equations
Non-homogeneous differential equations
First and second-order responses
Non-linear system elements
Design case
Objectives:
To develop explicit equations that describe a system response.
To recognize first and second-order equation forms.
3.1 INTRODUCTION
In the previous chapter we derived differential equations of motion for translating
systems. These equations can be used to analyze the behavior of the system and make
design decisions. The most basic method is to select a standard input type (a forcing function) and initial conditions, and then solve the differential equation. It is also possible to
estimate the system response without solving the differential equation as will be discussed
later.
Figure 3.1 shows an abstract description of a system. The basic concept is that the
system changes the inputs to outputs. Say, for example, that the system to be analyzed is
an elevator. Inputs to the system would be the mass of human riders and desired elevator
height. The output response of the system would be the actual height of the elevator. For
analysis, the system model could be developed using differential equations for the motor,
elastic lift cable, mass of the car, etc. A basic test would involve assuming that the elevator
starts at the ground floor and must travel to the top floor. Using assumed initial values and
input functions the differential equation could be solved to get an explicit equation for elevator height. This output response can then be used as a guide to modify design choices
(parameters). In practice, many of the assumptions and tests are mandated by law or by
groups such as Underwriters Laboratories (UL), Canadian Standards Association (CSA)
and the European Commission (CE).
inputs
system
forcing
function
differential
equations
Figure 3.1
outputs
response
function
There are several standard input types used to test a system. These are listed below
in order of relative popularity with brief explanations.
step - a sudden change of input, such as very rapidly changing a desired speed
from 0Hz to 50Hz.
ramp - a continuously increasing input, such as a motor speed that increases constantly at 10Hz per minute.
sinusoidal - a cyclic input that varies continuously, such as wave height that is
continually oscillating at 1Hz.
parabolic - an exponentially increasing input, such as a motor speed that is 2Hz at
1 second, 4rad/s at 2 seconds, 8rad/s at 3 seconds, etc.
After the system has been modeled, an input type has been chosen, and the initial
conditions have been selected, the system can be analyzed to determine its behavior. The
most fundamental technique is to integrate the differential equation(s) for the system.
Ax + Bx = 0
Ax + Bx = Cf ( t )
Ax + Bx + Cx = 0
Ax + Bx + Cx = Df ( t )
Figure 3.2
first-order homogeneous
first-order non-homogeneous
second-order homogeneous
second-order non-homogeneous
Yt
) + B ( Xe
Yt
) = 0
A ( Y ) + B = 0
Y = B
--A
Therefore the general form is,
B
---t
A
x h = Xe
initial condition
Figure 3.3
B
--- t
A
Yt
ans.
Figure 3.4
x ( t ) = 3e
2t
Given,
Ax + Bx + Cx = 0
x ( 0 ) = x0
and
x ( 0 ) = v 0
Guess a general equation form and substitute it into the differential equation,
2 Yt
Yt
Yt
xh = Y Xe
x h = Xe
x h = YXe
2 Yt
Yt
Yt
A ( Y Xe ) + B ( YXe ) + C ( Xe ) = 0
2
A(Y ) + B( Y) + C = 0
2
2
(
B
)
4
(
AC
)
B
4AC
Y = ---------------------------------------------------- = -----------------------------------------2A
2A
Note: There are three possible outcomes of finding the roots of the equations: two different real roots, two identical real roots, or two complex
roots. Therefore there are three fundamentally different results.
If the values for Y are both real, but different, the general form is,
R1 t
R2 t
xh = X1 e
+ X2 e
Note: The initial conditions are then used to find the values for X1 and X2.
Y = R 1, R 2
If the values for Y are both real, and identical, the general form is,
R1t
R1 t
xh = X1 e
+ X 2 te
The initial conditions are then used to find the values for X1 and X2.
Y = R 1, R 1
t
x h = X 3 e cos ( t + X 4 )
The initial conditions are then used to find the values of X3 and X4.
Y = j
Figure 3.5
Consider the situation where the results of a homogeneous solution are the complex
conjugate pair.
Y = R Cj
This gives the general result, as shown below:
( R + C j )t
( R Cj )t
x = X1 e
+ X2e
Rt Cjt
Rt Cjt
x = X1 e e
+ X2 e e
Rt
Cjt
Cjt
x = e X1 e
+ X2 e
Rt
x = e ( X 1 ( cos ( Ct ) + j sin ( Ct ) ) + X 2 ( cos ( C t ) + j sin ( C t ) ) )
Rt
x = e ( X 1 ( cos ( Ct ) + j sin ( Ct ) ) + X 2 ( cos ( ( Ct ) j sin ( Ct ) ) ) )
Rt
x = e ( ( X1 + X 2 ) cos ( Ct ) + j ( X 1 X 2 ) sin ( Ct ) )
Rt
x = e ( ( X1 + X 2 ) cos ( Ct ) + j ( X 1 X 2 ) sin ( Ct ) )
2 2
2
Rt ( X 1 + X 2 ) + j ( X 1 X 2 )
x = e ------------------------------------------------------------------------- ( ( X 1 + X 2 ) cos ( Ct ) + j ( X 1 X2 ) sin ( Ct ) )
2 2
2
( X1 + X2 ) + j ( X1 X2 )
2
2
2
2
X 1 + 2X 1 X 2 + X 2 X 1 + 2 X 1 X 2 + X 2
Rt
x = e ------------------------------------------------------------------------------------------------------------- ( ( X 1 + X 2 ) cos ( Ct ) + j ( X 1 X 2 ) sin ( C
2
2
2
2
X 1 + 2X 1 X 2 + X 2 X 1 + 2 X 1 X 2 + X 2
Rt 4X 1 X 2
--------------------- ( ( X 1 + X 2 ) cos ( Ct ) + j ( X 1 X 2 ) sin ( Ct ) )
4X 1 X 2
( X1 X2 )
( X1 + X2 )
Rt
x = e
4X 1 X2 ------------------------ cos ( Ct ) + j ----------------------- sin ( Ct )
4X X
4X 1 X 2
1 2
( X1 X2 )
Rt
x = e
4X 1 X 2 cos Ct + atan -------------------------
( X1 + X 2 )
x = e
x = e
Rt
X3 cos ( Ct + X 4 )
frequency
Figure 3.6
phase shift
where, X =
3
4X 1 X 2
( X1 X2 )
X 4 = atan -------------------------
( X 1 + X 2 )
Note: Occasionally a problem solution might consist of both a sine and cosine term with
the same frequency. These should normally be combined to a single term with a phase
shift as shown below.
Recall the double angle formula,
sin ( t + ) = sin t cos + sin cos t
This can be written in a more common form,
A ( sin t cos + sin cos t ) = A sin ( t + )
A cos sin t + A sin cos t = A sin ( t + )
B sin t + C cos t = A sin ( t + )
where,
B
C
A = ------------ = ----------cos
sin
sin C
----------= ---cos
B
A =
B = A cos
C = A sin
C
= atan ----
B
B +C
Figure 3.7
4
2
2
3 + 4 sin 5t + atan --- = 5 sin ( 5t + 0.927 )
3
x( 0) = 1
x ( 0 ) = 2
x + 2x + x = 0
ans.
Figure 3.8
x ( t ) = e + 3te
The methods for solving non-homogeneous differential equations builds upon the
methods used for the solution of homogeneous equations. This process adds a step to find
the particular solution of the equation. An example of the solution of a first-order nonhomogeneous equation is shown in Figure 3.9. To find the homogeneous solution the nonhomogeneous part of the equation is set to zero. To find the particular solution the final
form must be guessed. This is then substituted into the equation, and the values of the
coefficients are found. Finally the homogeneous and particular solutions are added to get
the final equation. The overall response of the system can be obtained by adding the
homogeneous and particular parts. This is acceptable because the equations are linear, and
the principle of superposition applies. The homogeneous equation deals with the response
to initial conditions, and the particular solution deals with the response to forced inputs.
Generally,
Ax + Bx = Cf ( t )
x ( 0 ) = x0
First, find the homogeneous solution as before, in Figure 3.3.
B
---t
A
xh = x0 e
Next, guess the particular solution by looking at the form of f(t). This step is highly
subjective, and if an incorrect guess is made, it will be unsolvable. When this happens, just make another guess and repeat the process. An example is given below. In
the case below the guess should be similar to the exponential forcing function.
For example, if we are given
4t
6x + 2x = 5e
A reasonable guess for the particular solution is,
4t
4t
xp = C1 e
\x p = 4C 1 e
Substitute these into the differential equation and solve for A.
4t
4t
4t
6 4C 1 e + 2 C 1 e = 5e
524C 1 + 2C 1 = 5
C 1 = ----26
Combine the particular and homogeneous solutions.
6--- t
5- e 4t + x e 2
x = x p + x h = ----0
26
Figure 3.9
Generally,
Ax + Bx + Cx = Df ( t )
x( 0 ) = x0
and
x ( 0 ) = v 0
1. Find the homogeneous solution as before.
t
x h = X 3 e cos ( t + X4 )
or
t
t
x h = X 1 e + X 2 te
or
1 t
2 t
+ X2 e
xh = X1 e
2. Guess the particular solution by looking at the form of f(t). This step is highly subjective, and if an incorrect guess is made it will be unsolvable. When this happens,
just make another guess and repeat the process. For the purpose of illustration an
example is given below. In the case below it should be similar to the sine function.
For example, if we are given
2x + 6x + 2x = 2 sin ( 3t + 4 )
A reasonable guess is,
x p = A sin ( 3t ) + B cos ( 3t )
x p = 3A cos ( 3t ) 3B sin ( 3t )
xp = 9A sin ( 3t ) 9B cos ( 3t )
Substitute these into the differential equation ans solve for A and B.
2 ( 9 A sin ( 3t ) 9B cos ( 3t ) ) + 6 ( 3A cos ( 3t ) 3B sin ( 3t ) ) +
2 ( A sin ( 3t ) + B cos ( 3t ) ) = 2 sin ( 3t + 4 )
( 18 A 18B + 2A ) sin ( 3t ) + ( 18 B + 18A + 2B ) cos ( 3t ) = 2 sin ( 3t + 4 )
( 16A 18B ) sin ( 3t ) + ( 18A 16B ) cos ( 3t ) = 2 ( sin 3t cos 4 + cos 3t sin 4 )
( 16A 18B ) sin ( 3t ) + ( 18A 16B ) cos ( 3t ) = ( 2 cos 4 ) sin ( 3t ) + ( 2 sin 4 ) cos ( 3t )
18A 16B = 2 sin 4
16A 18B = 2 cos 4
16 18 A = 2 cos 4 A = 16 18
18 16 B
2 sin 4 B
18 16
1.307 = 0.0109
1.514
0.0823
0.0823
2
2
0.0109 + 0.0823 sin 3t + atan ------------------- + ---
0.0109
2
3. Use the initial conditions to determine the coefficients in the homogeneous solution.
xp =
Figure 3.10
When guessing particular solutions, the forms in Figure 3.11 can be helpful.
Forcing Function
Guess
Ax + B
Ax
e
Cx + D
Ax or
Ax
Ce
Cxe
B sin ( Ax )
Figure 3.11
or B cos ( Ax )
C sin ( Ax ) + D cos ( Ax )
or Cx sin ( Ax ) + xD cos ( Ax )
ans.
Figure 3.12
x ( t ) = e te + 1
the homogeneous equation. Real roots are assumed thus allowing the problem solution to
continue in Figure 3.14.
M
Assume the system illustrated to the right starts
from rest at a height h. At time t=0 the system is released and allowed to move.
Fg
Ks
y
Kd
Mg
My
+
K d y
Ksy
Fy
= Mg + K s y + K d y = M y
My + K d y + K s y = Mg
The solution continues by assuming a particular solution and calculating values for
the coefficients using the initial conditions in Figure 3.14. The final result is a secondorder system that is overdamped, with no oscillation.
yp = C
yh = 0
yh = 0
M ( 0 ) + K d ( 0 ) + K s ( C ) = Mg
Mg
C = -------Ks
Now, add the homogeneous and particular solutions and solve for the unknowns
using the initial conditions.
R t
R t
Mg
y ( t ) = y p + y h = -------- + C 1 e 1 + C 2 e 2
Ks
y(0) = h
y' ( 0 ) = 0
Mg
0
0
h = -------- + C 1 e + C 2 e
Ks
Mg
C 1 + C 2 = h -------Ks
y' ( t ) = R 1 C 1 e
R1t
+ R2 C2 e
0 = R1 C1 e + R2 C2 e
R2 t
0 = R1 C 1 + R2 C2
R2
C 1 = --------- C 2
R1
R2
Mg
------ C 2 + C 2 = h -------R1
Ks
K s h Mg
R1
C 2 = ----------------------- ------------------
K s R 1 R 2
R2 K s h Mg
R1
C 1 = --------- ----------------------- ------------------
R1
Ks
R1 R2
Now, combine the solutions and solve for the unknowns using the initial conditions.
K s h Mg
R1
R1
R t R 2 K s h Mg
R t
Mg
y ( t ) = -------- + ----------------------- ------------------ e 1 + --------- ----------------------- ------------------ e 2
Ks
R1 R2
R1
Ks
R1 R 2
Ks
K s h Mg
K s h Mg
R1
R2
R1 t
R2 t
Mg
y ( t ) = -------- + ----------------------- ------------------ e + ----------------------- ------------------ e
Ks
R1 R2
Ks
R1 R 2
Ks
Figure 3.14
Given,
A sin ( t + )
To find theta,
B- = --------------A sin - = tan
--C
A cos
= atan ----
C
B
Figure 3.15
3.3 RESPONSES
Solving differential equations tends to yield one of two basic equation forms. The
e-to-the-negative-t forms are the first-order responses and slowly decay over time. They
never naturally oscillate, and only oscillate if forced to do so. The second-order forms may
3.3.1 First-order
A first-order system is described with a first-order differential equation. The
response function for these systems is natural decay or growth as shown in Figure 3.16.
The time constant for the system can be found directly from the differential equation. It is
a measure of how quickly the system responds to a change. When an input to a system has
changed, the system output will be approximately 63% of the way to its final value when
the elapsed time equals the time constant. The initial and final values of the function can
be determined algebraically to find the first-order response with little effort.
If we have experimental results for a system, we can calculate the time constant,
initial and final values. The time constant can be found two ways, one by extending the
slope of the first part of the curve until it intersects the final value line. That time at the
intersection is the time constant. The other method is to look for the time when the output
value has shifted 63.2% of the way from the initial to final values for the system. Assuming the change started at t=0, the time at this point corresponds to the time constant.
y
t
-
y0
y ( t ) = y 1 + ( y 0 y 1 )e
y + --1- y = f ( t )
time constant
y1
OR
y
y1
y ( ) = y 1 + ( y 0 y 1 )e
y0
Figure 3.16
y ( ) = y 1 + ( y 0 y 1 )e
y ( ) = y 1 + ( y 0 y 1 )0.368
y ( ) = 1 + ( 0 1 )0.368
y ( ) = 0.632
The example in Figure 3.17 calculates the coefficients for a first-order differential
equation given a graphical output response to an input. The differential equation is for a
permanent magnet DC motor, and will be examined in a later chapter. If we consider the
steady state when the speed is steady at 1400RPM, the first derivative will be zero. This
simplifies the equation and allows us to calculate a value for the parameter K in the differential equation. The time constant can be found by drawing a line asymptotic to the start of
the motor curve, and finding the point where it intercepts the steady-state value. This gives
an approximate time constant of 0.8 s. This can then be used to calculate the remaining
coefficient. Some additional numerical calculation leads to the final differential equation
as shown.
For the motor, use the differential equation and the speed curve when Vs=10V is applied:
K 2
d-
K-
--- ---------- dt + JR = JR V s
1400 RPM
1s
2s
3s
For steady-state
1
d-
--- = 1400RPM = 146.6rads
dt = 0
K 2
K
0 + ------- 146.6 = ------ 10
JR
JR
K = 0.0682
1400 RPM
0.8s
1s
K 2
1
------- = ---------0.8s
JR
K
10.0682 ------ = -------- JR
0.8s
K- = 18.328
----JR
1 + ----- = 18.328Vs
0.8
Figure 3.17
1800 RPM
d-
K
K-
--- ---- ---- dt + JR = JR V s
0.2s
0.4s
0.6s
ans.
Figure 3.18
d-
1
1800
--- + ---------- = --------------------- V s
dt
0.15
12 ( 0.15 )
A simple mechanical example is given in Figure 3.19. The modeling starts with a
FBD and a sum of forces. After this, the homogenous solution is found by setting the nonhomogeneous part to zero and solving. Next, the particular solution is found, and the two
solutions are combined. The initial conditions are used to find the remaining unknown
coefficients.
F
y
d- y = 0
+ F y = F + K s y + K d ---dt
d
K d ----- y
Ks y
dt
K d y + K s y = F
Find the homogeneous solution.
Bt
Bt
y h = Ae
y h = ABe
Bt
Bt
K d ( ABe ) + K s ( Ae ) = 0
Kd B + K s = 0
K
B = ---------s
Kd
Next, find the particular solution.
y p = 0
yp = C
FC = ----Ks
Combine the solutions, and find the remaining unknown.
Ks
---------t
Kd
Fy ( t ) = y p + y h = Ae
+ ----Ks
y(0) = h
0 F
F
h = Ae + ----A = h -----K
Ks
s
Kd ( 0 ) + K s ( C ) = F
Figure 3.19
Ks
--------- t
Kd
F
+ ----Ks
Ks
Kd
Use the general form given below to solve the problem in Figure 3.19 without solving
the differential equation. Assume the system starts at y=-20.
y + 10y = 5
y ( t ) = y 1 + ( y 0 y 1 )e
ans.
Figure 3.20
t
-
y ( t ) = 0.5 19.5e
10t
Drill problem: Developing the final equation using the first-order model
form
3.3.2 Second-order
A second-order system response typically contains two first-order responses, or a
first-order response and a sinusoidal component. A typical sinusoidal second-order
response is shown in Figure 3.21. Notice that the coefficients of the differential equation
include a damping coefficient and a natural frequency. These can be used to develop the
final response, given the initial conditions and forcing function. Notice that the damped
frequency of oscillation is the actual frequency of oscillation. The damped frequency will
be lower than the natural frequency when the damping coefficient is between 0 and 1. If
the damping coefficient is greater than one the damped frequency becomes negative, and
the system will not oscillate because it is overdamped.
= n
y ( t ) = y ss + ( y 0 y ss )e
d = n 1
cos ( d t )
y ss
y0
Figure 3.21
When only the damping coefficient is increased, the frequency of oscillation, and
overall response time will slow, as seen in Figure 3.22. When the damping coefficient is 0
the system will oscillate indefinitely. Critical damping occurs when the damping coeffi-
cient is 1. At this point both roots of the differential equation are equal. The system will
not oscillate if the damping coefficient is greater than or equal to 1.
= 0
= 0.5
(underdamped)
1- = 0.707
= -----2
= 1 (critical)
(overdamped)
Figure 3.22
When observing second-order systems it is more common to use more direct measurements of the response. Some of these measures are shown in Figure 3.23. The rise
time is the time it takes to go from 10% to 90% of the total displacement, and is comparable to a first order time constant. The settling time indicates how long it takes for the system to pass within a tolerance band around the final value. The permissible zone shown is
2%, but if it were larger the system would have a shorter settling time. The period of oscillation can be measured directly as the time between peaks of the oscillation, the inverse is
the damping frequency. (Note: dont forget to convert to radians.) The damped frequency
can also be found using the time to the first peak, as half the period. The overshoot is the
height of the first peak. Using the time to the first peak, and the overshoot the damping
coefficient can be found.
tp
T
e ss
0.02 x
0.02 x
0.9 x
tr
where,
t r = rise time (from 10% to 90%)
t s = settling time (to within 2-5% typ.)
x = total displacement
T, f d = period and frequency - damped
ts
x
0.5 x
b = overshoot
0.1 x
cos ( d t )
d ---tp
(1)
= n
tp
(2)
b- = e
----x
1
= ---------------------------- - 2
------ t p + 1
Figure 3.23
(3)
(4)
Note: We can calculate these relationships using the complex homogenous form, and the
generic second order equation form.
A + 2 n A + n = 0
2
2 n 4 n 4 n
A = ----------------------------------------------------------= j d
2
n = -----
2n
---------------- = = n
2
2
(1)
4 n 4 n
----------------------------------= j d
2
2
4 n 4 n = 4 ( 1 ) d
2
n n = d
2
n 1 = d
(2)
------ 2
------ = 2d
2
2
1
= -------------------2
d
------ + 1
2
d
1--------- + 1
=
2
2
(3)
The time to the first peak can be used to find the approximate decay constant
x ( t ) = C1e
d = ---tp
b xe
(4)
t p
(1)
b
ln ------
x
= -----------------tp
Figure 3.24
cos ( d t + C 2 )
(5)
2.2
y(t)
2.0
t
0
5.0
4.0
Write a function of time for the graph. (Note: measure, using a ruler, to get values.) Find the natural frequency and damping coefficient to develop the differential equation. Using the dashed lines determine the settling time.
Figure 3.25
ans.
t<4
y(t) = 0
t4
y(t) =
Figure 3.26
The basic techniques used for solving first and second-order differential equations
can be applied to higher order differential equations, although the solutions will start to
become complicated for systems with much higher orders. The example in Figure 3.27
shows a fourth order differential equation. In this case the resulting homogeneous solution
yields four roots. The result in this case are two real roots, and a complex pair. The two
real roots result in e-to-the-t terms, while the complex pair results in a damped sinusoid.
The particular solution is relatively simple to find in this example because the non-homogeneous term is a constant.
4 At
d- 4
---x
=
A
e
dt h
Substitute the values into the differential equation and find a value for the unknown.
4 At
3 At
2 At
At
At
A e + 13A e + 34A e + 42Ae + 20e = 0
4
3
2
A + 13A + 34A + 42A + 20 = 0
A = 1, 10, 1 j, 1 + j
xh = C1 e
+ C2 e
10t
t
+ C 3 e cos ( t + C 4 )
Figure 3.27
d- 4
--- dt x p = 0
A = 0.25
The example is continued in Figure 3.28 and Figure 3.29 where the initial conditions are used to find values for the coefficients in the homogeneous solution.
Solve for the unknowns, assuming the system starts at rest and undeflected.
t
10t
t
x ( t ) = C1 e + C 2 e
+ C 3 e cos ( t + C 4 ) + 0.25
0 = C 1 + C 2 + C 3 cos ( C 4 ) + 0.25
(1)
(2)
C 3 cos ( C 4 ) = C 1 C 2 0.25
d- x ( t ) = C e t 10C e 10t C e t cos ( t + C ) C e t sin ( t + C )
---1
2
3
4
3
4
dt h
(3)
0 = C 1 10C 2 C 3 cos ( C 4 ) C 3 sin ( C 4 )
Equations (1) and (3) can be added to get the simplified equation below.
0 = 9C 2 C 3 sin ( C 4 ) + 0.25
C 3 sin ( C 4 ) = 9C 2 + 0.25
(4)
t
10t
t
t
d- 2
---x
(
t
)
=
C
e
+
100C
e
+
C
e
cos
(
t
+
C
)
+
C
e
1
2
3
4
3 sin ( t + C 4 ) +
dt h
t
t
C 3 e sin ( t + C 4 ) C 3 e cos ( t + C 4 )
0 = C1 + 100C 2 + C 3 cos ( C 4 ) + C 3 sin ( C 4 ) + C 3 sin ( C 4 ) C 3 cos ( C 4 )
(5)
0 = C 1 + 100C 2 + 2C 3 sin ( C 4 )
Equations (4) and (5) can be combined.
0 = C 1 + 100C 2 + 2 ( 9C 2 + 0.25 )
0 = 17C 1 + 100C 2 + 0.5
(6)
t
10t
t
t
d- 3
---x ( t ) = C 1 e + ( 1000 )C 2 e
2C 3 e sin ( t + C 4 ) + 2C 3 e cos ( t + C 4 )
dt h
(7)
0 = C 1 + ( 1000 )C 2 2C 3 sin ( C 4 ) + 2C 3 cos ( C 4 )
Figure 3.28
(8)
C 2 = 0.00109
984
1
C 1 = --------- ( 0.00109 ) --C 1 = 0.0242
3
3
Equations (2) and (4) can be combined.
C 3 sin ( C 4 )
9C 2 + 0.25
--------------------------- = --------------------------------------C 3 cos ( C 4 )
C 1 C 2 0.25
9 ( 0.00109 ) + 0.25
C 4 = 0.760
tan ( C 4 ) = -------------------------------------------------------------------------- ( 0.0242 ) ( 0.00109 ) 0.25
Equation (4) can be used the find the remaining unknown.
C 3 sin ( 0.760 ) = 9 ( 0.00109 ) + 0.25
C 3 = 0.377
Figure 3.29
In some cases we will have systems with multiple differential equations, or nonlinear terms. In these cases explicit analysis of the equations may not be feasible. In these
cases we may use other techniques, such as numerical integration, which will be covered
in later chapters.
cases an unstable system may be the objective, such as an explosive device. Simple methods for determining the stability of a system are listed below:
1. If a step input causes the system to go to infinity, it will be inherently unstable.
2. A ramp input might cause the system to go to infinity; if this is the case, the system might not respond well to constant change.
3. If the response to a sinusoidal input grows with each cycle, the system is probably resonating, and will become unstable.
Beyond establishing the stability of a system, we must also consider general performance. This includes the time constant for a first-order system, or damping coefficient
and natural frequency for a second-order system. For example, assume we have designed
an elevator that is a second-order system. If it is under damped the elevator will oscillate,
possibly leading to motion sickness, or worse. If the elevator is over damped it will take
longer to get to floors. If it is critically damped it will reach the floors quickly, without
overshoot.
Engineers distinguish between initial setting effects (transient) and long term
effects (steady-state). The transient effects are closely related to the homogeneous solution
to the differential equations and the initial conditions. The steady-state effects occur after
some period of time when the system is acting in a repeatable or non-changing form. Figure 3.30 shows a system response. The transient effects at the beginning include a quick
rise time and an overshoot. The steady-state response settles down to a constant amplitude
sine wave.
Steady-state
Transient
Note: the transient response is predicted with the homogeneous solution. The
steady state response in mainly predicted with the particular solution,
although in some cases the homogeneous solution might have steady state
effects, such as a non-decaying oscillation.
Figure 3.30
x+x = 5
2
x +x = 5
x + log ( x ) = 5
x + ( 5t )x = 5
Figure 3.31
Consider the differential equation for a 100kg human ejected from an airplane. The
aerodynamic drag will introduce a squared variable, therefore making the equation
non-linear.
2
2
) 2 Mg = My
0.8 Ns
--------- ( y' )
F
=
0.8
(
y
2
y
m
2 2
N100kgy + 0.8 Ns
--------- ( y ) = 100kg9.81 ----2
kg
m
2 2
100kgy + 0.8 Ns
--------- ( y ) = 981N
y
2
Mg
m
2
m- -----s - ( y ) 2 = 981kg ---m100kgy + 0.8kg ---2 2
2
m
s
s
1 2
2
100y + 0.8m ( y ) = 981ms
2
2
3 1
y + 8 10 m ( y ) = 9.81ms
The terminal velocity can be found be setting the acceleration to zero.
2
2
3 1
( 0 ) + 8 10 m ( y ) = 9.81ms
2
9.81ms - = ------------------9.81 - m 2 s 2 = 35.0 m
y = --------------------------------- = 126 km
------3
s
h
3 1
8
10
8 10 m
Figure 3.32
The equation can also be solved using explicit integration, as shown in Figure
3.33. In this case the equation is separated and rearranged to isolate the v terms on the
left, and time on the right. The term is then integrated in Figure 3.34 and Figure 3.35. The
final form of the equation is non-trivial, but contains e-to-t terms, as we would expect.
An explicit solution can begin by replacing the position variable with a velocity variable
and rewriting the equation as a separable differential equation.
1 2
2
dv
1 2
2
100 ------ + 0.8m v = 981ms
dt
dv
2
1 2
100 ------ = 981ms 0.8m v
dt
100
------------------------------------------------dv = dt
2
1 2
981ms 0.8m v
100
-------------------1
0.8m
------------------------------------------- 981 2 2- dv = dt
-------------------- ms + v
1
0.8m
125m
- dv = t + C1
-------------------------------------------2
2 2
v 1226.25m s
125m
- dv
----------------------------------------------------------------m
m
Figure 3.33
Developing an integral
= t + C1
s
s
m
Av A 35.02 ---- + Bv + B 35.02 m
---- = 125m
s
s
m
v ( A + B ) + 35.02 ---- ( A + B ) = 125m
s
A = B
A+B = 0
35.02 m
---- ( A + B ) = 125m
s
125 - s
( ( B ) + B ) = -----------35.02
B = 1.785s
A = 1.785s
1.785s
1.785s - dv = t + C
--------------------------------- + ------------------------------1
v + 35.02 m
v 35.02 m
----
s
s
The integral can then be solved using an identity from the integral table. In this case
the integration constants can be left off because they are redundant with the one on
the right hand side.
1.785s ln v + 35.02 m
---- 1.785s ln v 35.02 m
---- = t + C 1
s
s
m
v + 35.02 ---s- = t + C
1
ln a + bx- + C
1.785s ln --------------------------( a + bx ) dx = ----------------------1
m
b
v 35.02 ---s
t -+C
--------------v + 35.02 m
---1
s- = e 1.785s
--------------------------v 35.02 m
---s
t --------------m
C 1.785s
v + 35.02 ---s- = e 1 e
--------------------------v 35.02 m
---s
Figure 3.34
m
t --------------v + 35.02 ---1.785s
s
---------------------------- = C 2 e
m
v 35.02 ---s
An initial velocity of zero can be assumed to find the value of the integration constant
m
0 --------------0 + 35.02 ---1.785s
s
---------------------------- = C 2 e
m
0 35.02 ---s
1 = C2
This can then be simplified, and the absolute value sign eliminated.
m
t v + 35.02 -----------------s
1.785s
---------------------------- = e
m
v 35.02 ---s
t ---------------
t ---------------
m
m 1.785s
1.785s
v + 35.02 ---- = ve
+ 35.02 ---- e
s
s
t
----------------
---------------
m
m
1.785s
1.785s
--v1
e
=
35.02
e
35.02 ---
+
+
s
s
t
---------------
1.785s
m
1 1
11
0
m
e
1
+
---------------0 = 35.02 ---- +
= ------------- = ---
v = 35.02 ---- ---------------------------
1 + 1
t 2
s 1+1
s
--------------
1.785s
1
+e
t --------------
1.785s
e
1
-
v = 35.02 ---- ----------------------t
s
----------------
1.785s
1 + e
Figure 3.35
As evident from the example, non-linear equations are involved and dont utilize
routine methods. Typically the numerical methods discussed in the next chapter are preferred.
nique to solve the equation, we will need to linearize the model before we can proceed. A
non-linear system can be approximated with a linear equation using the following method.
1. Pick an operating point or range for the component.
2. Find a constant value that relates a change in the input to a change in the output.
3. Develop a linear equation.
4. Use the linear equation for the analysis.
A linearized differential equation can be approximately solved using known techniques as long as the system doesnt travel too far from the linearized point. The example
in Figure 3.36 shows the linearization of a non-linear equation about a given operating
point. This equation will be approximately correct as long as the first derivative doesnt
move too far from 100. When this value does, the new velocity can be calculated.
Assume we have the non-linear differential equation below. It can be solved by linearizing the value about the operating point
Given,
2
y + 4y = 200
y ( 0 ) = 10
We can make the equation linear by replacing the velocity squared term with the
velocity times the actual velocity. As long as the system doesnt vary too much
from the given velocity the model should be reasonably accurate.
y = 200 4y
y ( 0 ) = 200 4 ( 10 ) = 12.65
12.65y + 4y = 20
This system may now be solved as a linear differential equation. If the velocity
(first derivative of y) changes significantly, then the differential equation should
be changed to reflect this.
Homogeneous:
12.65y + 4y = 0
12.65A + 4 = 0
0.316t
y h = Ce
Particular:
yp = A
12.65 ( 0 ) + 4A = 200
Initial conditions:
0.316t
y ( t ) = Ce
+ 50
0
10 = Ce + 50
0.316t
y ( t ) = 40e
+ 50
Figure 3.36
A = 0.316
A = 50
C = 40
0.316 ( 0.1 )
+ 50 = 11.24
12.25y' + 4y = 20
Now recalculate the solution to the differential equation.
Homogeneous:
12.25y + 4y = 0
12.25A + 4 = 0
0.327t
y h = Ce
A = 0.327
Particular:
yp = A
12.25 ( 0 ) + 4A = 200
Initial conditions:
0.327t
y ( t ) = Ce
+ 50
0.1
11.24 = Ce
+ 50
0.316t
y ( t ) = 35.07e
+ 50
A = 50
C = 35.070575
Notice that the values have shifted slightly, and as the analysis progresses the
equations will adjust slowly. Higher accuracy can be obtained using smaller
steps in time.
Figure 3.37
right hand side of the cable is being pulled at a constant rate, while the block is free to
move, only restricted by friction forces and inertia. At the beginning all components are at
rest and undeflected.
x2
k = 0.1
s = 0.3
NK s = 1000 --m
x 1 = 0.1 m
---- t
s
M = 100kg
An FBD and equation can be developed for the system. The friction force will be left as
a variable at this point.
Ks ( x1 x2 )
F x = F F + K s ( x 1 x 2 ) = Mx 2
M = 100kg
Mx 2
N- 0.1 m
F F + 1000 ------ t x 2 = 100kgx2
m
s
M = 100kg
Fx = FF = Mx2
Mx 2
F F = 100kgx 2
100kgx 2 = F F
FF
x 2 = --------------100kg
Figure 3.38
x2
k = 0.1
s = 0.3
NK s = 1000 --m
x 1 = 0.1 m
---- t
s
M = 100kg
An FBD and equation can be developed for the system. The friction force will be left as
a variable at this point.
N- = 981N
N = 100kg9.81 ----kg
static friction
d- x 0 m
------dt 2
s
d- x > 0 m
------dt 2
s
F F = k N = 98.1N
Figure 3.39
The analysis of the system begins by assuming the system starts at rest and undeflected.
In this case the cable/spring will be undeflected with no force, and the mass will be
experiencing static friction. Therefore the block will stay in place until the cable
stretches enough to overcome the static friction.
x2 = 0
x2 = 0
F F = 294.3N
FF
2
m
x 2 + 10s x 2 = 1 ----- t --------------3 100kg
s
2
m-t 294.3N
----------------0 + 10s 0 = 1 ---100kg
3
s
m- t = 294.3kgm
-----------------------1 ---3
2
s
100kgs
t = 2.943s
Therefore the system is static from 0 to 2.943s
Figure 3.40
After motion begins the object will only experience kinetic friction, and continue to
accelerate until the cable/spring becomes loose in compression. This stage of motion
requires the solution of a differential equation.
2
x 2 + 10s x 2 = 0
2
A + 10s
= 0
x h = C 1 sin ( 3.16t + C 2 )
A = 3.16js
xp = A
xp = 0
0 + 10s
Figure 3.41
A = 0.1 m
---s
B = 0.0981m
d- x ( 2.943s ) = 0 m
------dt
s
C 2 = ( 7.889 + n )rad
0.1962
C 1 = --------------------------------------------------= 0.199m
sin ( 9.29988 7.889 )
x ( t ) = 0.199 m sin ( 3.16t 7.889rad ) + 0.1 m
---- t 0.0981m
s
d
----- x ( t ) = 0.199 ( 3.16 )m cos ( 3.16t 7.889rad ) + 0.1 m
---dt
s
Figure 3.42
nI
The equation of motion changes after the cable becomes slack. This point in time can
be determined when the displacement of the block equals the displacement of the
cable/spring end.
m
m
0.1 ---- t = 0.199 m sin ( 3.16t 7.889rad ) + 0.1 ---- t 0.0981m
s
s
0.199 m sin ( 3.16t 7.889rad ) = 0.0981m
t = 3.328s
3.16t 7.889 + n = 0.51549413
x ( 3.328 ) = 0.137m
dm
---x ( 2.333 ) = 0.648 ---dt
s
m
x 2 = 0.981 ----2 t + C 1
s
m
m
0.648 ---- = 0.981 ----2 ( 3.328s ) + C 1
s
s
m
C 1 = 3.913 ---s
0.981
m
m
2
2 s
s
0.981 m
m
2
0.137m = ------------- ---2- ( 3.328s ) + 3.913 ---- ( 3.328s ) + C 2
2
s
s
C 2 = 7.453m
0.981 m 2
m
x 2 ( t ) = ------------- ----2 t + 3.913 ---- t 7.453m
2 s
s
This motion continues until the block stops moving.
m
m
0 = 0.981 ----2 t + 3.913 ---
s
s
t = 3.989s
The solution can continue, considering when to switch the analysis conditions.
Figure 3.43
Figure 3.44
There are a number of elements to the design and analysis of this system, but as
usual the best place to begin is by developing a free body diagram, and a differential equation. This is done in Figure 3.45.
Fy
= F 4K s y 4K d y + Mg = My
My' + 4K d y + 4K s y = F + Mg
4K d y 4K s y
F- + g
y + ------------ + ------------ = ---4K s y 4K d y
M
M
M
4K d y
4K s y
1000N - sin ( 2 ( 2 )t ) + 9.81ms 2
y + --------------------- + --------------------- = --------------------10000Kg 10000Kg
10000Kg
Mg
1
1
2
2
y + 0.0004Kg K d y + 0.0004Kg K s y = 0.1ms sin ( 4 t ) + 9.81ms
Figure 3.45
Using the differential equation, the spring values can be found by assuming the
machine is at rest. This is done in Figure 3.46.
When the system is at rest the equation is simplified; the acceleration and velocity
terms both become zero. In addition, we will assume that the cyclic force is not
applied for the unloaded/loaded case. This simplifies the differential equation by
eliminating several terms.
1
0.0004Kg K s y = 9.81ms
Now we can consider that when unloaded the spring is 0.30m long, and after loading
the spring is 0.25m long. This will result in a downward compression of 0.05m, in
the positive y direction.
1
K s = 491KNm
Figure 3.46
The remaining unknown is the damping coefficient. At this point we have determined the range of motion of the mass. This can be done by developing the particular
solution of the differential equation, as it will contain the steady-state oscillations caused
by the forces as shown in Figure 3.47.
1
1
1
2
2
The particular solution can now be found by guessing a value, and solving for the
coefficients. (Note: The units in the expression are uniform (i.e., the same in each
term) and will be omitted for brevity.)
y = A sin ( 4 t ) + B cos ( 4 t ) + C
y' = 4 A cos ( 4 t ) 4 B sin ( 4 t )
2
16 B + 0.0004K d 4 A + 196A = 0
6
B = A ( 31.8 10 K d + 1.24 )
2
16 A + 0.0004K d ( 4 B ) + 196A = 0.1
2
3.18 10 K d 0.124
B = -------------------------------------------------------------------------------------------------------2
9
3
K d ( 159 10 ) + K d ( 6.2 10 ) + 38.1
C = 9.81ms
Figure 3.47
The particular solution can be used to find a damping coefficient that will give an
overall oscillation of 0.02m, as shown in Figure 3.48. In this case Mathcad was used to
find the solution, although it could have also been found by factoring out the algebra, and
finding the roots of the resulting polynomial.
In the previous particular solution the values were split into cosine and sine components. The magnitude of oscillation can be calculated with the Pythagorean formula.
2
2
A +B
2
2
6
( 0.1 ) + ( 3.18 10 )K d 0.124
magnitude = --------------------------------------------------------------------------------------------------------------2
9
3
K d ( 159 10 ) + K d ( ( 6.2 10 ) ) + 38.1
The design requirements call for a maximum oscillation of 0.02m, or a magnitude of
0.01m.
2
2
6
( 0.1 ) + ( 3.18 10 )K d 0.124
0.01 = --------------------------------------------------------------------------------------------------------------2
9
3
K d ( 159 10 ) + K d ( ( 6.2 10 ) ) + 38.1
A given-find block was used in Mathcad to obtain a damper value of,
magnitude =
K d = 3411N ---sm
Figure 3.48
The values of the spring and damping coefficients can be used to select actual
components. Some companies will design and build their own components. Components
can also be acquired by searching catalogs, or requesting custom designs from other companies.
3.7 SUMMARY
First and second-order differential equations were analyzed explicitly.
First and second-order responses were examined.
The topic of analysis was discussed.
A case study looked at a second-order system.
Non-linear systems can be analyzed by making them linear.
2. The following differential equation was derived for a mass suspended with a spring. At time 0s
the system is released and allowed to drop. It then oscillates. Solve the differential equation to
find the motion as a function of time.
+
N
K s = 100 ---m
y
M = 1Kg
Ks y
FBD:
M
Mg
Fy
= K s y Mg = My
N-
N
m
Kg
N
1 Nm
-------- y + 100 ---- y = 9.81N
2
m
s
KgmKgm-
s
ms
2
2
y + ( 100s )y = 9.81ms
1
y 0 = 0ms
y = 0m
0
3. Solve the following differential equation with the three given cases. All of the systems have a
x + 2n x + n x = y
initial conditions
case 1:
= 0.5
n = 10
case 2:
= 1
n = 10
case 3:
= 2
n = 10
x = 0
x = 0
y = 1
4. Solve the following differential equation with the given initial conditions and draw a sketch of
the first 5 seconds. The input is a step function that turns on at t=0.
Vo = 0
5. Solve the following differential equation with the given initial conditions and draw a sketch of
the first 5 seconds. The input is a step function that turns on at t=0.
Vo = 1
6. a) Write the differential equations for the system below. Solve the equations for x assuming that
the system is at rest and undeflected before t=0. Also assume that gravity is present.
Kd1 = 1 Ns/m
Ks1 = 1 N/m
Kd = 1 Ns/m
M = 1 kg
x1
Kd2 = 1Ns/m
Ks2 = 1 N/m
F=1N
x2
Ks = 1 N/m
F=1N
b) State whether each system is first or second-order. If the system if first-order
find the time constant. If it is second-order find the natural frequency and
damping ratio.
7. Solve the following differential equation with the three given cases. All of the systems have a
sinusoidal input y and start undeflected and at rest.
initial conditions x = 0
x + 2n x + n x = y
x = 0
y = sin ( t )
= 0.5
n = 10
case 1:
case 2:
= 1
n = 10
case 3:
= 2
n = 10
8. A spring damper system supports a mass of 34N. If it has a spring constant of 20.6N/cm, what
is the systems natural frequency?
9. Using a standard lumped parameter model the weight is 36N, stiffness is 2.06*103 N/m and
damping is 100Ns/m. What are the natural frequency (Hz) and damping ratio?
10. What is the differential equation for a second-order system that responds to a step input with
an overshoot of 20%, with a delay of 0.4 seconds to the first peak?
11. A system is to be approximated with a mass-spring-damper model using the following parameters: weight 28N, viscous damping 6Ns/m, and stiffness 36N/m. Calculate the undamped natural frequency (Hz) of the system, the damping ratio and describe the type of response you
would expect if the mass were displaced and released. What additional damping would be
required to make the system critically damped?
Mx + K d x + K s x = F
12. Solve the differential equation below using homogeneous and particular solutions. Assume
the system starts undeflected and at rest.
+ 40 + 20 + 2 = 4
13. What would the displacement amplitude after 100ms for a system having a natural frequency
of 13 rads/sec and a damping ratio of 0.20. Assume an initial displacement of 50mm, and a
steady state displacement of 0mm. (Hint: Find the response as a function of time.)
14. Determine the first order differential equation given the graphical response shown below.
Assume the input is a step function.
x
4
t(s)
0
15. Explain with graphs how to develop first and second-order equations using experimental data.
16. The second order response below was obtained experimentally. Determine the parameters of
the differential equation that resulted in the response assuming the input was a step function.
0.5s
1s
10
t(s)
17. Develop equations (function of time) for the first and second order responses shown below.
x (m)
5m
0m
0.1s
0.2s
0.3s
0.4s
0.5s
0.6s
0.7s
0.8s
t (s)
x (m)
5m
0m
1s
-10m
3s
t (s)
FM FM- M t F
e
x ( t ) = ---------- --- t + -------2
2
B
B
B
2.
homogeneous:
guess
yh = e
At
y h = Ae
At
2 At
A e + ( 100s )e
2
At
2 At
yh = A e
= 0
A = 100s
A = 10 js
y h = C 1 cos ( 10t + C 2 )
particular:
guess
yp = 0
yp = A
2
( 0 ) + ( 100s )A = 9.81ms
2
2
( 100s )A = 9.81ms
2
9.81ms = 0.0981m
A = --------------------2
100s
y p = 0.0981m
initial conditions:
yp = 0
2
C2 = 0
0 = C 1 cos ( 10 ( 0 ) + ( 0 ) ) + 0.0981m
0.0981m = C 1 cos ( 0 )
y ( t ) = ( 0.0981m ) cos ( 10t ) + 0.0981m
C 1 = 0.0981m
3.
5t
case 1:
x ( t ) = 0.0115e
case 2:
x ( t ) = 0.010e
case 3:
x ( t ) = 775 10 e
10t
6 37.32t
10t
+ 0.010
0.0108e
2.679t
+ 0.010
4.
or
V o ( t ) = 8.465e
0.6t
V o ( t ) = 8.465e
0.6t
V 0 ( t ) = 8.331e
0.6t
5.
6.
a)
b)
t
x1 ( t ) = e 1
x ( t ) = 12.485e
x 2 ( t ) = 2e 2
= 0.5
= 1
0.5t
n = 1
7.
5t
case 1:
x ( t ) = 0.00117e
case 2:
x ( t ) = ( 1.96 10 )e
case 3:
x ( t ) = ( 3.5 10 )e
8. 24.37 rad/sec
9. fn=3.77Hz, damp.=.575
10t
37.32t
+ ( 9.9 10 )te
2.679t
( 18 10 )e
10.
x + 8.048x + 77.88x = F ( t )
11.
Given
s
K d = 6N ---m
N
K s = 36 ---m
28N
M = ----------------- = 2.85kg
N
9.81 -----kg
The typical transfer function for a mass-spring-damper systems is,
Ks
F
K d
x + x ------ + x ----- = ---- M
M
M
The second order parameters can be calculated from this.
2
x + x ( 2 n ) + x ( n ) = y ( t )
n =
Ks
----- =
M
N
36 ---m =
---------------2.85kg
kgm
36 ---------2ms - =
---------------2.85kg
12.63s
rad
= 3.55 --------- = 0.6Hz
s
K d
s
----6N --- M
m
= ------------ = --------------------------------------------- = 0.296
2 n
rad
2 ( 3.55 ) --------- 2.85kg
s
2
rad
d = n 1 = 3.39 --------s
If pulled and released the system would have a decaying oscillation about 0.52Hz
A critically damped system would require a damping coefficient of....
K
-----d-
Kd
M
= ------------ = --------------------------------------------- = 1.00
2 n
rad
2 ( 3.55 ) --------- 2.85kg
s
Ns
K d = 20.2 -----m
12.
6 39.50t
( t ) = 66 10 e
3.216e
0.1383t
+ 1.216e
0.3368t
+ 2.00
13.
y ( t ) = 0.05e
2.6t
2.6t
14.
x
4
t(s)
0
1
2
= 1
Given the equation form,
1
x + --- x = A
x = 0
x = 4
So the unknown A can be calculated.
1
A = 4
0 + --- 4 = A
1
1
x + --- x = 4
1
x+x = 4
15.
Key points:
First-order:
find initial final values
find time constant with 63% or by slope
use these in standard equation
Second-order:
find damped frequency from graph
find time to first peak
use these in cosine equation
16.
For the first peak:
t
b----= e p
x
d ---tp
2 0.5
----= e
10
2
ln ------ = 0.5
10
2
= 2 ln ------ = 3.219
10
For the damped frequency:
1
= ---------------------------- - 2
------+1
t p
2
d = ------ = 2
1s
These values can be used to find the damping coefficient and natural frequency
= n
3.219
n = ------------
d = n 1
3.219
2
2 = ------------- 1
2
2 - 2
1 -----------= ------------2
3.219
2 - 2
1
-----------+ 1 = ----2 3.219
1
------------------------------ = 0.4560
2 - 2
----------- 3.219 + 1
3.219
3.219
n = ------------- = ---------------- = 7.059
0.4560
This leads to the final equation using the steady state value of 10
2
x + 2n x + n x = F
2
x + 6.438x + 49.83x = F
( 0 ) + 6.438 ( 0 ) + 49.83 ( 10 ) = F
F = 498.3
17.
t
----------
0.18
x(t) = 51 e
x ( t ) = 10e
0.693 t
cos ( t )
4. NUMERICAL ANALYSIS
Topics:
State variable form for differential equations
Numerical integration with software and calculators
Numerical integration theory: first-order, Taylor series and Runge-Kutta
Using tabular data
A design case
Objectives:
To be able to solve systems of differential equations using numerical methods.
4.1 INTRODUCTION
For engineering analysis it is always preferable to develop explicit equations that
include symbols, but this is not always practical. In cases where the equations are too
costly to develop, numerical methods can be used. As their name suggests, numerical
methods use numerical calculations (i.e., numbers not symbols) to develop a unique solution to a differential equation. The solution is often in the form of a set of numbers, or a
graph. This can then be used to analyze a particular design case. The solution is often
returned quickly so that trial and error design techniques may be used. But, without a symbolic equation the system can be harder to understand and manipulate.
This chapter focuses on techniques that can be used for numerically integrating
systems of differential equations.
d-
--- dt x = Ax + Bu
y = Cx + Du
where,
x = state/output vector (variables such as position)
u = input vector (variables such as input forces)
A = transition matrix relating outputs/states
B = matrix relating inputs to outputs/states
y = non-state value that can be found directly (i.e. no integration)
C = transition matrix relating outputs/states
D = matrix relating inputs to outputs/states
Figure 4.1
Given the FBD shown below, the differential equation for the system is,
F
K d x
x
Mx
Ksx
Fx
= F K d x K s x = Mx
F K d x K s x = Mx
The equation is second-order, so two state variables will be needed. One obvious
choice for a state variable in this equation is x. The other choice can be the velocity, v. Equation (1) defines the velocity variable. The velocity variable can then be
substituted into the differential equation for the system to reduce it to first-order.
x = v
(1)
Mx = F K d x K s x
Mv = F K d v K s x
Ks
Kd
F
v = x --------- + v ---------- + ------
M
M
M
(2)
Equations (1) and (2) can also be put into a matrix form similar to that given in Figure 4.1.
0
1
0
d
x
= K K x + F
----dt v
-------------s- ---------d- v
M
M M
Note: To have a set of differential equations that is solvable, there must be the
same number of state equations as variables. If there are too few equations,
then an additional equation must be developed using an unexploited relationship. If there are too many equations, a redundancy or over constraint must be
eliminated.
Figure 4.2
d-
x
F = F = M ---dt
ans. x = v
v = ----M
0
d- x
0
1
x
---=
+ F
dt v
----0 0 v
M
Figure 4.3
Consider the two cart problem in Figure 4.4. The carts are separated from each
other and the wall by springs, and a force is applied to the left hand side. Free body diagrams are developed for each of the carts, and differential equations developed. For each
cart a velocity state variable is created. The equations are then manipulated to convert the
second-order differential equations to first-order state equations. The four resulting equations are then put into the state variable matrix form.
x1
x2
K s1
K s2
M1
K s1 ( x 1 x 2 )
F
M1
K s1 ( x 1 x 2 )
M2
M1 x1
M2
K s2 ( x 2 )
M2 x2
Fx
= F K s1 ( x 1 x 2 ) = M 1 x 1
M 1 x 1 + K s1 x 1 K s1 x 2 = F
x1 = v1
M 1 v 1 + K s1 x 1 K s1 x 2 = F
F- K
s1 + K
s1
v 1 = ---------------x
---------x
1
M1 M1
M1 2
+
Fx
(2)
= K s1 ( x 1 x 2 ) K s2 ( x 2 ) = M 2 x 2
M 2 x 2 + ( K s1 + K s2 )x 2 K s1 x 1 = 0
x2 = v2
(3)
M 2 v 2 + ( K s1 + K s2 )x 2 K s1 x 1 = 0
K s1
K s1 + K s2
v 2 = ---------x
-------------------------- x 2
(4)
M2
M2 1
K
v2
v2
s1
s1
s2
0
--------- 0 ----------------------------- 0
M2
M2
Figure 4.4
(1)
x
Kd
M
Ks
ans.
x = v
Ks
Kd
v = v ------ + x -----
M
M
Figure 4.5
x1
x2
K d1
F
M1
K s1
K s2
M2
ans. x = v
1
1
x2 = v2
K d1
K s1
K d1
K s1
F
v 1 = v 1 ------------ + x 1 ----------- + v 2 --------- + x 2 -------- + ------M1
M1
M1
M1
M1
K d1
K s1 K s2
K d1
K s1
M2
M2
M2
Figure 4.6
q = 2y
Step 3: solve the equation using the dummy variable, then solve for y as an output eqn.
q + 5u
q = 2y
y = --------------3
Figure 4.7
At other times it is possible to eliminate redundant terms through algebraic manipulation, as shown in Figure 4.8. In this case the force on both sides of the damper is the
same, so it is substituted into the equation for the cart. But, the effects on the damper must
also be integrated, so a dummy variable is created for the integration. An output equation
was created to calculate the value for x1.
x1
F
x2
Kd
M
Kd ( x1 x2 )
Fx
= F Kd ( x1 x2 ) = 0
Kd ( x 1 x 2 ) = F
(1)
q = x1 x2
Kd ( q ) = F
Kd ( x1 x2 )
Mx 2
Fq = -----Kd
x1 = x2 + q
M
(2)
(3)
= K d ( x 1 x 2 ) = Mx2
F = Mx 2
x2 = v2
Fx
Fv 2 = ---M
(4)
(5)
The state equations (2, 4, 5) can be put in matrix form. The output equation (2) can
also be put in matrix form.
F-----q
q
Kd
0 0 0
d
x
x
----- 2 = 0 0 1 2 +
0
dt
v2
0 0 0 v2
F---M
Figure 4.8
q
x1 = 1 1 0 x2 + 0
v2
K d y
Ks y
Kd
M
y
M
My
d-
d- 2
--- ---F
=
K
y
K
y
=
M
y
d dt
s
dt y
d
d 2
F + K d ----- y + K s y + M ----- y = 0
dt
dt
F
Step 2: We need to identify state variables. In this case the height is clearly a defining variable. We will also need to use the vertical velocity, because the acceleration is a second
derivative (we can only have first derivatives). Using the height, y, and velocity, v, as state
variables we may now proceed to rewriting the equations. (Note: this is just an algebraic
trick, but essential when setting up these matrices.)
Step 3:
d-
---y = v
dt
F Kd v K s y
d-
---v = -------------------------------------- dt
M
0
1
F= K K y + 0 ---S
1 M
---------- ---------d- v
d-
---v
M M
dt
Figure 4.9
Figure 4.10 shows the method for solving state equations on a TI-86 graphing calculator. (Note: this also works on other TI-8x calculators with minor modifications.) In the
example a sinusoidal input force, F, is used to make the solution more interesting. The
next step is to put the equation in the form expected by the calculator. When solving with
the TI calculator the state variables must be replaced with the predefined names Q1, Q2,
etc. The steps that follow describe the button sequences required to enter and analyze the
equations. The result is a graph that shows the solution of the equation. Points can then be
taken from the graph using the cursors. (Note: large solutions can sometimes take a few
minutes to solve.)
First, we select some parameter values for the equations of Figure 4.9. The input force
will be a decaying sine wave.
d-
---y = vy
dt
F Kd vy Ks y
d-
0.5t
---v y = ---------------------------------------- = 4e
sin ( t ) 2v y 5y
dt
M
Next, the calculator requires that the state variables be Q1, Q2, ..., Q9, so we replace y
with Q1 and v with Q2.
Q1' = Q2
Q2' = 4e
0.5t
Now, we enter the equations into the calculator and solve. To do this roughly follow
the steps below. Look at the calculator manual for additional details.
1. Put the calculator in differential equation mode
[2nd][MODE][DifEq][ENTER]
2. Go to graph mode and enter the equations above [GRAPH][F1]
3. Set up the axis for the graph [GRAPH][F2] so that time and the xaxis is from 0 to 10 with a time step of 0.5, and the y height is from
+3 to -3.
4. Enter the initial conditions for the system [GRAPH][F3] as Q1=0,
Q2=0
5. Set the axis [GRAPH][F4] as x=t and y=Q
6. (TI-86 only) Set up the format [GRAPH][MORE][F1][FldOff][ENTER]
7. Draw the graph [GRAPH][F5]
8. Find points on the graph [GRAPH][MORE][F4]. Move the left/right
cursor to move along the trace, use the up/down cursor to move
between traces.
Figure 4.10
First, we select some parameter values for the equations of Figure 4.9. The input force
will be a decaying sine wave.
d-
---y = vy
dt
F Kd vy Ks y
0.5t
d-
---v y = -----------------------------------------= 4e
sin ( t ) 2v y 5y
dt
M
Next, the calculator requires that the state variables be Q1, Q2, ..., Q9, so we replace y
with Q1 and v with Q2.
Q1' = Q2
Q2' = 4e
0.5t
Now, we enter the equations into the calculator and solve. To do this roughly follow
the steps below. Look at the calculator manual for additional details.
1. Put the calculator in differential equation mode
[2nd][MODE][DifEq][ENTER]
2. Go to graph mode and enter the equations above [GRAPH][F1]
3. Set up the axis for the graph [GRAPH][F2] so that time and the xaxis is from 0 to 10 with a time step of 0.5, and the y height is from
+3 to -3.
4. Enter the initial conditions for the system [GRAPH][F3] as Q1=0,
Q2=0
5. Set the axis [GRAPH][F4] as x=t and y=Q
6. (TI-86 only) Set up the format [GRAPH][MORE][F1][FldOff][ENTER]
7. Draw the graph [GRAPH][F5]
8. Find points on the graph [GRAPH][MORE][F4]. Move the left/right
cursor to move along the trace, use the up/down cursor to move
between traces.
Figure 4.11
State equations can also be solved in Mathcad using built-in functions, as shown in
Figure 4.12. The first step is to enter the state equations as a function, D(t, Q), where t
is the time and Q is the state variable vector. (Note: the equations are in a vector, but it is
not the matrix form.) The state variables in the vector Q replace the original state variables in the equations. The rkfixed function is then used to obtain a solution. The arguments for the function, in sequence are; the state vector, the start time, the end time, the
number of steps, and the state equation function. In this case the 10 second time interval is
divided into 100 parts each 0.1s in duration. This time is chosen because of the general
response time for the system. If the time step is too large the solution may become unstable and go to infinity. A time step that is too small will increase the computation time marginally. When in doubt, run the calculator again using a smaller time step.
Figure 4.12
Note: Notice that for the TI calculators the variables start at Q1, while in Mathcad the
arrays start at Q0. Many students encounter problems because they forget this.
y(t + h)
d
y ( t + h ) y ( t ) + h ----- y ( t )
dt
d---y( t)
dt
y(t)
t
Figure 4.13
t+h
The example in Figure 4.14 shows the solution of Newtons equation using Eulers
method. In this example we are determining velocity by integrating the acceleration
caused by a force. The acceleration is put directly into Eulers equation. This is then used
to calculate values iteratively in the table. Notice that the values start before zero so that
initial conditions can be used. If the system was second-order we would need two previous
values for the calculations.
d
v ( t + h ) = v ( t ) + h ----- v ( t )
dt
F(t)
v ( t + h ) = v ( t ) + h ----------
M
We can now use the equation to estimate the system response. We will assume that the
system is initially at rest and that a force of 1N will be applied to the 1kg mass for
4 seconds. After this time the force will rise to 2N. A time step of 2 seconds will be
used.
i
t (sec)
F (N) d/dt vi
vi
-1
0
1
2
3
4
5
6
7
8
-2
0
2
4
6
8
10
12
14
16
0
1
1
2
2
2
2
2
2
2
0
0
2
4
8
12
16
20
etc
Figure 4.14
0
1
1
2
2
2
2
2
etc
x + 0.1x = 5
ans.
x ( t + h ) = x ( t ) + h ( 0.1x ( t ) + 5 )
x ( 10 ) =
Figure 4.15
An example of solving the previous example with a traditional programming language is shown in Figure 4.16. In this example the results will be written to a text file
out.txt. The solution iteratively integrates from 0 to 10 seconds with time steps of 0.1s.
The force value is varied over the time period with if statements. The integration is done
with a separate function.
h = 0.1,
M = 1.0,
F;
FILE
*fp;
double v,
t;
if( ( fp = fopen("out.txt", "w")) != NULL){
v = 0.0;
for( t = 0.0; t < 10.0; t += h ){
if((t >= 0.0) && (t < 4.0)) F = 1.0;
if(t > 4.0) F = 2.0;
v = step(v, h, F/M);
fprintf(fp, "%f, %f, %f\n", t, v, F, M);
}
}
fclose(fp);
}
double step(double v, double h, double slope){
double
v_new;
v_new = v + h * slope;
return v_new;
}
Figure 4.16
Figure 4.17
The program below is for Scilab (a Matlab clone). The state variable function is
defined first. This is followed by a definition of the parameters to be used for the numerical integration. Finally the function is integrated with rectangular, trapezoidal and Simpsons rule forms.
//
//
//
//
//
//
//
//
//
first_order.sce
A first order integration of an accelerating mass
To run this in Scilab use 'File' then 'Exec'.
by: H. Jack Sept., 16, 2002
// initial conditions
// Set the time length and step size for the integration
steps = 100;
t_start = 1;
t_end = 100;
h = (t_end - t_start) / steps;
Figure 4.18
//
// Loop for integration
//
for i=1:steps,
X = [X ; X($,:) + h*f(X, i*h)];
end
printf("The value at the end of first order integration is (x, v) = (%f, %f)\n", ...
X($,1), ...
X($,2));
//
// Explicit equation
//
function x=position(x0, v0, a0, t)
x = (0.5 * a0 * t^2) + (v0 * t) + x0;
endfunction
function v=velocity(v0, a0, t)
v = (a0 * t) + v0;
endfunction
printf("The value with integration is (x, v) = (%f, %f)\n", ...
position(x0, v0, force/mass, t_end), ...
velocity(v0, force/mass, t_end));
//
//
//
//
//
//
//
//
//
//
Figure 4.19
d
1 2 d 2
1 3 d 3
1 4 d 4
x ( t + h ) = x ( t ) + h ----- x ( t ) + ----- h ----- x ( t ) + ----- h ----- x ( t ) + ----- h ----- x ( t ) +
dt
2!
dt
3!
dt
4!
dt
Figure 4.20
An example of the application of the Taylor series is shown in Figure 4.21. Given
the differential equation, we must first determine the derivatives and substitute these into
Taylors equation. The resulting equation is then used to iteratively calculate values.
20t 3
x x = 1 + e
+t
Given
20t 3
d
+t +x
We can write, ----- x = 1 + e
dt
20t
2
d- 2
---+ 3t
dt x = e
20t
d- 3
---x
=
e
+ 6t
dt
In the Taylor series this becomes,
x(t + h) = x( t) + h(1 + e
Thus
20t
3
1- h 2 ( e t + 3t 2 ) + ---1- h 3 ( e t + 6t
+ t + x ) + ---2!
3!
x0 = 0
t (s)
x(t)
h = 0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
0
Figure 4.21
Recall that the state variable equations are first-order equations. But, to obtain
accuracy the Taylor method also requires higher order derivatives, thus making is unsuitable for use with state variable equations.
F 1 = hf ( t, x )
F
F 2 = hf t + h---, x + ------1
2
2
F
F 3 = hf t + h---, x + ------2
2
2
F 4 = hf ( t + h, x + F 3 )
x ( t + h ) = x ( t ) + 1--- ( F1 + 2F 2 + 2F 3 + F 4 )
6
where,
x = the state variables
f = the differential function or (d/dt) x
t = current point in time
h = the time step to the next integration point
Figure 4.22
d- x = v
---dt
y = 2 (assumed input)
d- v = 3 + 4v + 5y
---dt
d- x = 0 1 x + 0 0 y
---dt v
0 4 v
5 3 1
v0 = 1
x0 = 3
h = 0.1
3 + 0.1
------0
1
2
0
0
2
= 0.185
F 2 = 0.1
+
5 3 1
2.04
0 4 1 + 1.7
------
3 + 0.185
------------0
1
2
0
0
2
= 0.202
F 3 = 0.1
+
5 3 1
2.108
0 4 1 + 2.04
---------
xi
+ 1--- ( F1 + 2F 2 + 2F 3 + F 4 )
6
vi
Figure 4.23
d 2
F = M ----- x
dt
Figure 4.24
use,
x(0) = 1
v(0) = 2
h = 0.5 s
F = 10
M=1
Kd
1
K d = 10Nsm
1
K s = 10Nm
M = 1kg
Ks
Figure 4.25
y 0 = 1m
y 0 = 0m
The program in Figure 4.26 and Figure 4.28 is used to perform a Runge-Kutta integration of a mass-spring-damper system. The main program loops through the time steps
and writes the value to a file. The step function performs one timestep integration for a
second order Runge-Kutta integration. It uses the functions add and multiply to manipulate the state matrix. The derivative function updates the state matrix with the new
derivative values.
#define
#define
#define
#define
#define
SIZE
Ks
Kd
Mass
Force
int main(){
FILE
*fp;
double h = 0.001;
double t;
int
j = 0;
double X[SIZE];// create state variable list
X[0] = 0;// set initial condition for x
X[1] = 0;// set initial condition for v
if( ( fp = fopen("out.txt", "w")) != NULL){
fprintf(fp, "
t(s)
x
v \n\n");
for( t = 0.0; t < 50.0; t += h ){
step(t, h, X);
if(j == 0) fprintf(fp, "%9.5f %9.5f %9.5f\n", t, X[0], X[1]);
j++; if(j >= 10) j = 0;
}
}
fclose(fp);
}
Figure 4.26
/* First order integration done here (could be replaced with runge kutta)*/
void step(double t, double h, double X[]){
double tmp[SIZE],
dX[SIZE],
F1[SIZE],
F2[SIZE],
F3[SIZE],
F4[SIZE];
/* Calculate F1 */
derivative(t, X, dX);
multiply(h, dX, F1);
/* Calculate F2 */
multiply(0.5, F1, tmp);
add(X, tmp, tmp);
derivative(t+h/2.0, tmp, dX);
multiply(h, dX, F2);
/* Calculate F3 */
multiply(0.5, F2, tmp);
add(X, tmp, tmp);
derivative(t+h/2.0, tmp, dX);
multiply(h, dX, F3);
/* Calculate F4 */
add(X, F3, tmp);
derivative(t+h, tmp, dX);
multiply(h, dX, F4);
/* Calculate the weighted sum */
add(F2, F3, tmp);
multiply(2.0, tmp, tmp);
add(F1, tmp, tmp);
add(F4, tmp, tmp);
multiply(1.0/6.0, tmp, tmp);
add(tmp, X, X);
}
/* State Equations Calculated Here */
void derivative(double t, double X[], double dX[]){
dX[0] = X[1];
dX[1] = (-Ks/Mass)*X[0] + (-Kd/Mass)*X[1] + (Force/Mass);
}
/* A subroutine to add vectors to simplify other equations */
void add(double X1[], double X2[], double R[]){
for(int i = 0; i < SIZE; i++) R[i] = X1[i] + X2[i];
}
/* A subroutine to multiply a vector by a scalar to simplify other equations*/
void multiply(double X, double V[], double R[]){
for(int i = 0; i < SIZE; i++) R[i] = X*V[i];
}
Figure 4.27
//
//
//
//
runge_kutta.sce
A first order integration of an accelerating mass
To run this in Scilab use 'File' then 'Exec'.
by: H. Jack Sept., 15, 2003
// initial conditions
Figure 4.28
details such as time constants and damped frequencies can be obtained by the same methods used for experimental analysis. In addition to these methods there is a technique that
can determine the steady-state response of the system.
0
1
0
d
x
x +
----=
Ks K d
Fdt v
--- ------ ------- v
M
M M
0
1
0
0 =
x +
Ks Kd
F---0
------ ------- v
M
M M
0
1
0
x =
Ks Kd
F
---- ------ ------- v
M
M M
1
F Kd
F-
--- ----- ------ M
M M
Fx = ------------------------------ = ------------ = -----K
KS
0
1
-----s-
M
Ks Kd
------ ------M M
Figure 4.29
0 0
Ks F
------ ----M M
0 - = 0
v = ------------------------------ = ----------K
0
1
-----s-
M
Ks Kd
------ ------M M
y(t)
yi + 1
yi
yi 1
ti 1
ti
ti + 1
ti
yi + yi 1
T
------------------------ ( t t
y
(
t
)
d
t
t
i
i i 1 ) = --2- ( y i + y i 1 )
2
i1
y i y i 1
y i + 1 y i
d---y ( t i ) ------------------------ = ------------------------- = --1- ( y i y i 1 ) = --1- ( y i + 1 y i )
dt
T
T
ti t i 1
ti + 1 ti
--1- ( y i + 1 y i ) --1- ( y i y i 1 )
2 yi + yi 1 + yi + 1
2
d-
T
T
--------------------------------------------------------------------------
=
----------------------------------------------------y
(
t
)
i
dt
T
2
T
Figure 4.30
An example of numerical integration using Scilab is given in Figure 4.31 and Figure 4.32.
//
//
//
//
//
//
//
//
//
integrate.sce
A simple program to integrate a function
To run this in Scilab use 'File' then 'Exec'.
by: H. Jack Sept., 9, 2002
//
// Loop for rectangular integration
//
total = 0; // set the initial sum to zero
for i=0:steps,
x = x_start + i * x_delta;
total = total + f(x);
end
total = total * x_delta;
printf("Rectangular integration value %f\n", total);
Figure 4.31
//
// Loop for Simpson's rule integration
//
total = 0; // set the initial sum to zero
even = 0;
for i=0:steps,
x = x_start + i * x_delta;
if i == 0 then
total = total + f(x);
elseif i == steps then
total = total + f(x);
else
even = even + 1;
if even > 1 then
total = total + 4 * f(x);
even = 0;
else
total = total + 2 * f(x);
end
end
end
total = total * x_delta / 3;
printf("Simpsons rule integration value %f\n", total);
Figure 4.32
ments are greater than or equal to zero, or off (a value of 0) when the argument is negative.
Examples of the use of switching functions are shown in Figure 4.33. By changing the values of the arguments we can change when a function turns on or off.
u(t)
1
t
u ( t )
1
t
u(t 1)
1
t
1
u(t + 1)
1
-1
u(t + 1) u(t 1)
1
-1
Figure 4.33
t
1
These switching functions can be multiplied with other functions to create a complex function by turning parts of the function on or off. An example of a curve created
with switching functions is shown in Figure 4.34.
f(t)
5
t
seconds
0
f ( t ) = 5t ( u ( t ) ( t 1 ) ) + 5 ( u ( t 1 ) u ( t 3 ) ) + ( 20 5t ) ( u ( t 3 ) u ( t 4 ) )
or
f ( t ) = 5tu ( t ) 5 ( t 1 )u ( t 1 ) 5 ( t 3 )u ( t 3 ) + 5 ( t 4 )u ( t 4 )
Figure 4.34
The unit step switching function is available in Mathcad and makes creation of
complex functions relatively trivial. Step functions are also easy to implement when writing computer programs, as shown in Figure 4.35.
5.0 *
- 5.0
- 5.0
+ 5.0
t
*
*
*
* u(t)
(t - 1.0) * u(t - 1.0)
(t - 3.0) * u(t - 3.0)
(t - 4.0) * u(t - 4.0);
return f;
}
Figure 4.35
valve angle
(deg.)
flow rate
(gpm)
0
10
20
30
40
50
60
70
80
90
0
0.1
0.4
1.2
2.0
2.3
2.4
2.4
2.4
2.4
Figure 4.36
Using tables of values to interpolate numerical values using the lever law
The subroutine in Figure 4.37 was written to return the numerical value for the
data table in Figure 4.36. In the subroutine the tabular data is examined to find the interval
that the flow rate value falls inside. Once this is found the valve angle is calculated as the
ratio between the two known values.
#define
double
SIZE
10;
data[SIZE][2] = {{0.0,
{10.0,
{20.0,
{30.0,
{40.0,
{50.0,
{60.0,
{70.0,
{80.0,
{90.0,
0.0},
0.1},
0.4},
1.2},
2.0},
2.3},
2.4},
2.4},
2.4},
2.4}};
Figure 4.37
The datapoints below might have been measured for the horsepower of an internal
combustion engine on a dynamometer.
S (RPM)
P (HP)
1000
4000
6000
105
205
110
In this case there are three datapoints, so we can fit the curve with a second (3-1)
order polynomial. The major task is to calculate the coefficients so that the
curve passes through all of the given points.
2
P ( S ) = AS + BS + C
Data values can be substituted into the equation,
2
P ( 1000 ) = A1000 + B1000 + C = 105
2
P ( 4000 ) = A4000 + B4000 + C = 205
2
P ( 6000 ) = A6000 + B6000 + C = 110
This can then be put in matrix form to find the coefficients,
2
1000 1000 1 A
105
=
2
205
4000 4000 1 B
110
C
2
6000 6000 1
1
A
1000000 1000 1
105
=
B
16000000 4000 1
205
C
36000000 6000 1
110
8
7
7
A
6.667x10
1.667x10
1.000x10
105
B =
4
3
4 205
6.667x10
1.167x10
5.000x10
C
110
1.600
1.000
0.400
5
A
1.617x10
B =
0.114
C
7.000
5 2
P ( S ) = ( 1.617x10 )S + 0.114S + 7.000
Figure 4.38
The order of the polynomial should match the number of points. Although, as the
number of points increases, the shape of the curve will become less smooth. A common
way for dealing with this problem is to fit the spline to a smaller number of points and then
verify that it matches the remaining points, or use a least squares method to find the best
approximation.
x
F
2
20x
Figure 4.39
Fx
2
= F 20x = Mx
x = v
(1)
(2)
yc
Mc
K s1
Kd
N
K s1 = 50000 ---m
N
K s2 = 200000 ---m
M c = 400Kg
M t = 20Kg
Mt
yt
L = yc yt
L
(cm)
Kd
(Ns/m)
30
0
-20
-30
1000
1600
2000
2250
K s2
yr
Figure 4.40
For our purposes we will focus only on the translation of the tire, and ignore its
rotational motion. The differential equations describing the system are developed in Figure 4.41.
Mc yc
yc
Mc
K s1 ( y t y c ) K d ( y t y c )
= K s1 ( y t y c ) + K d ( y t y c ) M c g = M c y c
M c y c = ( K s1 )y t + ( K s1 )y c + ( K d )y t + ( K d )y c M c g
d- y = v
---c
dt c
d
----- y t = v t
dt
K s1
K s1
Kd
K d
-------------------------vc =
y +
y + v + ---------- v + ( g )
M c t M c c M c t M c c
(1)
(2)
(3)
Kd ( yt yc )
K s1 ( y t y c )
Mt
Mt yt
K s2 ( y r y t )
= K s2 ( y r y t ) K s1 ( y t y c ) K d ( y t y c ) M t g = M t y t
M t y t = ( K s2 )y r + ( K s2 K s1 )y t + ( K s1 )y c + ( K d )y t + ( K d )y c M t g
K s2
K s2 K s1
K s1
K d
K d
Mt
Mt
M t
Figure 4.41
(4)
The damping force must be converted from a tabular form to equation form. This
is done in Figure 4.42.
K d ( L ) = AL + BL + CL + D
L
(cm)
Kd
(Ns/m)
30
0
-20
-30
1000
1600
2000
2250
The four data points can now be written in equation form, and then put into matrix form.
3
1
1
1
1
A
1000
B = 1600
C
2000
D
2250
The matrix can be solved to find the coefficients, and the final equation written.
A
2778
B = 277.8
C
1833
D
1600
3
The system is to be tested for overall deflection when exposed to obstacles on the
road. For the initial conditions we need to find the resting heights for the tire and car body.
This can be done by setting the accelerations and velocities to zero, and finding the resulting heights.
The initial accelerations and velocities are set to zero, assuming the car has settled to
a steady state height. This then yields equations that can be used to calculate the
initial deflections. Assume the road height is also zero to begin with.
K s1
K s1
Kd
K d
0 = -------- y t + ----------- y c + ------- 0 + --------- 0 + ( g )
Mc
Mc
Mc
Mc
Mc
y c = y t g -------K s1
K s2
K s2 K s1
K s1
Kd
Kd
0 = -------- 0 + -------------------------- y t + -------- y c + --------- 0 + ------ 0 + ( g )
Mt
Mt
Mt
Mt
Mt
gM t = ( K s2 K s1 )y t + ( K s1 )y c
Mc
gM t = ( K s2 K s1 )y t + ( K s1 ) y t g --------
K s1
Mc + Mt
Mc
y c = g -------------------- g -------K s2
K s1
Figure 4.43
Mc + Mt
y t = g -------------------K s2
Mc + Mt Mc
y c = g -------------------- + --------
K s2
K s1
The resulting calculations can then be written in a computer program for analysis,
as shown in Figure 4.44.
#include <stdio.h>
#include <math.h>
#define
#define
#define
#define
#define
SIZE
y_c
y_t
v_c
v_t
4
0
1
2
3
Ks1
Ks2
Mc
Mt
grav
// number of steps
// define step size
Figure 4.44
=
=
=
=
%f \n",
Figure 4.45
This program was then used to test various design cases by selecting input types
for changes in the road height, and then calculating how the tire and vehicle heights would
change as a result. Some of these results are seen in Figure 4.46. These results were
obtained by running the program, and then graphing the results in a spreadsheet program.
The input of zero for the road height was used to test the program. As shown the height of
the vehicle changes, indicating that the initial height calculations are correct, and the
model is stable. The step function shows some oscillations that settle out to a stable final
value. The oscillation is relatively slow, and is fully transmitted to the automobile. The
ramp function shows that the car follows the rise of the slope with small transient effects
at the start.
y r ( t ) = 0.0
y r ( t ) = 0.2m
0.3
0
1
15
22
29
36
43
50
57
64
71
78
85
92
99
0.25
-0.02
0.2
-0.04
0.15
Series1
-0.06
Series2
0.1
Series
Series
0.05
0
-0.08
15
22
29
36
43
50
57
64
71
78
85
92
99
-0.05
-0.1
-0.1
-0.12
-0.15
y r ( t ) = 0.2m sin ( t )
y r ( t ) = ( 0.2m )t
0.3
2.5
0.2
2
0.1
1.5
0
1
15
22
29
36
43
50
57
64
71
78
85
92
99
S eries1
S eries2
Series
Series
-0.1
0.5
-0.2
-0.3
-0.4
15
22
29
36
43
50
57
64
71
78
85
92
99
-0.5
Figure 4.46
4.8 SUMMARY
State variable equations are used to reduced to first order differential equations.
First order equations can be integrated numerically.
Higher order integration, such as Runge-Kutta increase the accuracy.
Switching functions allow functions terms to be turned on and off to provide
more complex function.
Tabular data can be used to get numerical values.
x + 2x + 3x + 5y = 3
y + y + 6y + 9x = sin ( t )
2. a) Put the differential equations given below in state variable form.
b) Put the state equations in matrix form
y 1 + 2y 1 + 3y 1 + 4y 2 + 5y 2 + 6y 3 + t + F = 0
y 1 + 12y 2 + 13y 3 = 0
y 1, y 2, y 3 = outputs
F = input
Ks
4. The system below is comprised of two masses. There is viscous damping between the masses
and between the bottom mass and the floor. The masses are also connected with a cable that is
run over a massless and frictionless pulley. Write the differential equations for the system, and
put them in state variable form.
x1
Ks
M1
B1
x2
M2
B2
5. Do a first order numerical integration of the derivative below from 0 to 10 seconds in one second steps. Assume the system starts undeflected. Write the equation for each time step.
d2
---x ( t ) = 5( t 4 )
dt
6. Given the differential equation below integrate the values numerically for the first ten seconds
with 1 second steps. Assume the initial value of x is 1. You may use first order or Runge Kutta
integration.
x + 0.25x = 3
7. Write a Scilab program to calculate the area under the function below using a numerical
method, such as Simpsons rule. Test it by using the range from x=1 to x=1.2. Assume the sine
function is in radians
sin x
f ( x ) = 5x + 2 ln ----------
x
8. Write a computer program in a language of your choice (C/C++, Java, Scilab script, etc.) that
will numerically integrate the differential equation below.
+ 3 + 9 = 10
9. Given the following differential equation and initial conditions, draw a sketch of the first 5 seconds of the output response. The input is a step function that turns on at t=0. Use at least two
different methods, and compare the results.
initial conditions V i ( t 0 ) = 5V
0.5V o + 0.6V o + 2.1V o = 3V i + 2
V o ( 0 ) = 0V
V o ( 0 ) = 1V
10. a) For the mass-spring-damper system below solve the differential equation as a function of
time. Assume the system starts at rest and undeflected. b) Also solve the problem using your
calculator (and state equations) to verify your solution. Sketch the results.
x
Kd
K s = 10
K d = 10
F
M = 10
M
Ks
F = 10
11. The mechanical system below is a mass-spring-damper system. A force F of 100N is applied
to the 10Kg cart at time t=0s. The motion is resisted by the spring and damper. The spring
coefficient is 1000N/m, and the damping coefficient is to be determined. Follow the steps
below to develop a solution to the problem. Assume the system always starts undeflected and
at rest.
a) Develop the differential equation for the system.
b) Solve the differential equation using damping coefficients of 100Ns/m and
10000Ns/m. Draw a graph of the results.
c) Develop the state equations for the system.
d) Solve the system with a first order numerical analysis using Scilab for damping
coefficients of 100Ns/m and 10000Ns/m. Draw a graph of the results.
e) Solve the system with a Runge Kutta numerical analysis using Scilab for damping coefficients of 100Ns/m and 10000Ns/m. Draw a graph of the results.
f) Write a computer program (in C, Java or Fortran) to do the Runge Kutta numerical integration in step e). Draw a graph of the results.
g) Compare all of the solutions found in the previous steps.
h) Select a damper value to give an overall system damping coefficient of 1. Verify
the results by numerically integrating.
x
Kd
F
M
Ks
12. For the mechanism illustrated in the figure below the values are Ks1=Ks2=100N/m,
M1=M2=1kg, F=1N. Assume that the system starts at rest, and the springs are undeformed initially.
x1
x2
K s1
M1
K s2
M2
= v
= u
= 2v 3x 5y + 3
= u 6y 9x + sin t
x
0
d- y
---= 0
dt v
3
u
9
0
0
5
6
1
0
2
0
0
1
0
1
x
0
y + 0
v
3
u
sin t
2.
y1
y2
y3
v1
v2
v3
= v1
= v2
= v3
= 2v 1 3y 1 4v 2 5y 2 6v 3 t F
v 13
= -------1- ------ v 3
12 12
v1 7
8
9
11
5
= -------- ------ y 1 ------ v 2 ------ y 2 ------ y 3 ------ cos ( 5t )
10
10
10
10
10 10
y1
y2
d- y 3
---=
dt v
1
v2
v3
0 0
0 0
0 0
3 5
0
0
0
0
7 9 11
------ ------ -----10 10 10
1
0
0
2
1
-----12
1
-----10
0 0
1 0
0 1
4 6
13
0 -----12
8
------ 0
10
y1
y2
y3
v1
v2
v3
0
0
0
+ tF
0
5
------ cos t
10
3.
x
x = v
Kd x
Ks x
F
M
K d
Ks
F
4.
+y
+x
T
Ks x 1
M1
= T Ks x1 B1 ( x 1 x2 ) = M1 x1
v1 = x1
v2 = x2
B1 ( x1 x2 )
K
B
B
T
B1 ( x1 x2 )
T
M2
= T + B 1 ( x1 x2 ) + F B2 x2 = M2 x2
B1
B1 B 2
T+F
B2 x2
5.
h = 1
x ( t + h ) = x ( t ) + h5 ( t 4 )
0
1
2
3
4
5
6
7
8
9
10
0
80
125
145
150
150
155
175
220
300
425
6.
x' = 3 0.25x
d
x ( t + h ) = x ( t ) + h ----- x ( t )
dt
0
1
2
3
4
5
6
7
8
9
10
1
3.75
5.81
7.36
8.52
9.39
10.0
10.5
10.9
11.2
11.4
2.75
2.06
1.55
1.16
0.870
0.652
0.489
0.367
0.275
0.206
x ( t + h ) = x ( t ) + 1 ( 3 0.25x ( t ) )
x ( t + h ) = 0.75x ( t ) + 3
7.
// integrate.sce - A simple program to integrate a function
// To run this in Scilab use 'File' then 'Exec'.
// by: H. Jack Sept., 9, 2002
// define the function
function foo=f(x)
foo = 5 * x + 2 * log(sin(x) / x);
endfunction
// Set the time length and step size
steps = 10;
x_start = 1;
x_end = 1.2;
x_delta = (x_end - x_start) / steps;
// Loop for rectangular integration
total = 0; // set the initial sum to zero
for i=0:steps,
x = x_start + i * x_delta;
total = total + f(x);
end
total = total * x_delta;
printf("Rectangular integration value %f\n", total);
// Loop for trapezoidal integration
total = 0; // set the initial sum to zero
for i=0:steps,
x = x_start + i * x_delta;
if i == 0 then
total = total + f(x);
elseif i == steps then
total = total + f(x);
else
total = total + 2 * f(x);
end
end
total = total * x_delta / 2;
printf("Trapezoidal integration value %f\n", total);
// Loop for Simpson's rule integration
total = 0; // set the initial sum to zero
even = 0;
for i=0:steps,
x = x_start + i * x_delta;
if i == 0 then
total = total + f(x);
elseif i == steps then
total = total + f(x);
else
even = even + 1;
if even > 1 then
total = total + 4 * f(x);
even = 0;
else
total = total + 2 * f(x);
end
end
end
total = total * x_delta / 3;
printf("Simpsons rule integration value %f\n", total);
8.
#include <stdio.h>
int main(){
int
steps = 100,
i;
double theta,
omega,
step_t,
theta_last,
omega_last;
theta = 0.0;
omega = 0.0;
step_t = 1.0;
for(i = 0; i < steps; i++){
theta_last = theta;
omega_last = omega;
theta = theta_last + step_t * omega_last;
omega = omega_last + step_t*(-3 * omega_last - 9 * theta_last + 10);
printf("%f %f %f \n", i+step_t, theta, omega);
}
9.
a)
b)
V o ( t ) = 8.331e
Vo = Yo
0.6t
Y o = 1.2Y o 4.2V o + 34
// assign4_1.sce
// by: H. Jack Sept., 23, 2003
v0 = 0;
y0 = 1;
X=[v0, y0];
// initial conditions
// Set the time length and step size for the integration
steps = 1000;
t_start = 0;
t_end = 10;
h = (t_end - t_start) / steps;
t = [t_start];
//
// Loop for integration
//
for i=1:steps,
t = [t ; t($,:) + h];
F1 = h * f(X($,:), t($,:));
F2 = h * f(X($,:) + F1/2.0, t($,:) + h/2.0);
F3 = h * f(X($,:) + F2/2.0, t($,:) + h/2.0);
F4 = h * f(X($,:) + F3, t($,:) + h);
X = [X ; X($,:) + (F1 + 2.0*F2 + 2.0*F3 + F4)/6.0];
end
//
// Graph the values
//
plot2d(t, X, [-2, -5], leg="position@velocity");
xtitle('Time (s)');
//
// Generate points from the given function
//
XX = [v0];
for i=1:steps,
tt = i * h;
XX = [XX ; -8.331 * exp(-0.6 * tt) * cos( 1.960 * tt - 0.238) + 8.095];
end
plot2d(t, XX, [-4], leg="explicit");
c) The two curves produced by the scilab program overlap, so the results agree.
10.
Kd x
F
M
Ks x
x
= F K d x K s x = Mx
Mx + K d x + K s x = F
homogeneous:
x+x+x = 0
2
A +A+1 = 0
1 1 4( 1 )( 1 )
3
A = ---------------------------------------------- = 0.5 j ------2(1)
2
3
0.5t
xh = C1 e
cos ------- t + C 2
2
particular:
xp = B
0+0+B = 1
xp = 1
initial conditions:
x = xh + xp = C1e
0.5t
3
cos ------- t + C 2 + 1
2
3
3
3
0.5t
0.5t
x' = ------- C e
sin ------- t + C 2 0.5C 1 e
cos ------- t + C 2
2
2
2 1
3
x' ( 0 ) = ------- C sin ( C 2 ) 0.5C 1 cos ( C 2 ) = 0
2 1
3
------- sin ( C 2 ) = 0.5 cos ( C 2 )
2
1
tan ( C 2 ) = ------3
1
C 2 = atan ------- = 0.5236
3
x ( 0 ) = C 1 cos ( 0.5236 ) + 1 = 0
1
C 1 = --------------------------------- = 1.155
cos ( 0.5236 )
x = 1.155 e
0.5t
3
sin ------- t + 1.047 + 1
2
11.
a)
b)
c)
g)
Ks
F
K d
x + ------ x + ----- x = ----M
M
M
5t
N
x 1 ( t ) = 0.115e cos ( 5 3t 0.524 ) + 0.10
K d = 100 ---m
0.1t
5 999.9t
N
x 1 ( t ) = 0.1e
+ 10 e
+ 0.10
K d = 10000 ---m
0
X0
d
----= K
dt X
---------s
1
M
1
0
X0
+
Kd
F------------ X 1
M
M
h)
Ns
K d = 200 -----m
12.
b)
x1 = v1
K s1
K s1
F
x 1 = x 1 ----------- + x 2 -------- + -------
M1
M1
M1
x2 = v2
K s1
K s1 K s2
x 1 = x 1 -------- + x 2 ---------------------------
M2
M2
c)
x1
d- v 1
---=
dt x
2
v2
0 1
K s1
----------- 0
M1
0
K s1
-------M1
0
0
x1
v1
0 0
0
1 x2
K s1
K s1 K s2
------- 0 -------------------------- 0 v2
M2
M2
0
F
------+ M1
0
0
e), f), g)
h)
//
// State Equations Calculated Here
//
void derivative(double t, double X[], double dX[]){
dX[0] = X[1];
dX[1] = -Ks1 / M1 * X[0] + Ks1 / M1 * X[2] + Force / M1;
dX[2] = X[3];
dX[3] = Ks1 / M2 * X[0] - (Ks1 + Ks2) / M2 * X[2];
}
x =
5iu ( i 20 )
i=0
where,
u(t) = 0
when
t0
u(t) = 1
when
t>0
2. Write a Scilab program to implement the following equation to calculate the value of x.
10
v =
F( t)
- dt
--------M
M = 10
where,
F ( t ) = 10
when
t5
F(t) = 0
when
t>5
3. Write a Scilab program to implement the following equation to calculate the value of x.
10
x =
f ( t ) dt
f ( t ) = 5 ln ( t )
4. Write a Scilab program to integrate the area under the function below using a numerical
method, such as Simpsons rule. Find the area from 1 to 2.
sin x
f ( x ) = 5x + 2 ln ----------
x
5. Numerically integrate one time step of the differential equation below using a) first order integration and b) Runge Kutta integration.
+ 3 + 9 = 10
6. Convert the third order differential equation below to state equation form. With a numerical
method of your choice, find the state of the system 1 second later. Show all calculations.
x + 4x + 2x + 5x = 10
7. The differential equation below describes a first order system that starts with an initial value of
k=20. Find the state at two milliseconds using a) explicit integration, b) first order numerical
integration and c) Runge-Kutta integration. For the numerical methods use a timestep of
h=0.001s. ----> The final results must be put in a table for easy comparison.
k + 10k = 5
k ( 0 ) = 20
8. For the mechanism shown in the figure below the values are Ks1=Ks2=100N/m, Kd1=10Nm/s,
M1=M2=1kg, F=1N. Assume that the system starts at rest, and the springs are undeformed ini-
tially.
x1
x2
K d1
F
M1
K s1
K s2
M2
v + 20v = 200
rotation - 5.1
5. ROTATION
Topics:
Basic laws of motion
Inertia, springs, dampers, levers, gears and belts
Design cases
Objectives:
To be able to develop and analyze differential equations for rotational systems.
5.1 INTRODUCTION
The equations of motion for a rotating mass are shown in Figure 5.1. Given the
angular position, the angular velocity can be found by differentiating once, the angular
acceleration can be found by differentiating again. The angular acceleration can be integrated to find the angular velocity, the angular velocity can be integrated to find the angular position. The angular acceleration is proportional to an applied torque, but inversely
proportional to the mass moment of inertia.
equations of motion
= -----
dt
d
(2)
( t ) = ( t ) dt =
(3)
( t ) = ( t ) dt
( t )( t) = T
--------where,
d 2
= ----- = -----
dt
dt
d
OR
(1)
( t ) dt dt
(4)
(5)
JM
Figure 5.1
rotation - 5.2
Note: A torque and moment are equivalent in terms of calculations. The main difference is that torque normally refers to a rotating moment.
Given the initial state of a rotating mass, find the state 5 seconds later.
0 = 1rad
rad
s
0 = 2 ---------
rad
= 3 -------2
s
ans.
Figure 5.2
( 5 ) = 86rad
rad
( 5 ) = 17 --------s
5.2 MODELING
Free Body Diagrams (FBDs) are required when analyzing rotational systems, as
they were for translating systems. The force components normally considered in a rotational system include,
inertia - opposes acceleration and deceleration
springs - resist deflection
dampers - oppose velocity
levers - rotate small angles
gears and belts - change rotational speeds and torques
rotation - 5.3
5.2.1 Inertia
When unbalanced torques are applied to a mass it will begin to accelerate, in rotation. The sum of applied torques is equal to the inertia forces shown in Figure 5.3.
, ,
J
T
= JM
J M = I xx + I yy
2
I xx = y dM
2
I yy = x dM
(6)
(7)
(8)
(9)
Note: The mass moment of inertia will be used when dealing with acceleration of a
mass. Later we will use the area moment of inertia for torsional springs.
Figure 5.3
The mass moment of inertia determines the resistance to acceleration. This can be
calculated using integration, or found in tables. When dealing with rotational acceleration
it is important to use the mass moment of inertia, not the area moment of inertia.
The center of rotation for free body rotation will be the centroid. Moment of inertia
values are typically calculated about the centroid. If the object is constrained to rotate
about some point, other than the centroid, the moment of inertia value must be recalculated. The parallel axis theorem provides the method to shift a moment of inertia from a
centroid to an arbitrary center of rotation, as shown in Figure 5.4.
rotation - 5.4
2
J M = JM + Mr
where,
J M = mass moment about the new point
= mass moment about the center of mass
JM
M = mass of the object
r = distance from the centroid to the new point
Figure 5.4
2
J A = JA + Ar
where,
J A = area moment about the new point
JA = area moment about the centroid
A = mass of the object
r = distance from the centroid to the new point
Figure 5.5
Aside: If forces do not pass through the center of an object, it will rotate. If the object
is made of a homogeneous material, the area and volume centroids can be used as
the center. If the object is made of different materials then the center of mass should
be used for the center. If the gravity varies over the length of the (very long) object
then the center of gravity should be used.
rotation - 5.5
4
-5
First find the density and calculate the moments of inertia about
the centroid.
I xx =
4 y
dM =
4 y
-4
1 y
3 4
2 ( 5m ) dy = 1.25Kgm ----3
3
5
-1
10Kg
2
= -------------------------------- = 0.125Kgm
2 ( 5m )2 ( 4m )
4
-2.5
( 4m )
1 ( 4m )
2
= 1.25Kgm --------------- ------------------ = 53.33Kgm
3
3
5
I yy =
5 x
dM =
5 x
1 x
2 ( 4m ) dx = 1Kgm ----3
( 5m )
1 ( 5m )
2
= 1Kgm --------------- ------------------ = 83.33Kgm
3
3
2
The centroid can now be shifted to the center of rotation using the parallel axis theorem.
2
2
2
2
2
J M = JM + Mr = 136.67Kgm + ( 10Kg ) ( ( 2.5m ) + ( 1m ) ) = 209.2Kgm
Figure 5.6
rotation - 5.6
4
-2.5
5
-1
-4
ans.
Figure 5.7
I M x = 66.33Kgm
2
I M y = 145.8Kgm
2
J M = 209.2Kgm
rotation - 5.7
The 20cm diameter 10 kg cylinder to the left is sitting in a depression that is effectively frictionless. If a torque of 10 Nm is applied for 5
seconds, what will the angular velocity be?
ans. ( 5s ) = 312.5rad
rad
( 5s ) = 125 --------s
Figure 5.8
5.2.2 Springs
Twisting a rotational spring will produce an opposing torque. This torque increases
as the deformation increases. A simple example of a solid rod torsional spring is shown in
Figure 5.9. The angle of rotation is determined by the applied torque, T, the shear modulus, G, the area moment of inertia, JA, and the length, L, of the rod. The constant parameters can be lumped into a single spring coefficient similar to that used for translational
springs.
rotation - 5.8
JAG
T = ----------
L
(8)
T = K S ( )
(9)
T
Note: Remember to use radians for these calculations. In fact you are advised to use
radians for all calculations. Dont forget to set your calculator to radians also.
Note: This calculation uses the area moment of inertia.
Figure 5.9
The spring constant for a torsional spring will be relatively constant, unless the
material is deformed outside the linear elastic range, or the geometry of the spring changes
significantly.
When dealing with strength of material properties the area moment of inertia is
required. The calculation for the area moment of inertia is similar to that for the mass
moment of inertia. An example of calculating the area moment of inertia is shown in Figure 5.10, and based on the previous example in Figure 5.6. The calculations are similar to
those for the mass moments of inertia, except for the formulation of the integration elements. Note the difference between the mass moment of inertia and area moment of inertia
for the part. The area moment of inertia can be converted to a mass moment of inertia simply by multiplying by the density. Also note the units.
rotation - 5.9
4
-5
-2.5
5
-1
-4
4m
4m ) 3 ( 4m ) 3
= 10m (-------------- ------------------- = 426.7
3
3
4m
5m
3
( 5m ) 3 ( 5m ) 3
4
5m 2
5m 2
x
x 2 ( 4m ) dx = 8m ----= 8m --------------- ------------------- = 666.7m
I yy =
x dA =
5m
5m
3
3
3
5m
4
4
JA = I xx + I yy = ( 426.7 + 666.7 )m = 1093.4m
I xx =
3
4m 2
4m 2
y----=
y
2
(
5m
)
d
y
=
10m
y
d
A
4m
4m
3
Next, shift the area moment of inertia from the centroid to the other point of rotation.
2
J A = JA + Ar
4
2
2
= 1093.4m + ( ( 4m ( 4m ) ) ( 5m ( 5m ) ) ) ( ( 1m ) + ( 2.5m ) )
= 1673m
I xx =
I yy =
y dA
2
x dA
J A = I xx + I yy
2
J A = JA + Ar
(8)
(9)
(10)
(11)
Note: You may notice that when the area moment of inertia is multiplied by the density of the material, the mass moment of inertia is the result. Therefore if you have
a table of area moments of inertia, multiplying by density will yield the mass
moment of inertia. Keep track of units when doing this.
Figure 5.10
rotation - 5.10
For a 1/2" 1020 steel rod that is 1 yard long, find the torsional spring coefficient.
ans.
Figure 5.11
Nm
K s = 215 --------rad
An example problem with torsional springs is shown in Figure 5.12. There are
three torsional springs between two rotating masses. The right hand spring is anchored
solidly in a wall, and will not move. A torque is applied to the left hand spring. Because
the torsional spring is considered massless the torque will be the same at the other end of
the spring, at mass J1. FBDs are drawn for both of the masses, and forces are summed.
(Note: the similarity in the methods used for torsional, and for translational springs.)
These equations are then rearranged into state variable equations, and finally put in matrix
form.
rotation - 5.11
J M1
K s1
J M2
K s2
K s3
Model the system above assuming that the center shaft is a torsional spring, and that a
torque is applied to the leftmost disk. Leave the results in state variable form.
1
K s2 ( 1 2 )
= K s2 ( 1 2 ) = J M1 1
J M1 1 = K s2 1 + K s2 2 +
1 = 1
K s2
K s2
1 = ----------- 1 + -------- 2 +
JM
JM
J M1
K s2 ( 2 1 )
= K s2 ( 2 1 ) K s3 2 = J M2 2
K s3 K s2
K s2
2 = --------------------------- 2 + -------- 1
JM
JM
K s3 2
1
d- 1
---=
dt
2
Figure 5.12
0 1
K s2
----------- 0
J M1
0
K s2
-------J M1
0
0
1
1
0 0
0
1 2
K s3 K s2
K s2
-------- 0 -------------------------- 0 2
J M2
J M2
0
+
0
0
(2)
K s3 K s2
K s2
2 = --------------------------- 2 + -------- 1
J M2
J M 2
2 = 2
J M2
(1)
(3)
(4)
rotation - 5.12
5.2.3 Damping
Rotational damping is normally caused by viscous fluids, such as oils, used for
lubrication. It opposes angular velocity with the relationships shown in Figure 5.13. The
first equation is used for a system with one rotating and one stationary part. The second
equation is used for damping between two rotating parts.
T = Kd
T = Kd ( 1 2 )
Figure 5.13
If a wheel (JM=5kg m2) is turning at 150 rpm and the damping coefficient is 1Nms/rad,
what is the deceleration?
ans.
rad
= 3.141 -------2
s
Figure 5.14
The example in Figure 5.12 is extended to include damping in Figure 5.15. The
primary addition from the previous example is the addition of the damping forces to the
FBDs. In this case the damping coefficients are indicated with B, but Kd could have
also been used. The state equations were developed in matrix form. Visual comparison of
the final matrices in this and the previous example reveal that the damping terms are the
rotation - 5.13
only addition.
J M1
K s1
J M2
K s2
B1
K s3
B2
Model the system above assuming that the center shaft is a torsional spring, and that a
torque is applied to the leftmost disk. Leave the results in state variable form.
1
K s2 ( 1 2 )
B1 1
+ M = K s2 ( 2 1 ) B 2 2 K s3 2 = J M 2 2
K s2 ( 2 1 )
K s3 K s2
K s2
B2
JM
J M2
2
2
(3)
2 = 2
J M2
K s3 2
1
d- 1
---=
dt
2
2
B2
K s3 K s2
K s2
B2 2
0
K s2
----------JM1
1
B1
--------J M1
0
K s2
-------JM2
Figure 5.15
= K s2 ( 1 2 ) B 1 1 = J M1 1
J M1 1 = B 1 1 K s2 1 + K s2 2 +
(1)
1 = 1
B
K s2
K s2
J M1
0
K s2
-------J M1
0
K s3 K s2
0 --------------------------J M2
0
0
1
1
1 2
B2
2
--------J M2
A System Example
0
------+ J M1
0
0
(4)
rotation - 5.14
5.2.4 Levers
The lever shown in Figure 5.16 can be used to amplify forces or motion. Although
theoretically a lever arm could rotate fully, it typically has a limited range of motion. The
amplification is determined by the ratio of arm lengths to the left and right of the center.
F2
d1
d2
2
F1
1
d1
----- = -----2- = ----1d2
F1
2
Figure 5.16
Given a lever set to lift a 1000 kg rock - if the lever is 2m long and the distance from the
fulcrum to the rock is 10cm, how much force is required to lift it?
ans.
F = 516.3N
Figure 5.17
rotation - 5.15
T 1 = Fc r1
T2 = Fc r2
Vc = 1 r1 = 2 r2
n1
n
----- = ----2r1
r2
T
r
n
--------1- = ----1 = ----1T2
r2
n2
n
----2 = ---------1 = --------1- = ------------1 = ----2r1
2
2
2
n1
where,
n = number of teeth on respective gears
r = radii of respective gears
Fc = force of contact between gear teeth
Vc = tangential velocity of gear teeth
T = torques on gears
Figure 5.18
For lower gear ratios a simple gear box with two gears can be constructed. For
higher gear ratios more gears can be added. To do this, compound gear sets are required.
In a compound gear set two or more gears are connected on a single shaft, as shown in
Figure 5.19. In this example the gear ratio on the left is 4:1, and the ratio for the set on the
right is 4:1. Together they give a gear ratio of 16:1.
rotation - 5.16
N 2 = 60
N2 N4
e = ------------- = 16
N3 N5
N 5 = 15
Output shaft
Input shaft
Compound gear set
N 3 = 15
N 4 = 60
Figure 5.19
A manual transmission is shown in Figure 5.20. In the transmission the gear ratio
is changed by sliding (left-right) some of the gears to change the sequence of gears transmitting the force. Notice that when in reverse an additional compound gear set is added to
reverse the direction of rotation.
rotation - 5.17
9
clutch
stem gear
2
8
7
Motor shaft
reverse idler
10
11
6
5
Speed (gear)
Gear Train
1
2
3
4
reverse
2-3-6-9
2-3-5-8
2-3-4-7
bypass gear train
2-3-6-10-11-9
Figure 5.20
A manual transmission
Rack and pinion gear sets are used for converting rotation to translation. A rack is
a long straight gear that is driven by a small mating gear called a pinion. The basic relationships are shown in Figure 5.21.
rotation - 5.18
T = Fr
V c = r
l = r
where,
r = radius of pinion
F = force of contact between gear teeth
Vc = tangential velocity of gear teeth and velocity of rack
T = torque on pinion
Figure 5.21
Belt based systems can be analyzed with methods similar to gears (with the exception of teeth). A belt wound around a drum will act like a rack and pinion gear pair. A belt
around two or more pulleys will act like gears.
A gear train has an input gear with 20 teeth, a center gear that has 100 teeth, and an
output gear that has 40 teeth. If the input shaft is rotating at 5 rad/sec what is the
rotation speed of the output shaft?
Figure 5.22
rotation - 5.19
5.2.6 Friction
Friction between rotating components is a major source of inefficiency in
machines. It is the result of contact surface materials and geometries. Calculating friction
values in rotating systems is more difficult than translating systems. Normally rotational
friction will be given as static and kinetic friction torques.
An example problem with rotational friction is shown in Figure 5.23. Basically
these problems require that the model be analyzed as if the friction surface is fixed. If the
friction force exceeds the maximum static friction the mechanism is then analyzed using
the dynamic friction torque. There is friction between the shaft and the hole in the wall.
The friction force is left as a variable for the derivation of the state equations. The friction
value must be calculated using the appropriate state equation. The result of this calculation
and the previous static or dynamic condition is then used to determine the new friction
value.
rotation - 5.20
JM
Ks
JM
FF
Ks
T s 10Nm
T k = 6Nm
= Ks TF = JM
JM = Ks T F
=
TF
Ks
= -------------- + ------
JM
JM
(1)
(2)
(3)
Next, the torque force must be calculated, and then used to determine the new torque force.
J M = K s T test
T test = K s J M
cases:
Figure 5.23
(4)
T F = T test
T F = 6Nm
T F = T test
T F = 6Nm
The friction example in Figure 5.23 can be analyzed using the C program in Figure
5.24. For the purposes of the example some component values are selected and the system
is assumed to be at rest initially. The program loops to integrate the state equations. Each
loop the friction conditions are checked and then used for a first-order solution to the state
equations.
rotation - 5.21
int main(){
double
int
FILE
Figure 5.24
rotation - 5.22
2
T load
d
K
K
The speed response of a permanent magnet DC motor is first-order. The steadystate velocity will be a straight line function of the torque applied to the motor, as shown
in Figure 5.26. In addition the line shifts outwards as the voltage applied to the motor
increases.
voltage/current increases
ss
Figure 5.26
rotation - 5.23
(5)
E = E K + EP
EK = JM
Figure 5.27
(6)
E P = T
(7)
P = T
(8)
coefficient units
Ks
Nm-------rad
K d, B
Nms
----------rad
JM
Kgm
rotation - 5.24
To begin the analysis the velocity curve in Figure 5.28 was obtained experimentally by applying a voltage of 15V to the motor with no load attached. In addition the resistance of the motor coils was measured and found to be 40 ohms. The steady-state speed
and time constant were used to determine the constants for the motor.
rpm
2400
R = 40
0.5s
K2
d-
K T
load
--- ----------------
=
+
m J R
s J R- ---------- dt m
JM
M
M
The steady-state velocity can be used to find the value of K.
2
rot K
K
( 0 ) + 2400 --------- ---------- = 15V ---------- ( 0 )
min J M R
JMR
rot- 1min
2400 -------------------- 2rad
---------------- ( K ) = 15V
K 1
1
--------= ---------- = 2s
JM R
0.5s
2
Vs-
39.8 10 3 -------
rad
- = 0.198005 10-4 = 19.8 10 6 Kgm 2
J = -------------------------------------------1
( 40 ) ( 2s )
T load
d-
1
1 2
--- m + m 2s = V s ( 50.3V s rad ) ----------------------------------------6
2
dt
19.8 10 Kgm
m' = m
(1)
1 2
50505Kg m T load
(2)
rotation - 5.25
Figure 5.28
The remaining equations describing the system are developed in Figure 5.29.
These calculations are done with the assumption that the inertial effects of the gears and
other components are insignificant.
rotation - 5.26
The long shaft must now be analyzed. This will require that angles at both ends be
defined, and the shaft be considered as a spring.
gear, gear = angular position and velocity of the shaft at the gear box
pinion, pinion = angular position and velocity of the shaft at the pinion
1
gear = ------ m
20
1
gear = ------ m
20
6 2
1
pinion = ( gear pinion )1.768 10 in Kg K s
1
0.0254in 2
6 2
1
pinion = ------ m pinion 1.768 10 in Kg K s ---------------------
20
1.0m
1
9
2
1
pinion = ------ m pinion ( 1.141 10 )m Kg K s
20
pinion = pinion
(3)
12 2
1
9 2
1
pinion = 57.1 10 m Kg K s m 1.141 10 m Kg K s pinion
(4)
Figure 5.29
If the gear box is assumed to have relatively small moment of inertia, then we can
say that the torque load on the motor is equal to the torque in the shaft. This then allows
rotation - 5.27
the equation for the motor shaft to be put into a useful form, as shown in Figure 5.30. Having this differential equation now allows the numerical analysis to proceed. The analysis
involves iteratively solving the equations and determining the point at which the system
begins to overshoot, indicating critical damping.
1 2
1
1 2
m = V s 50.3V s rad m 2s 50505Kg m K s ( gear pinion )
1
1 2
1
1 2
m = V s 50.3V s rad m 2s 50505Kg m K s ------ m pinion
20
1 2
1 2
m = ( V s 50.3V s rad ) + pinion ( 50505Kg m K s )
1
+ m ( 2 s ) + m ( 2525 Kg m K s )
The state equations can then be put in matrix form for clarity. The units will be eliminated for brevity, but acknowledging that they are consistent.
m
d- m
---=
dt
pinion
pinion
0
2525 K s
1
2
0
50505K s
57.1 10
12
0
0
m
m
1 pinion
9
K s 0 1.141 10 K s 0 pinion
0
Vs 50.3
0
0
The state equations for the system are then analyzed using a computer for the parameters below to find the Ks value that gives a response that approximates critical
damping for a step input from 0 to 10V.
Ks
(rad/Nm)
Overshoot
(rad)
100
Figure 5.30
These results indicate that a spring value of XXX is required to have the system
rotation - 5.28
behave as if it is critically damped. (Note: Clearly this system is not second order, but in
the absence of another characteristics we approximate it as second order.)
5.5 SUMMARY
The basic equations of motion were discussed.
Mass and area moment of inertia are used for inertia and springs.
Rotational dampers and springs.
A design case was presented.
JM2
JM1
Ks1
Ks2
B
rotation - 5.29
2. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form.
R1
Kd1
R2
J1
x2
F
Ks1
M1
x1
3. Draw the FBDs and write the differential equations for the mechanism below.
2
R1
JM1
1
R2
JM2
Ks
Kd
x1
x2
M1
4. The system below consists of two masses hanging by a cable over mass J. There is a spring in
the cable near M2. The cable doesnt slip on J.
a) Derive the differential equations for the following system.
rotation - 5.30
JM
K s1
K s2
x1
M1
F
M2
x2
5. Write the state equations for the system to relate the applied force F to the displacement x.
Note that the rotating mass also experiences a rotational damping force indicated with Kd1
K d1
F
x
JM
K s1
K s2
K s3
K d2
6. For the system pictured below a) write the differential equations (assume small angular deflec-
rotation - 5.31
R1
R2
R3
R4
J1
F1
Kd1
Ks1
M1
x1
7. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form.
R1
R2
J M1, N 1
x2
F1
x3
J M2, N 2
Ks1
Kd1
M3
x1
8. For the system pictured below a) write the differential equations (assume small angular deflec-
rotation - 5.32
Kd1
Ks1
x1
M1
Ks2
x2
J M2, R 2
M2
2
F2
9. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form.
Kd3
Ks3
M3
Ks2
x3
F1
R2,J2,M2
x2
1
R1,J1
10. For the system pictured below a) write the differential equations (assume small angular
rotation - 5.33
Ks1
R2
x2
Kd2
Ks2
F1
x3
11. For the system pictured below a) write the differential equations (assume small angular deflections) and b) put the equations in state variable form.
R2
R1
Two gears have a force on
one side, and a mass
Kd1
N
on the other, both sus2
N1
pended from moment
arms. There is a rotaJ1
J2
tional damping on one
F1
Ks1
1
of the gears.
2
M1
x1
12. Find the polar moments of inertia of area and mass for a round cross section with known
radius and mass per unit area. How are they related?
13. The rotational spring is connected between a mass J, and the wall where it is rigidly held.
The mass has an applied torque T, and also experiences damping B.
a) Derive the differential equation for the rotational system shown.
b) Put the equation in state variable form (using variables) and then plot the position (not velocity) as a function of time for the first 5 seconds with your calcula-
rotation - 5.34
tor using the parameters below. Assume the system starts at rest.
Nm
K s = 10 --------rad
Nms
B = 1 ----------rad
J M = 1Kgm
T
T = 10Nm
Ks
B
c) A differential equation for the rotating mass with a spring and damper is given
below. Solve the differential equation to get a function of time. Assume the system starts at rest.
1
2
2
'' + ( 1s )' + ( 10s ) = 10s
14. Find the response as a function of time (i.e. solve the differential equation to get a function of
time.). Assume the system starts undeflected and at rest.
= 10Nm
Ks
JM
Kd
J M = 1Kgm
Ns
K d = 3 -----m
N
K s = 9 ---m
rotation - 5.35
2
K s1 ( 1 2 )
T
J M1
J M2
K s1 ( 1 2 )
M1
K s2 2
= T K s1 ( 1 2 ) = J M1 1
K s1
K s1
T
M2
= K s1 ( 1 2 ) B 2 K s2 2 = J M 2 2
K s1 + K s2
K s1
B
2 + 2 -------- + 2 ----------------------- + 1 ----------- = 0
JM2
J M2
J M2
2.
x1 = v1
K s1
K s1
x 1 = x 1 ----------- + x 2 -------- + g
M1
M1
x2 = v2
2
R 2 K d1
R 2 K s1
R 2 K s1 FR 1
B 2
rotation - 5.36
3.
T1
T2
M1
J M1
gM 1
if T1,T2,Kdx1 > 0
J M2
T1
T2
x 1
1 = -------R2
x1
2 = -----R1
Fy
= T 1 gM 1 = M 1 x 2
T1
x 2 = g ------M1
M1
= T 1 R 1 + T 2 R 1 = J M1 2
T1 R1 T2 R1
2 = -----------------------------J M1
M2
= T 2 R 2 R 2 K d x 1 = J M2 1
T2 R2
R 2 Kd
1 + x 1 ---------------- = ----------- JM
JM
2
6 equations, 6 unknowns
Kd x1
T1 = Ks ( x2 x1 )
rotation - 5.37
4.
K s1
a)
RT
K s2 ( x 2 R )
JM
M1
RK s2 ( x 2 R )
M2
M1 g
FM1 = T M1 g F = M1 x1
T = M 1 x 1 + M 1 g + F = M 1 R + M 1 g + F
MJ = RT Ks1 + RKs2 ( x2 R ) = JM
M2 g
if(T < 0) T=0
if(T >= 0) R = x 1
2
2
R ( M 1 g + F ) ( K s1 + R K s2 ) + ( RK s2 )x 2 = ( J M + R M 1 )
2
FM2
R ( M1 g + F )
K s1 + R K s2
R K s2
--------------------------- = -------------------------------+
x
+ ----------------------------
2
2
JM + R2 M1
J M + R 2 M 1
JM + R M1
= K s2 ( x 2 R ) M 2 g = M 2 x 2
K s2
R K s2
x 2 + x 2 -------- + --------------- = g
M2
M2
b)
=
2
K s1 R K s2
RK s2 RM 1 g RF
- + x 2 -------------------------- + ---------------------------------
= -------------------------------- JM + R2 M1
J M + R 2 M 1 J M + R 2 M 1
x2 = v2
RK s2
K s2
v 2 = ------------ + x 2 ----------- + g
M2
M2
(1)
(2)
rotation - 5.38
5.
=
2
K d1
K s2 r
K s1 r K s2
Fr
M
M
M
M
x = v
K s2 r
K d2
K s2 K s3
M
6.
x1 = v1
K d1
K s1
K d1 R 2 R 4
K s1 R2 R 4
1 = 1
2 2
2 2
K d1 R 2 R 4
K s1 R 2 R 4
K d1 R 2 R 4
K s1 R2 R 4 F 1 R 1
1 3
rotation - 5.39
7.
2
R 2 K d1 ( x 1 x 3 )
F1 R1
JM1
J M1 1
JM2
1 N 1 = 2 N2
N2
-----N1
K d1 ( x 1 x 3 )
K s1 ( x 1 x 3 )
J M2 2
M3
R 2 K s1 ( x 1 x 3 )
x
2 = -----3R2
M3 x1
M3g
x1 = v1
R 2 K s1
K d1
K s1
R2 K d1
2 = 2
N2
2
2
v 1 ( R2 K d1 ) + x 1 ( R 2 K s1 ) + 2 ( R 2 K d1 ) + 1 ( R2 K s1 ) + F --------- R 1
N1
2 = ----------------------------------------------------------------------------------------------------------------------------------------------------------------2
N2
J M1 -----2- + J M2
N1
rotation - 5.40
8.
K d1 x 1
K s2 ( x 1 x 2 )
K s1 x 1
x1
x1
k N -------x1
M1
M2 x2
K s2 ( x 1 x 2 )
N = M 1 g cos ( 3 )
M 1 g sin ( 3 )
J M2, R 2
M2
Fc
x
2 = -----2R2
x2
2
F2
M 2 g sin ( 3 )
x1 = v1
v1 =
x2 = v2
v2 =
9.
x2 = v2
2 2
2 2
2 2
R 1 R 2 ( 2F 1 + M 2 g )
R 1 R 2 K s2
R 1 R 2 ( K s2 K s3 )
rotation - 5.41
10.
state equations
x1 = v1
R1 + R 2
v 1 = F 1 ------------------
R1 M1
K s2
K s2
F1
output equations
R1 + R 2
x 2 = ( x 1 p ) ------------------
R1
x3 = q + x2
11.
x1 = v1
K s1
R 2 K s1 N 1
v 1 = x 1 ----------- + 1 --------------------- + g
M1
M1 N2
1 = 1
2
F1 R1 N2
R 2 K s1 N 1 N 2
K d1 N 1
R 2 K s1 N 1
For area:
J area
r
= r dA = r ( 2r dr ) = 2 r dr = 2 ---4
2
R
= --------2
MFor mass: = M
----- = -------2
A
R
R
J mass
r
= r dM = r ( 2r dr ) = 2 r dr = 2 ---4
2
R
MR
= --------- = ---------- 2
2
The mass moment can be found by multiplying the area moment by the area density.
rotation - 5.42
13.
T
a)
M = T K s B = J M
+
J M + B + K s = T
Ks
T
B
+ ------ + ------ = -----JM
JM
JM
JM
Ks
B
T Ks
' =
b)
2
1
0
5s
Nm
10
s
- 10 10
--------- ------- = --------------------
2
rad
rad
Kgm
10
s
2
= s 10 --------- ---------
rad rad
rotation - 5.43
(c
homogeneous:
1
2
+ ( 1s ) + ( 10s ) = 0
At
guess: = e At
h = Ae
h
2 At
At
2 At
h = A e
2
A e + ( 1s )Ae + ( 10s )e
2
A + ( 1s )A + 10s
1
At
= 0
1 2
= 0
1s ( 1s ) 4 ( 1 ) ( 10s )
A = ------------------------------------------------------------------------------2( 1)
1s 1s 40s
1
A = ------------------------------------------------------- = ( 0.5 j3.123 )s
2(1)
h = C1 e
0.5s t
cos ( 3.123s t + C 2 )
particular:
1
p = 0
p = A
guess:
1
p = 0
( 0 ) + ( 1s ) ( 0 ) + ( 10s ) ( A ) = 10s
p = 1
10s = 1
A = -----------2
10s
Initial conditions:
( t ) = C1e
0.5s t
( 0 ) = C1 e
cos ( 3.123s t + C 2 ) + 1
0.5s 0
cos ( 3.123s 0 + C 2 ) + 1 = 0
C 1 cos ( C 2 ) + 1 = 0
1
' ( t ) = 0.5s C 1 e
0.5s t
0.5s t
C 2 = 0.159
C 1 cos ( 0.159 ) + 1 = 0
1
C 1 = ------------------------------ = 1.013
cos ( 0.159 )
( t ) = 1.013 e
0.5s t
14.
10
27
1.5t
( t ) = ------ + ( 1.283 )e
cos ---------- t 0.524
9
2
sin ( 3.123s t + C 2 )
rotation - 5.44
2
N1
x2
Kd1
J1
M1
N2
x1
J2
T1
Ks1
2. Draw FBDs for the following mechanical system containing two gears.
Ks1
J 1, R 1
J 2, R 2
y
M
Ks2
Kd2
rotation - 5.45
3. Draw FBDs for the following mechanical system. Consider both friction cases.
Ks1
J,R
x
Ks1
M
s, k
Ks2
Kd2
5. Develop a differential equation of motion for the system below assuming that the cable always
rotation - 5.46
remains tight.
J,R
Ks1
J,R
y
Ks2
Kd2
6. Analyze the system pictured below assuming the rope remains tight.
J,R
F = 10N
M 1 = 1kg
M1
Ks1
Kd1
M 2 = 1kg
y
k=0.2
R = 0.1m
M2
=45
Ks2
J = 10Kgm
N
K s1 = 100 ---m
N
K s2 = 100 ---m
Ns
K d1 = 50 -----m
Ns
K d2 = 50 -----m
Kd2
a) Draw FBDs and write the differential equations for the individual masses.
b) Combine the equations in input-output form with y as the output and F as the
input.
d) Write the equations in state variable matrix form.
c) Use Runge-Kutta to find the system state after 1 second.
rotation - 5.47
7. Analyze the system pictured below assuming the rope remains tight.
J,R
J,R
F = 10N
M 1 = 1kg
M 2 = 1kg
R = 0.1m
J = 10Kgm
N
K s1 = 100 ---m
N
K s2 = 100 ---m
Ns
K d1 = 50 -----m
Ns
K d2 = 50 -----m
J,R
F
M1
Ks1 Kd1
M2
Ks2 Kd2
a) Draw FBDs and write the differential equations for the individual masses.
b) Combine the equations in input-output form with y as the output and F as the
input.
c) Write the equations in state variable matrix form.
d) Use Runge-Kutta to find the system state after 1 second.
6. INPUT-OUTPUT EQUATIONS
Topics:
The differential operator, input-output equations
Design case - vibration isolation
Objectives:
To be able to develop input-output equations for mechanical systems.
6.1 INTRODUCTION
To solve a set of differential equations we have two choices, solve them numerically or symbolically. For a symbolic solution the system of differential equations must be
manipulated into a single differential equation. In this chapter we will look at methods for
manipulating differential equations into useful forms.
d- x = Dx
---dt
basic definition
n
n
d
-------- = D
n
dt
1- x =
--D
algebraic manipulation Dx + Dy = D ( x + y )
Dx + Dy = Dy + Dx
Dx + ( Dy + Dz ) = ( Dx + Dy ) + Dz
n
nm
D
---------x- = D
x
m
D
x--------------------( D + a )- = x
(D + a)
simplification
at
at
x(t) -----------------= e x ( t )e dt + C
(D + a)
first-order axiom
Figure 6.1
Note:
x + Ax = y ( t )
xD + Ax = y ( t )
At
At
At
e xD + e Ax = e y ( t )
At
At
At
At
xD + Ax = y ( t )
x( D + A) = y( t)
y( t)
x = -------------D+A
e xD = e y ( t )
e xD = e y ( t )
At
e x =
x = e
A t
At
y ( t )dt + C
at
e y ( t )dt + C
At
At
y(t )A t
------------= e e y ( t )dt + C
D+A
Figure 6.2
d- at
at d
at
---( e x ) = e ----- x + ae x
dt
dt
at
at
De x = e Dx + ae x
x dt
d
d- 2
---x + ----- x + 5x = 5t
dt
dt
2
D x + Dx + 5x = 5t
2
x ( D + D + 5 ) = 5t
5t
x = -------------------------2
D +D+5
5
x = t ------------------------- 2
D +D+5
Figure 6.3
d---x + 5x = 3t
dt
Given,
x ( 0 ) = 10
Dx + 5x = 3t
3t
x = ------------D+5
x = e
5 t
x = e
5 t
d- 5t 5t
---( te e )
dt
guess,
5t
5t
= 5te + e e
5t
3tdt + C
5t
5t
te e
3 --------------------- + C
Initial conditions,
= 5te
5t
5t
d- 5t 5t
5t
---( te e ) = 5e t
dt
5t
5t
e te
-------------------=
5
5t
t dt
(0)( 1) ( 1)
x ( 0 ) = ( 1 ) 3 ----------------------------- + C = 10
5
C = 10.6
x = e
5 t
5t
5t
te e -
-------------------
3
+ 10.6
5
x = 0.6 ( t 1 ) + 10.6e
x = 0.6t 0.6 + 10.6e
Figure 6.4
5t
5t
e.g.,
2y 1 + y 1 + y 1 + 4y 1 = u 1 + u 1 + 3u 2 + u 3 + u 3
y 2 + 6y 2 + y 2 = u 1 + 3u 2 + u 2 + 0.5u 2 + u 3
where,
y = outputs
u = inputs
Figure 6.5
y 1 = 3y 1 + 2y 2 + u1 + 2u 2
y 2 = 2y 1 + y 2 + u1
(1)
(2)
Dy 2 = 2y 1 + y 2 + Du 1
y 2 ( D 1 ) = 2y 1 + Du 1
D+3
y 1 ------------- 0.5u 1 Du2 ( D 1 ) = 2y 1 + Du 1
2
2
D + 2D 3
2
y 1 ----------------------------- 2 0.5Du 1 + 0.5u 1 D u 2 + Du2 = Du 1
2
2
Figure 6.6
Find the second equation for the example in Figure 6.6 for the output y2.
Figure 6.7
x1
F
x2
Ks
M1
M2
Ks ( x2 x 1 )
M1
Fx
= F + Ks ( x2 x1 ) = M1 D x1
2
x1 ( M1 D + K s ) = F + Ks x2
Ks ( x2 x 1 )
Fx
M2
= Ks ( x2 x1 ) = M2 D x2
2
Ks x1 = x 2 ( M2 D + Ks )
The equations can be combined to eliminate x2.
Ks x 1
2
-
x 1 ( M 1 D + K s ) = F + K s ------------------------- M 2 D 2 + K s
2
x1 ( ( M1 D + K s ) ( M2 D + Ks ) Ks ) = F ( M2 D + Ks )
4
x1 ( D M1 M2 + D Ks ( M1 + M2 ) + Ks Ks ) = F ( M 2 D + Ks )
2
x1 ( D M1 M2 + D Ks ( M1 + M2 ) ) = F ( M2 D + Ks )
d 2
d- 4
d 2
---x 1 M 1 M 2 + ----- x 1 K s ( M 1 + M 2 ) = ----- FM + FK s
dt
dt
dt
2
Figure 6.8
The equation is then converted to state variable form, including a step to calculate
a second derivative of the input, as shown in Figure 6.9.
This can then be written in state variable form by creating dummy variables for integrating the function F.
d-
--- dt x 1 = v 1
d-
---v = a1
dt 1
d-
--- dt a 1 = d 1
d-
--- dt d 1 M 1 M 2 + a 1 K s ( M 1 + M 2 ) = a F M 2 + FK s
K s ( M 1 + M 2 )
Ks
1 -
d
------------- ---- + aF ------ d 1 = a 1 ---------------------------------+
F
M 1
M 1 M 2
dt
M1 M2
The approximate value of the second derivative of a unit time step can be calculated using
the time step.
1
a F0 = -----2
T
1
a F 1 = ----2T
1
1
v F = TaF 0 + Ta F1 = T -----2 + T ----2- = 0
T
T
2
(to verify)
T
T
T 1
T 1
1 1
x F = ----- a F + ----- aF = ----- -----2 + ----- -----2 = --- + --- = 1
2 2
2
2
2 T 2 T
These equations can then be written in matrix form.
x1
d- v 1
---dt a
1
d1
Figure 6.9
0 1
0 0
= 0 0
0
1
0
Ks ( M1 + M2 )
0 0 ----------------------------------M1 M2
0 x
0
0
1
0
0
0
v1
aF
+
1
0
0
a1
F
Ks
1 - ------------------0 d
1
M1 M1 M2
x + 3x + 2x = 4F + 5F
The solution begins by evaluating the homogeneous equation as normal.
x + 3x + 2x = 0
2
A + 3A + 2 = 0
2
3 3 4 ( 2 )
A = ------------------------------------- = 1, 2
2
t
2t
xh = C1 e + C 2 e
The particular solution can also be found as normal, assuming F is unit step function.
Guess,
x = A
x = 0
x = 0
0 + 3 ( 0 ) + 2A = 2 ( 0 ) + 5 ( 1 )
5
A = --2
The derivative in the non-homogeneous solution must now be used to find the initial conditions. However the initial position and velocity are known to be zero.
2
d
d
d-
---x + 3 ----- x 0 + 2x 0 = 4 ----- F + 5F 0
dt 0
dt
dt 1
2
d-
d-
--- --- dt x 0 + 3 ( 0 ) + 2 ( 0 ) = 4 dt 1 + 5 ( 0 )
u ( t ) = ----dt
d-
d-
--- --- dt x 0 = 4 dt 1
d---x = 4
dt 0
Figure 6.10
The initial conditions can then be used to find the values of the coefficients. It will be
assumed that the system starts undeflected and at rest.
t
x ( t ) = C 1 e + C2 e
x ( 0 ) = C1e
2t
+ C2e
5
+ --2
5
+ --- = 0
2
5
C 1 = C 2 --2
t
2t
x ( t ) = C 1 e 2C 2 e
x ( 0 ) = C 1 2C 2 = 4
5
C 2 --- 2C 2 = 4
2
3
C 2 = --2
3
5
C 1 = --- --- = 1
2 2
Figure 6.11
The example in Figure 6.10 is reconsidered with a sinusoidal input in Figure 6.12.
In this case the initial acceleration is found to be non-zero. In practical terms, this can be
ignored because only the initial position and velocity will be used to find the coefficients.
F ( t ) = cos ( t )
F(0) = 1
d
d- 2
---x 0 + 3 ----- x 0 + 2x 0 = 4 ( F ( 0 ) ) + 5F ( 0 )
dt
dt
2
d-
--- dt x 0 + 3 ( 0 ) + 2 ( 0 ) = 4 ( 1 ) + 5 ( 0 )
2
d-
--- dt x 0 = 4
Figure 6.12
dxd x
+ K s x = M -------2 Fy = F + Kd ----dt
dt
Ks
Kd
dx
d x
F = M -------2- + K d ------ + K s x
dt
dt
x
F = MD x + K d Dx + K s x
F
2
--- = MD x + K d Dx + K s x
x
F
OR
1
--x- = ----------------------------------------2
F
MD + K d D + K s
Aside: An important concept that is ubiquitous yet largely unrecognized is the use of
functional design. We look at parts of systems as self contained modules that use
inputs to produce outputs. Some systems (such a mechanisms) are reversible, others
are not (consider a internal combustion engine, turning the crank does not produce
gasoline). An input is typically something we can change, an output is the resulting
change in a system. For the example above F over x implies that we are changing the input x, and there is some change in F. We know this could easily be
reversed mathematically and practically.
Figure 6.13
Aside: Keep in mind that the mathematical expression F/x is a ratio between input
(displacement action) and output (reaction force). When shown with differentials it
is obvious that the ratio is not simple, and is a function of time. Also keep in mind
that if we were given a force applied to the system it would become the input (action
force) and the output would be the displacement (resulting motion). To do this all
we need to do is flip the numerators and denominators in the transfer function.
Mass-spring-damper systems are often used when doing vibration analysis and
design work. The first stage of such analysis involves finding the actual displacement for a
given displacement or force. A system experiencing a sinusoidal oscillating force is given
in Figure 6.14. Numerical values are substituted and the homogeneous solution to the
equation is found.
N
K s = 2 ---m
Ns
K d = 0.5 -----m
F = 5 sin ( 6t ) N
The differential equation for the mass-spring damper system can be written.
2
dx
N
d x- Ns
------1Kg 2 + 0.5 ------ ------ + 2 ---- x = 5 sin ( 6t ) N
m dt
m
dt
The homogeneous solution can be determined.
2
d x
Ns dx
N
1Kg -------2- + 0.5 ------ ------ + 2 ---- x = 0
m dt
m
dt
Ns
Ns 2
N
0.5 ------ 0.5 ------ 4 ( 1Kg ) 2 ----
m
m
m
A = ------------------------------------------------------------------------------------------2 ( 1Kg )
A = 0.5 ( 0.5 0.25 8 )s
A = ( 0.25 1.392j )s
xh = C1 e
Figure 6.14
0.25t
cos ( 1.392t + C2 )
The solution continues in Figure 6.15 where the particular solution is found and
put in phase shift form.
( 0.1460 ) + ( 0.01288 )
x p = ------------------------------------------------------------------- ( ( 0.1460 ) sin 6t + ( 0.01288 ) cos 6t )
2
2
( 0.1460 ) + ( 0.01288 )
x p = 0.1466 ( 0.9961 sin 6t 0.08788 cos 6t )
0.9961
x p = 0.1466 sin 6t + atan ----------------------
0.08788
x p = 0.1466 sin ( 6t + 1.483 )
Figure 6.15
The system is assumed to be at rest initially, and this is used to find the constants in
the homogeneous solution in Figure 6.16. Finally the displacement of the mass is used to
find the force exerted through the spring on the ground. In this case there are two force frequency components at 1.392rad/s and 6rad/s. The steady-state force at 6rad/s will have a
magnitude of .2932N. The transient effects have a time constant of 4 seconds (1/0.25), and
should be negligible within a few seconds of starting the machine.
0.25t
0.25t
0.25t
sin ( 1.392t + C 2 ) )
+ 6 ( 0.1466 cos ( 6t + 1.483 ) )
0.25t
The displacement can then be used to calculate the force transmitted to the ground,
assuming the spring is massless.
F = Ks x
N
0.25t
F = 2 ---- ( 0.1673 e
cos ( 1.392t 0.5099 ) + 0.1466 sin ( 6t + 1.483 ) )m
m
F = ( 0.3346 e
Figure 6.16
0.25t
A decision has been made to reduce the vibration magnitude transmitted to the
ground to 0.1N. This can be done by adding a mass-spring isolator, as shown in Figure
6.17. In the figure the bottom mass-spring-damper combination is the original system. The
mass and spring above have been added to reduce the vibration that will reach the ground.
Values must be selected for the mass and spring. The design begins by developing the differential equations for both masses.
K s2
x2
M2
Kd
K s1
M1
x1
F
K s2 x 2
= K s2 x 2 K d ( x 2 x 1 ) K s1 ( x 2 x 1 ) = M 2 x 2
K s2 x 2 K d ( x 2 D x1 D ) K s1 ( x 2 x 1 ) = M 2 x 2 D
M2
Kd ( x2 x1 )
K s1 ( x 2 x 1 )
x 2 ( K s2 K d D K s1 M 2 D ) = x 1 ( K d D K s1 )
2
K s2 + K d D + K s1 + M 2 D
x 1 = x 2 -------------------------------------------------------------
K d D + K s1
Kd ( x2 x1 )
K s1 ( x 2 x 1 )
= K d ( x 2 x 1 ) + K s1 ( x 2 x 1 ) F = M 1 x 1
K d ( x 2 D x 1 D ) + K s1 ( x 2 x 1 ) F = M 1 x 1 D
M1
F
Figure 6.17
(1)
x 1 ( K d D K s1 M 1 D ) + x 2 ( K d D + K s1 ) = F
(2)
For the design we are only interested in the upper spring, as it determines the force
on the ground. An input-output equation for that spring is developed in Figure 6.18. The
given values for the mass-spring-damper system are used. In addition a value for the upper
mass is selected. This is arbitrarily chosen to be the same as the lower mass. This choice
may need to be changed later if the resulting spring constant is not practical.
The solution begins by combining equations (1) and (2) and inserting the.numerical values for the lower mass, spring and damper. We can also limit the problem by selecting a
mass value for the upper mass.
2
K s2 + K d D + K s1 + M 2 D
2
x 2 ------------------------------------------------------------ ( K d D K s1 M 1 D ) + x 2 ( K d D + K s1 ) = F
K
D
+
K
d
s1
N
Ns
K s1 = 2 ---K d = 0.5 -----M 1 = 1Kg
M 2 = 1Kg
m
m
2
K s2 + 0.5D + 2 + D
2
x 2 ------------------------------------------------- ( 0.5D 2 D ) + x 2 ( 0.5D + 2 ) = F
0.5D + 2
x 2 ( D ( 1 ) + D ( K s2 ) + D ( 0.5K s2 ) + ( 2K s2 ) ) = F ( 0.5D + 2 )
This can now be converted back to a differential equation and combined with the force.
d 4
d 2
d
d
----- x + K s2 ----- x 2 0.5K s2 ----- x 2 2K s2 x 2 = 0.5 ----- F + 2F
dt 2
dt
dt
dt
d 4
d 2
d
d
----- x + K s2 ----- x 2 0.5K s2 ----- x 2 2K s2 x 2 = 0.5 ----- 5 sin ( 6t ) + 2 ( 5 ) sin ( 6t )
dt 2
dt
dt
dt
d 2
d
d 4
----- x + K s2 ----- x 2 0.5K s2 ----- x 2 2K s2 x 2 = 15 cos ( 6t ) + 10 sin ( 6t )
dt
dt
dt 2
Figure 6.18
This particular solution of the differential equation will yield the steady-state displacement of the upper mass. This can then be used to find the needed spring coefficient.
d-
--- dt x p = 1296A sin 6t + 1296B cos 6t
sin ( 6t ) ( 1296A 36AK s2 + 0.5K s2 6B 2K s2 A ) = 10 sin ( 6t )
A ( 1296 38K s2 ) + B ( 3K s2 ) = 10
10 + A ( 1296 + 38K s2 )
B = -----------------------------------------------------3K s2
cos ( 6t ) ( 1296B 36BK s2 + ( 0.5 )K s2 6A 2K s2 B ) = 15 cos ( 6t )
A ( 3K s2 ) + B ( 1296 38K s2 ) = 15
10 + A ( 1296 + 38K s2 )
A ( 3K s2 ) + ------------------------------------------------------ ( 1296 38K s2 ) = 15
3K s2
2
45K s2
9K s2
Figure 6.19
- ( 1296 + 38K s2 )
10 + -------------------------------------------------------------------------------- 1453K 22s 98496K 2s 1679616
B = -----------------------------------------------------------------------------------------------------------------------------------------3K s2
2
1620K 2s + 2.097563 10 K 2s
B = -------------------------------------------------------------------------------------------------2
3K s2 ( 1453K 2s 98496K 2s 1679616 )
540K 2s + 69918768
B = --------------------------------------------------------------------------------2
1453K 2s 98496K 2s 1679616
Figure 6.20
Finally the magnitude of the particular solution is calculated and set to the desired
amplitude of 0.1N. This is then used to calculate the spring coefficient.
amplitude =
0.1 =
A +B
425K 2s + 12960
540K 2s + 69918768
2
2
- + ----------------------------------------------------------------------------------
-------------------------------------------------------------------------------- 1453K 22s 98496K 2s 1679616
1453K 22s 98496K 2s 1679616
A value for the spring coefficient was then found using Mathcad to get a value of 662N/m.
Figure 6.21
6.5 SUMMARY
The differential operator can be manipulated algebraically
Equations can be manipulated into input-output forms and solved as normal differential equations
y+y+x = 3
x+x+y = 0
3. Find the input-output form for the following equations.
x 1 + x 1 + 2x 1 x 2 x 2 = 0
x1 x1 + x2 + x2 + x2 = F
4. The following differential equations were converted to the matrix form shown. Use Cramers
rule to find an input-output equation for y.
F
y + 2x = -----10
7y + 4y + 9x + 3x = 0
F
y = -----10
2
( 7D + 4 ) ( 9D + 3 ) x
0
2
(D )
(2)
5. Find the input output equation for y2. Ignore the effects of gravity.
K s1
y1
M1
K s2
M2
y2
F
6. Find the input-output equations for the systems below. Here the input is the torque on the left
hand side.
1
J1
K s1
B1
K s2
J2
B2
7. Write the input-output equations for the mechanical system below. The input is force F, and
the output is y or the angle theta (give both equations). Include the inertia of both masses, and
JM
M
F
8. The applied force F is the input to the system, and the output is the displacement x.
K1 = 500 N/m
K2 = 1000 N/m
M = 10kg
F
B
x + x ----- = ---- M
M
2.
x + 2x = 3
y + 2y = 3
3.
d 3
d 2
d
d
d- 4
---x + 2 ----- x + 3 ----- x + ----- x + x 1 = ----- F + F
dt 1
dt 1
dt 1 dt 1
dt
d- 4
d- 3
d- 2
d-
d- 2
d-
--- --- --- --- --- ---x
+
2
x
+
3
x
+
x
+
x
=
2
dt 2
dt 2
dt 2 dt 2
dt F + dt F + 2F
4.
F----10
(2)
F2
2
----( 9D + 3 )
2
0 ( 9D + 3 )
10
F ( 0.9D + 0.3 ) --------------------------------------------------y = -------------------------------------------------------- = -------------------------------------------------------------=
2
2
4
2
9D + 3D 14D 8
D ( 9D + 3 ) 2 ( 7D + 4 )
2
(D )
(2)
2
( 7D + 4 ) ( 9D + 3 )
4
14
)
+
y
(
8
)
=
dt
dt
dt
dt F ( 0.9 ) + F ( 0.3 )
d 2 1
d 2 1
d 1 14
8
1
d- 4
---y + ----- y --- + ----- y --------- + y ------ = ----- F ------ + F ---
dt
dt
3
dt
9
9
dt
10
3
5.
K s1 K s2
d- 4
d 2 K s1 M 2 + K s2 M 2 + K s2 M 1
---y 2 + ----- y 2 ------------------------------------------------------------- + y 2 ----------------- =
dt
dt
M1 M2
M1 M2
M1
K s1 + K s2
d- 2 ------------- ---- + F ---------------------F
dt M 1 M 2
M1 M2
6.
d- 4
d- 3 J---------------------------d- 2 J---------------------------------------------------1 B 2 + J 2 B 1
1 K s2 + J 2 K s2 + B 1 B 2
--- --- ---- +
+
+
2
2
2
dt
dt
dt
J1 J2
J1 J2
K s2
d B 1 K s2 + B 2 K s2
---- 2 ------------------------------------ = -------- dt
J1 J2
J1 J2
J1 B2 + J2 B1
J 1 K s2 + J 2 K s2 + B1 B 2
d- 4
d 3
d 2
--- 1 + ----- 1 ----------------------------- + ----- 1 ----------------------------------------------------- +
dt
dt
dt
J1 J2
J1 J2
B2
K s2
1
d- B
d- 2 ---d
1 K s2 + B 2 K s2
-------------------------------------- = --- - + ----- ---------- + ----------
1
dt
dt J 1 dt J 1 J 2
J 1 J 2
J1 J2
7.
R ( K ( R ) )
= R ( K ( R ) ) + R ( K ( y R ) ) = J M
2
2
R K + RKy + R K = J M
JM
R ( K ( y R ) )
K ( y R )
2R K + J M D
---------------------------------- = y
RK
= K ( y R ) F Mg = My
K ( y + R ) + F + Mg = M yD
( KR ) + F + Mg = y ( MD K )
2
( KR ) + y ( MD + K ) = F Mg
F+Mg
for the theta output equation;
2
2R K + J M D
2
( KR ) + ---------------------------------- ( MD + K ) = F Mg
RK
2 2
2
2
2
( K R ) + ( 2R K + J M D ) ( MD + K ) = FKR + MgKR
2 2
RK
- ( KR ) + y ( MD 2 + K ) = F Mg
y -------------------------------- 2R2 K + J M D 2
4
2
2 JM D M
2 JM D
2
y 2R MD + ------------------- + K2R + ------------- R K =
K
K
2
2
2 JM D
2 JMD
F 2R + ------------- Mg 2R + -------------
K
K
4
2
M
J
J
d
d- ----------2
M
---- + y ( R2 K ) =
y M - + ----- y 2R M + ---- dt K dt
K
J M
d- 2 --------2
2
---F
+ F ( 2 R ) + ( 2 MgR )
dt K
2
R 2 K 2
d- 4
d- 2 2KR
1-
---
------------------+
+ ------------ =
dt y dt y J M + M y J M M
2 KR 2 2 gKR2
d- 2 1
-------F
+ F ----------------- + --------------------
dt M
JM M JM
8.
K1 K 2
K2
K2
b)
c)
y + y + 5x = 3
x+x+y = 7
2. For the system pictured below find the output response as a function of time for y1 and y2 using
a) integration to find an explicit function, b) numerical analysis using Scilab or C.
N
K s1 = 100 ---m
K s1
y1
M1
K s2
M2
N
K s2 = 1000 ---m
M 2 = 0.1Kg
y2
F
M 1 = 1Kg
6.9 REFERENCES
Irwin, J.D., and Graf, E.R., Industrial Noise and Vibration Control, Prentice Hall Publishers,
1979.
Close, C.M. and Frederick, D.K., Modeling and Analysis of Dynamic Systems, second edition,
John Wiley and Sons, Inc., 1995.
circuits - 7.1
7. ELECTRICAL SYSTEMS
Topics:
Basic components; resistors, power sources, capacitors, inductors and op-amps
Device impedance
Example circuits
Objectives:
To apply analysis techniques to circuits
7.1 INTRODUCTION
A voltage is a pull or push acting on electrons. The voltage will produce a current
when the electrons can flow through a conductor. The more freely the electrons can flow,
the lower the resistance of a material. Most electrical components are used to control this
flow.
7.2 MODELING
Kirchoffs voltage and current laws are shown in Figure 7.1. The node current law
holds true because the current flow in and out of a node must total zero. If the sum of currents was not zero then electrons would be appearing and disappearing at that node, thus
violating the law of conservation of matter. The loop voltage law states that the sum of all
rises and drops around a loop must total zero.
Inode = 0
Vloop = 0
Figure 7.1
node current
loop voltage
Kirchoffs laws
The simplest form of circuit analysis is for DC circuits, typically only requiring
algebraic manipulation. In AC circuit analysis we consider the steady-state response to a
circuits - 7.2
sinusoidal input. Finally the most complex is transient analysis, often requiring integration, or similar techniques.
DC (Direct Current) - find the response for a constant input.
AC (Alternating Current) - find the steady-state response to an AC input.
Transient - find the initial response to changes.
There is a wide range of components used in circuits. The simplest components are
passive, such as resistors, capacitors and inductors. Active components are capable of
changing their behaviors, such as op-amps and transistors. A list of components that will
be discussed in this chapter are listed below.
resistors - reduce current flow as described with ohms law
voltage/current sources - deliver power to a circuit
capacitors - pass current based on current flow, these block DC currents
inductors - resist changes in current flow, these block high frequencies
op-amps - very high gain amplifiers useful in many forms
7.2.1 Resistors
Resistance is a natural phenomenon found in all materials except superconductors.
A resistor will oppose current flow as described by ohms law in Figure 7.2. The resistance value is assumed to be linear, but in actuality it varies with conductor temperature.
I
V
I = --R
+
V
R
-
Figure 7.2
V = IR
Ohms law
The voltage divider example in Figure 7.3 illustrates the methods for analysis of
circuits using resistors. In this circuit an input voltage is supplied on the left hand side.
The output voltage on the right hand side will be some fraction of the input voltage. If the
output resistance is very large, no current will flow, and the ratio of output to input voltages is determined by the ratio of the resistance between R1 and R2. To prove this the cur-
circuits - 7.3
rents into the center node are summed and set equal to zero. The equations are then
manipulated to produce the final relationship.
I1 + I2 + I3 = 0
+
R1
Vi Vo
0 Vo
----------------- + I 2 + --------------- = 0
R1
R2
I1
Vi
I2
R2
I3
+
Vo
-
Vi V o
Vo
----------------- + --------- = 0
R1
R2
1
1
1
V o ------ + ------ = V i ------
R1 R2
R1
R 1 + R 2
1-
- = V i ----V o ---------------- R 1 R2
R 1
Vo
R2
------ = -----------------Vi
R 1 + R2
Figure 7.3
If two resistors are in parallel or series they can be replaced with a single equivalent resistance, as shown in Figure 7.4.
circuits - 7.4
R1
series resistors
R eq = R 1 + R 2
R2
parallel resistors
R1
R2
11
1
------= ------ + -----R eq
R1 R2
1
R eq = ------------------1
1------ + ----R1 R2
R1 R 2
R eq = -----------------R1 + R2
Figure 7.4
circuits - 7.5
+
V
-
Figure 7.5
+
V
V+
A circuit containing a voltage source and resistors is shown in Figure 7.6. The circuit is analyzed using the node voltage method.
R1
+
Vi
+
-
R2
R3
Vo
-
Examining the circuit there are two loops, but only one node, so the node current methods is the most suitable for calculations. The currents into the upper right node, Vo,
will be solved.
Vo Vi Vo Vo
- + ------ + ------ = 0
=
I
---------------R1
R2 R3
1
1
1
1
V o ------ + ------ + ------ = V i -------
R 1 R 2 R 3
R1
R2 R 3 + R1 R 3 + R 1 R2
1-
- = V i -----V o ------------------------------------------------
R1
R 1 R2 R 3
R2 R 3
V o = Vi --------------------------------------------------
R 2 R 3 + R1 R 3 + R 1 R2
Figure 7.6
A circuit calculation
circuits - 7.6
Evaluate the circuit in Figure 7.6 using the loop voltage method.
R2 R3
V o = V i --------------------------------------------------
R1 R2 + R1 R3 + R2 R3
Figure 7.7
circuits - 7.7
Evaluate the voltage divider in Figure 7.3 using the loop current method. Hint: Put a voltage supply on the left, and an output resistor on the right. Remember that the output
resistance should be infinite.
Figure 7.8
Dependant (variable) current and voltage sources are shown in Figure 7.9. The
voltage and current values of these supplies are determined by their relationship to some
other circuit voltage or current. The dependant voltage source will be accompanied by a
+ and - symbol, while the current source has an arrow inside.
V = f( )
+
I = f( )
Figure 7.9
circuits - 7.8
2ohm
Ii
+
V1
-
3ohm
Io
- V =f(V )
2
1
f(V1) = 3V1
Figure 7.10
7.2.3 Capacitors
Capacitors are composed of two isolated metal plates very close together. When a
voltage is applied across the capacitor, electrons will be forced into one plate, and forced
out of the other plate. Temporarily this creates a small current flow until the plates reach
equilibrium. So, any voltage change will result in some current flow. In practical terms
this means that the capacitor will block any DC voltages, except for transient effects. But,
high frequency AC currents will pass through the device. The equation for a capacitor and
schematic symbols are given in Figure 7.11.
circuits - 7.9
+
+
d
I = C ----- V = CDV
dt
C
V
Figure 7.11
Capacitors
The symbol on the left is for an electrolytic capacitor. These contain a special fluid
that increases the effective capacitance of the device but requires that the positive and negative sides must be observed in the circuit. (Warning: reversing the polarity on an electrolytic capacitor can make them leak, fail and possibly explode.) The other capacitor symbol
is for a regular capacitor, normally with values under a microfarad.
Figure 7.12
C=1uF
I
circuits - 7.10
7.2.4 Inductors
While a capacitor will block a DC current, an inductor will pass it. Inductors are
basically coils of wire. When a current flows through the coils, a magnetic field is generated. If the current through the inductor changes then the magnetic field must change, otherwise the field is maintained without effort (i.e., no voltage). Therefore the inductor
resists changes in the current. The schematic symbol and relationship for an inductor are
shown in Figure 7.13.
+
L
I
V
d
V = L ----- I = LDI
dt
Figure 7.13
An inductor
circuits - 7.11
Figure 7.14
L=1mH
I
7.2.5 Op-Amps
The ideal model of an op-amp is shown in Figure 7.15. On the left hand side are
the inverting and non-inverting inputs. Both of these inputs are assumed to have infinite
impedance, and so no current will flow. Op-amp application circuits are designed so that
the inverting and non-inverting inputs are driven to the same voltage level. The output of
the op-amp is shown on the right. In circuits op-amps are used with feedback to perform
standard operations such as those listed below.
adders, subtractors, multipliers, and dividers - simple analog math operations
amplifiers - increase the amplitude of a signal
impedance isolators - hide the resistance of a circuit while passing a voltage
circuits - 7.12
I-
V+
I+
V+
Figure 7.15
Vo
An ideal op-amp
A simple op-amp example is given in Figure 7.16. As expected both of the op-amp
input voltages are the same. This is a function of the circuit design. (Note: most op-amp
circuits are designed to force both inputs to have the same voltage, so it is normally reasonable to assume they are the same.) The non-inverting input is connected directly to
ground, so it will force both of the inputs to 0V. When the currents are summed at the
inverting input, an equation including the input and output voltages is obtained. The final
equation shows the system is a simple multiplier, or amplifier. The gain of the amplifier is
determined by the ratio of the input and feedback resistors.
circuits - 7.13
R2
R1
+
Vi
-
+
Vo
-
The voltage at the non-inverting input will be 0V, by design the voltage at the inverting input will be the same.
V + = 0V
V - = V + = 0V
The currents at the inverting input can be summed.
V- Vi V- Vo
---------------- + ----------------- = 0
I
=
V
R1
R2
0 Vi 0 Vo
-------------- + --------------- = 0
R1
R2
R 2 Vi
V o = -------------R1
R2
V o = --------- Vi
R1
Figure 7.16
circuits - 7.14
R2
Find the input/output ratio,
R1
+
+
Vi
-
Figure 7.17
R5
+
Vref
-
R3
R4
+
Vo
-
Op-amp example
For ideal op-amp problems the node voltage method is normally the best choice.
The equations for the circuit in Figure 7.17 and derived in Figure 7.18. The general
approach to this solution is to sum the currents into the inverting and non-inverting input
nodes. Notice that the current into the op-amp is assumed to be zero. Both the inverting
and non-inverting input voltages are then set to be equal. After that, algebraic manipulation results in a final expression for the op-amp. Notice that if all of the resistor values are
the same then the circuit becomes a simple subtractor.
circuits - 7.15
Note: normally node voltage methods work best with op-amp circuits, although others
can be used if the non-ideal op-amp model is used.
First sum the currents at the inverting and non-inverting op-amp terminals.
V+ Vi V + V o
----------------- + ------------------ = 0
I
=
V+
R1
R2
1
1
1
1
V + ------ + ------ = V i ------ + V o ------
R 1 R 2
R 1
R 2
R1 + R2
1
1
V + ------------------ = V i ------ + V o ------
R1 R2
R1
R2
R2
R1
V + = V i ------------------ + V o ------------------
R1 + R2
R 1 + R2
(1)
V - V ref V- + ------ = 0
I
=
V- ------------------R5
R4
1
1
1
V - ------ + ------ = V ref ------
R4 R5
R5
R4
V - = V ref ------------------
R4 + R5
(2)
Figure 7.18
(3)
circuits - 7.16
100KHz. When the op-amp is being used for high frequencies or large gains, the model of
the op-amp in Figure 7.19 should be used. This model includes a large resistance between
the inverting and non-inverting inputs. The voltage difference drives a dependent voltage
source with a large gain. The output resistance will limit the maximum current that the
device can produce, normally less than 100mA.
V
rn
+
A ( V+ V- )
Figure 7.19
Vo
ro
typically,
r n > 1M
5
A > 10
r o < 100
7.3 IMPEDANCE
Circuit components can be represented in impedance form as shown in Figure
7.20. When represented this way the circuit solutions can focus on impedances, Z,
instead of resistances, R. Notice that the primary difference is that the differential operator has been replaced. In this form we can use impedances as if they are resistances.
Device
Time domain
Impedance
Resistor
V ( t ) = RI ( t )
Z = R
Capacitor
1
V ( t ) = ---- I ( t ) dt
C
1
Z = -------DC
Inductor
d
V ( t ) = L ----- I ( t )
dt
Z = LD
Figure 7.20
circuits - 7.17
When representing component values with impedances the circuit solution is done
as if all circuit components are resistors. An example of this is shown in Figure 7.21.
Notice that the two impedances at the right (resistor and capacitor) are equivalent to two
resistors in parallel, and the overall circuit is a voltage divider. The impedances are written
beside the circuit elements.
DL
t=0sec
50VDC
+
-
R
1/DC
+
Vo
-
50VDC
1
1
R
-----------------------= ------------------ = ---------------------1 - --11
RCD + 1
------------+
DC + --R
R
1 -
------ DC
+
-
R
D RLC + DL + R
DL + ----------------------
1 + DCR
Figure 7.21
circuits - 7.18
coefficient units
C
As
-----V
Vs
-----A
V
--A
The circuit in Figure 7.22 could be solved with two loops, or two nodes. An arbitrary decision is made to use the current loop method. The voltages around each loop are
summed to provide equations for each loop.
circuits - 7.19
R1
+
V
-
L
I1
I2
+
Vo
-
R2
VL1
= V + R 1 I 1 + L ( DI 1 DI 2 ) = 0
( R 1 + LD )I 1 = V + ( LD )I 2
(1)
V
LD
I 1 = -------------------- + -------------------- I 2
R 1 + LD
R 1 + LD
(2)
I2
------- + R2 I2 = 0
V
=
L
(
DI
DI
)
+
L2
2
1
CD
I2
L ( DI 2 DI 1 ) + -------- + R 2 I 2 = 0
CD
1
( DL )I 1 = LD + -------- + R2 I 2
CD
R2
1
I 1 = 1 + --------------2 + -------- I 2
LD
CLD
Figure 7.22
(3)
(4)
Example problem
The equations in Figure 7.22 are manipulated further in Figure 7.23 to develop an
input-output equation for the second current loop. This current can be used to find the current through the output resistor R2. The output voltage can then be found by multiplying
the R2 and I2.
circuits - 7.20
First, sum the voltages around the loops and then eliminate I1.
R2
V
LD
1
I 1 = -------------------- + -------------------- I 2 = 1 + --------------2 + -------- I 2
R 1 + LD R 1 + LD
LD
CLD
R2
V 1
LD
------------------= 1 + --------------2 + -------- -------------------- I 2
R 1 + LD
LD R 1 + LD
CLD
2
( CLD + 1 + CDR2 ) ( R 1 + LD ) CL D
V ------------------= -------------------------------------------------------------------------------------------------- I2
2
R 1 + LD
CLD
2
CLD
- V
I 2 = -------------------------------------------------------------------------------------------------------------------------------- ( R 1 + LD ) ( ( CLD 2 + 1 + CDR 2 ) ( R 1 + LD ) CL 2 D 3 )
2
CLD
- V
I 2 = -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- CL ( R1 + R 2 )D 3 + L ( CR 21 + 2CR 1 R 2 + L )D 2 + R 1 ( CR1 R 2 + 2L )D + ( R 12 )
Figure 7.23
The equations can also be manipulated into state equations, as shown in Figure
7.24. In this case a dummy variable is required to replace the two first derivatives in the
first equation. The dummy variable is used in place of I1, which now becomes an output
variable. In the remaining state equations I1 is replaced by q1. In the final matrix form the
state equations are in one matrix, and the output variable must be calculated separately.
circuits - 7.21
State equations can also be developed using equations (1) and (3).
R 1 I 1 + LI 1 = V + LI 2
(1) becomes
LI 1 LI 2 = V R 1 I 1
V R1
I 1 I 2 = --- ------ I 1
L L
q1 = I1 I2
(10)
V R1
q 1 = --- ------ I 1
L L
(11)
I1 = q1 + I2
V R1
q 1 = --- ------ ( q 1 + I 2 )
L L
R1
R1
V
q 1 = q 1 ------ + I 2 ------ + --L
L
L
I
LI 2 LI 1 + ---2- + R 2 I 2 = 0
C
I2
V R1 I 1 + ---- + R 2 I 2 = 0
C
I2
V R 1 ( q1 + I 2 ) + ---- = R 2 I 2
C
(3) becomes
1
I 2 = I 2 R 1 + ---- + q 1 ( R 1 ) + V
C
These can be put in matrix form,
d- q 1
---dt I
2
Figure 7.24
R1
-----L
=
R1
-----L
1
R 1 R1 + ---C
q1
I2
V
--+ L
V
(12)
(13)
I1 = 1 1
q1
I2
circuits - 7.22
Evaluate the circuit in Figure 7.22 using the node voltage method.
Figure 7.25
circuits - 7.23
Vin
Figure 7.26
Ri
+
Vout
circuits - 7.24
The circuit in Figure 7.27 can be evaluated as a voltage divider when the capacitor
is represented as an impedance. In this case the result is a first-order differential equation.
t=0
V s = 3 cos t
+
-
+
Vo
-
As normal we relate the source voltage to the output voltage. The we find the values
for the various terms in the frequency domain.
ZR
V o = Vs ----------------- Z R + Z C
where,
ZR = R
1
Z C = -------DC
V o = Vs ------------------
1
R + --------
DC
CRD
V o = Vs ----------------------
CRD + 1
V o ( CRD + 1 ) = V s ( CRD )
V o ( CR ) + V o = Vs ( CR )
V o + Vo -------- = V s
CR
Figure 7.27
circuits - 7.25
-------
1
CR
yields
V o + Vo -------- = 0
V
=
C
e
h
1
CR
Next, the particular solution can be determined, starting with a guess.
1
d
V o + Vo -------- = ----- ( 3 cos t ) = 3 sin t
CR
dt
V p = A sin t + B cos t
V p' = A cos t B sin t
1
( A cos t B sin t ) + ( A sin t + B cos t ) -------- = 3 sin t
CR
1
A + B -------- = 0
CR
1
A = B --------
CR
1
B + A -------- = 3
CR
1
1
B + B -------- -------- = 3
CR CR
1
B ----------- 2 2 + 1 = 3
C R
2 2
3C 2 R 2 1
3C R 3 CR B = --------------------- -------- = --------------------A
=
--------------------2 2
2 2 CR
2 2
1+C R
1 + C R
1+C R
B
2
2
A + B sin t + atan ---
A
The homogeneous and particular solutions can now be combined. The system will be
assumed to be at rest initially.
Vp =
t
-------CR
B
2
2
+ A + B sin t + atan ---
A
B
0
2
2
0 = C 1 e + A + B sin 0 + atan ---
A
V 0 = Vh + V p = C1 e
B
2
2
C 1 = A + B sin 0 + atan ---
Figure 7.28
circuits - 7.26
voltage/current increases
Figure 7.29
The basic equivalent circuit model is shown in Figure 7.30, includes the rotational
inertia of the rotor and any attached loads. On the left hand side is the resistance of the
motor and the back emf dependent voltage source. On the right hand side the inertia
components are shown. The rotational inertia J1 is the motor rotor, and the second inertia
is an attached disk.
circuits - 7.27
I
Voltage
Supply
J1
J2
T,
+
Vm +
-
Vs
-
Because a motor is basically wires in a magnetic field, the electron flow (current) in
the wire will push against the magnetic field. And, the torque (force) generated
will be proportional to the current.
Tm
T m = KI
I = -----K
Next, consider the power in the motor,
P = V m I = T = KI
V m = K
Consider the dynamics of the rotating masses by summing moments.
M
Figure 7.30
d
= T m T load = J -----
dt
d
T m = J ----- + T load
dt
These basic equations can be manipulated into the first-order differential equation
in Figure 7.31.
The current-voltage relationship for the left hand side of the equation can be written and manipulated to relate voltage and angular velocity.
Vs V m
I = -----------------R
Tm
V s K
------ = -------------------R
K
d
J ----- + T load
dt
V s K
------------------------------------ = ------------------R
K
2
T load
d
K
K
----- + ------ = Vs ------ -----------dt
JR
JR
J
Figure 7.31
circuits - 7.28
Ls
Rs
Lr
Lm
Rr
IL
RL
1f
R L = ---------- R r
f
Figure 7.32
The torque relationship for AC motors is given in Figure 7.33. These can be combined with the equivalent circuit model to determine the response of the motor to a load.
circuits - 7.29
d
= T rotor T load = J -----
dt
Va
windings on stator
Vb
N
Vc
Figure 7.33
S
permanent
magnet rotor
circuits - 7.30
The basic relationships for brushless DC motors are given in Figure 7.34.
d
V t = R m + ----- L I m + E
dt
E = Ke
T = Kt Im
where,
V t = terminal voltage across motor windings
R m = resistance of a motor winding
L = phase to phase inductance
I m = current in winding
E = back e.m.f. of motor
K e = motor speed constant
= motor speed
Figure 7.34
circuits - 7.31
d
T
V t = R m + ----- L ----- + K e
dt K t
d
M = T Tload = J ---dt
d
T = J ----- + T load
dt
where,
J = combined moments of inertia for the rotor and external loads
T load = the applied torque in the system
d
J ----- + T load
d
dt
V t = R m + ----- L ------------------------------- + K e
Kt
dt
JR m d
Rm
LJ d 2
L d
V t = ---------- ----- + ------ ----- + ------- T load + ----- ----- T load + K e
K t dt
K t dt
Kt
K t dt
d 2
d
d
( LJ ) ----- + ( JR m ) ----- + K e K t = K t V t L ----- T load R m T load
dt
dt
dt
Kt V
Rm d
Ke Kt
Rm T load
d- 2
d ---- ----- + ----------- = ----------t 1--- ---T load ------------------ + ----- dt
L dt
LJ
LJ J dt
LJ
Figure 7.35
To rotate the motor at a constant velocity the waveform in Figure 7.36 would be
applied to each phase. Although each phase would be 120 degrees apart for a three pole
motor. A more sophisticated motor controller design would smooth the waves more to
approach a sinusoidal shape.
Vt
t(ms)
circuits - 7.32
Figure 7.36
7.6 FILTERS
Filters are useful when processing data signals. Low pass often used to eliminate
noise, high pass filters eliminate static signals and leave dynamic signals. Band pass filters
reject all frequencies outside a desired frequency band. A low pass filter is shown in Figure 7.37. At high frequencies the capacitor, C, has a very low impedance, and grounds the
input signal. At low frequencies the capacitor impedance is high, increasing the gain of the
op-amp circuit. This is easier to conceptualize if the R1-C pair are viewed as a voltage
divider.
Rf
Ri
+
R1
+
Vi
-
Figure 7.37
+
Vo
-
Low-Pass Filter
A high pass filter is shown in Figure 7.38. In this case the voltage divider in the
previous circuit is reversed. In this circuit the gain will increase for signals with higher frequencies.
circuits - 7.33
Rf
Ri
+
C
+
Vi
-
Figure 7.38
+
Vo
-
High-Pass Filter
V
P = IV = I R = -----R
2
Figure 7.39
E = Pt
7.8 SUMMARY
Basic circuit components are resistors, capacitors, inductors op-amps.
node and loop methods can be used to analyze circuits.
circuits - 7.34
R
+
Vi
+
-
Vo
R
+
Vi
+
-
Vo
-
L
Vi
Vo
R
-
circuits - 7.35
5. Develop differential equations and the input-output equation for the electrical system below.
R1
Vi
+
-
+
C
Vo
R2
-
6. Consider the following circuit. Develop a differential equation for the circuit.
L
Vi
R1
C
R2
R3
Vo
-
7. Find the input-output equation for the circuit below, and then find the natural frequency and
damping coefficient.
R1
R2
Vi
+
-
+
Vo
-
8. a) Find the differential equation for the circuit below where the input is Vi, and the output is Vo.
L
+
Vi
+
-
Vo
-
circuits - 7.36
c) Solve the differential equation found in part b) using the numerical values given
below. Assume at time t=0, the circuit has the voltage Vo and the first derivative
shown below.
L = 10mH
C = 1F
R = 1K
V i = 10V
V
at t=0s
V o = 2V
V o = 3 --s
9.
a) Write the differential equations for the system pictured below.
b) Put the equations in input-output form.
R2
R1
Vi
Vo
10. Given the circuit below, find the ratio of the output over the input (this is also known as a
transfer function). Simplify the results.
R2
Vi
R1
C
+
Vo
circuits - 7.37
11. Examine the following circuit and then derive the differential equation.
L
R2
R1
+
Vi
+
Vo
12. Examine the following circuit and then derive the differential equation.
L
R2
R1
+
Vi
+
Vo
13.
a) Find the differential equation for the circuit below.
1000
1F
0.001H
1000
+
Vi
+
-
+
10
Vo
-
circuits - 7.38
b) Put the differential equation in state variable form and a numerical method to
produce a detailed sketch of the output voltage Vo. Assume the system starts at
rest, and the input is Vi=5V.
14.
a) Write the differential equations for the system pictured below.
b) Put the equations in state variable form.
c) Use numerical methods to find the ratio between input and output voltages for a
range of frequencies. The general method is put in a voltage such as
Vi=1sin(___t), and see what the magnitude of the output is. Divide the magnitude of the output sine wave by the input magnitude. Note: This should act as a
high pass or low pass filter.
d) Plot a graph of gain against the frequency of the input.
R2
R1
Vi
Vo
C=1uF
R1=1K
R2=1K
R series = R 1 + R 2
C 1 C2
C series = ------------------C1 + C 2
C parallel = C 1 + C 2
L1 L2
L parallel = -----------------L1 + L2
L series = L 1 + L 2
circuits - 7.39
2.
Vo
------ = 1--5
Vi
3.
R
1
1
V o + V o --- + V o ------- = V i -------
L
LC
LC
4.
a)
1
1
1
V o + Vo -------- + V o ------- = V i -------
RC
LC
LC
b)
V o = Xo
1
1
1
X o = Xo -------- + V o ------- + V i -------
RC
LC
LC
5.
1
1
1
1
V o + V o ---------- + ---------- + V o ------- = V i ----------
CR 1 CR 2
LC
CR 1
6.
R2
R2 R 3
R2 R3 C + L
1
1
1
1
V o + Vo ---------- + ---------- + V o ------- = V i ----------
CR 1 CR2
LC
CR 1
b)
n =
1-----LC
L ( R1 + R2 )
= ------------------------------2 CR 1 R 2
circuits - 7.40
8.
(a.
(b.
1
1
1
V o + V o -------- + V o ------- = V i -------
CR
LC
LC
circuits - 7.41
(c.
2 6
10
V o ( 10 10 ) + V o --------- + V o = 10
10 3
8
5
V o ( 10 ) + V o ( 10 ) + V o = 10
L
V o ( LC ) + V o --- + V o = V i
R
3
8
9
V o + V o ( 10 ) + V o ( 10 ) = ( 10 )
homogeneous;
2 At
At
At
A e + Ae ( 10 ) + e ( 10 ) = 0
3 2
10 ( 10 ) 4 ( 10 )
A + A ( 10 ) + ( 10 ) = 0
A = -------------------------------------------------------------- = 500 9987j
2
500t
Vh = C1 e
cos ( 9987t + C2 )
2
particular;
guess
3
( 0 ) + ( 0 ) ( 10 ) + ( A ) ( 10 ) = ( 10 )
Vp = A
A = 10
V p = 10
for initial conditions,
Vo = C1 e
500t
cos ( 9987t + C 2 ) + 10
500 ( 0 )
2 = C1 e
cos ( 9987 ( 0 ) + C 2 ) + 10
8 = C 1 cos ( C 2 )
(1)
500t
500t
V o = 500C 1 e
cos ( 9987t + C 2 ) 9987C 1 e
sin ( 9987t + C 2 )
for t=0, d/dt Vo=3V
V o = 8.01 e
500t
C 2 = 0.050
circuits - 7.42
9.
a)
R1
V i + V i ( R 1 C ) + Vo ------ = 0
R 2
b)
R 2
Vo = V i ( CR 2 ) + Vi ---------
R1
10.
Vo
R2
------ = --------------------------------Vi
R 1 + DR 1 R 2 C
11.
R
L
V o = Vi ------ + V i --------2-
R 1
R1
12.
R2
R2
V o + Vo ------ = V i ---------
L
R1
13.
(a.
Create a node between the inductor and resistor Va, and use the node voltage method
IV
( VA Vi ) ( VA V )
= ---------------------- + -------------------------- = 0
0.001D
1000
V = V
1000000 ( V A V i ) + V A D = 0
= 0V
XXXXADD UNITSXXXXX
1000000
V A = V i -------------------------------
1000000 + D
( V VA ) ( V Vo )
- + ------------------------- + ( V V o ) ( 0.000001D ) = 0
=
I
V
------------------------1000
1000
Vo )
(---------- 1 )- 1000000 (------------- + ( V o ) ( 0.000001D ) = 0
V i ------------------------------- +
1000 1000000 + D 1000
1000000
V o ( 1 0.001D ) = V i -------------------------------
1000000 + D
2
circuits - 7.43
(b.
d
---Vo = Vo
dt
d-
14.
a)
b)
c)
R2
V o = Vi ( CR 2 ) + V i ---------
R1
Not a state equation
assume Vi=1sin(wt)
Vo =
1
2
2
1 + ( 1000 ) sin t + atan ----------------
1000
d)
R
+
Vi
-
+
Vo
-
circuits - 7.44
R1
+
Vi
+
Vo
+
Vo
4. Develop the differential equation(s) for the system below, and use them to find the response to
the following inputs. Assume that the circuit is off initially.
R3
+
Vi
R1=R2=R3=R4=1Kohm
C=1uF
a)
V i = 5 sin ( 100t )
b)
c)
V i = 5 sin ( 1000000t )
Vi = 5
+
-
Vo
C
R4
R1
R2
circuits - 7.45
+
-
C2
R2
R1
Vo
6. Study the circuit below. Assume that for t<0s the circuit is discharged and off. Starting at t=0s
an input of Vi=5sin(100,000t) is applied.
1K
Vi
+
1uF
+
Vo
-
a) Write a differential equation and then a transfer function describing the circuit.
b) Find the output of the circuit using explicit integration (i.e., homogeneous and
particular solutions).
Topics:
Transfer functions, block diagrams and simplification
Feedback controllers
Control system design
Objectives:
To be able to represent a control system with block diagrams.
To be able to select controller parameters to meet design objectives.
8.1 INTRODUCTION
Every engineered component has some function. A function can be described as a
transformation of inputs to outputs. For example it could be an amplifier that accepts a signal from a sensor and amplifies it. Or, consider a mechanical gear box with an input and
output shaft. A manual transmission has an input shaft from the motor and from the
shifter. When analyzing systems we will often use transfer functions that describe a system as a ratio of output to input.
An example
output
----------------- = f ( D )
input
x
4+D
--- = --------------------------------2
F
D + 4D + 16
Figure 8.1
If both sides of the example were inverted then the output would become F, and
the input x. This ability to invert a transfer function is called reversibility. In reality
many systems are not reversible.
There is a direct relationship between transfer functions and differential equations.
This is shown for the second-order differential equation in Figure 8.2. The homogeneous
equation (the left hand side) ends up as the denominator of the transfer function. The nonhomogeneous solution ends up as the numerator of the expression.
f
2
x + 2n x + n x = ----M
2
2
f
xD + 2 n xD + n x = ----M
2
2
f
x ( D + 2 n D + n ) = ----M
1-
--- M
x
-- = -------------------------------------------2
2
f
D + 2 n D + n
Figure 8.2
particular
homogeneous
The transfer function for a first-order differential equation is shown in Figure 8.3.
As before the homogeneous and non-homogeneous parts of the equation becomes the
denominator and the numerator of the transfer function.
1
x + --- x = f
1
xD + --- x = f
1
x D + --- = f
x
1
-- = ------------f
1
D + --
Figure 8.3
Control variable
INPUT
(e.g. gas)
vdesired
verror
+
OUTPUT
(e.g. velocity)
SYSTEM
(e.g. a car)
control
function
gas
car
vactual
Note: The arrows in the diagram indicate directions so that outputs and inputs
are unambiguous. Each block in the diagram represents a transfer function.
Figure 8.4
There are two main types of feedback control systems: negative feedback and positive feedback. In a positive feedback control system the setpoint and output values are
added. In a negative feedback control the setpoint and output values are subtracted. As a
rule negative feedback systems are more stable than positive feedback systems. Negative
feedback also makes systems more immune to random variations in component values and
inputs.
The control function in Figure 8.4 can be defined many ways. A possible set of
rules for controlling the system is given in Figure 8.5. Recall that the system error is the
difference between the setpoint and actual output. When the system output matches the
setpoint the error is zero. Larger differences between the setpoint and output will result in
larger errors. For example if the desired velocity is 50mph and the actual velocity 60mph,
the error is -10mph, and the car should be slowed down. The rules in the figure give a general idea of how a control function might work for a cruise control system.
In following sections we will examine mathematical control functions that are easy
to implement in actual control systems.
de
u = K p e + K i edt + K d ------
dt
Figure 8.6
Figure 8.7 shows a basic PID controller in block diagram form. In this case the
potentiometer on the left is used as a voltage divider, providing a setpoint voltage. At the
output the motor shaft drives a potentiometer, also used as a voltage divider. The voltages
from the setpoint and output are subtracted at the summation block to calculate the feedback error. The resulting error is used in the PID function. In the proportional branch the
error is multiplied by a constant, to provide a longterm output for the motor (a ballpark
guess). If an error is largely positive or negative for a while the integral branch value will
become large and push the system towards zero. When there is a sudden change occurs in
the error value the differential branch will give a quick response. The results of all three
branches are added together in the second summation block. This result is then amplified
to drive the motor. The overall performance of the system can be changed by adjusting the
gains in the three branches of the PID function.
proportional
PID function
Kp ( e )
integral
e
+
-
Figure 8.7
Ki ( e )
+V
+
u
amp
motor
derivative
d
K d ----- e
dt
-V
There are other variations on the basic PID controller shown in Figure 8.8. A PI
controller results when the derivative gain is set to zero. (Recall the second order
response.) This controller is generally good for eliminating long term errors, but it is prone
to overshoot. In a P controller only the proportional gain in non-zero. This controller will
generally work, but often cannot eliminate errors. The PD controller does not deal with
longterm errors, but is very responsive to system changes.
For a PI Controller
gas = K p v error + K i v error dt
For a P Controller
gas = K p v error
For a PD Controller
dv error
gas = K p v error + K d ----------------
dt
Figure 8.8
Aside: The manual process for tuning a PID controlled is to set all gains to zero. The proportional gain is then adjusted until the system is responding to input changes without
excessive overshoot. After that the integral gain is increased until the longterm errors
disappear. The differential gain will be increased last to make the system respond
faster.
c
+
G( D)
c
is equal to
H( D)
G( D)
-----------------------------------1 + G ( D )H ( D )
r = G ( D )e
r = G ( D ) ( c H ( D )r )
r ( 1 + G ( D )H ( D ) ) = G ( D )c
G( D)
r = ------------------------------------- c
1 + G ( D )H ( D )
Figure 8.9
Other block diagram equivalencies are shown in Figure 8.10 to Figure 8.16. In all
cases these operations are reversible. Proofs are provided, except for the cases where the
equivalence is obvious.
c
+
G( D)
c
is equal to
G( D)
-----------------------------------1 G ( D )H ( D )
r = G ( D )e
H( D)
r = G ( D ) ( c + H ( D )r )
r ( 1 G ( D )H ( D ) ) = G ( D )c
G( D)
r = ------------------------------------- c
1 G ( D )H ( D )
Figure 8.10
G( D)
is equal to
1
------------G( D)
r = G ( D )c
1
c = r -------------
G(D )
Figure 8.11
G( D)
G( D)
is equal to
G( D)
r
Figure 8.12
G( D)
is equal to
H( D)
G ( D )H ( D )
x = G ( D )c
b = H ( D )x
b = H ( D )G ( D ) )c
Figure 8.13
G( D)
1
------------G( D)
is equal to
c
Figure 8.14
G( D)
G( D)
c
is equal to
G( D)
b
e = G ( D )c b
b
e = G ( D ) c -----------------
G ( D )
Figure 8.15
1
------------G( D)
G( D)
G(D)
is equal to
d
r = G ( D ) ( c d )
G( D)
r = cG ( D ) dG ( D )
Figure 8.16
Recall the example of a cruise control system for an automobile presented in Figure 8.4. This example is extended in Figure 8.17 to include mathematical models for each
of the function blocks. This block diagram is first simplified by multiplying the blocks in
sequence. The feedback loop is then reduced to a single block. Notice that the feedback
line doesnt have a function block on it, so by default the function is 1 - everything that
goes in, comes out.
Ki
K p + ----- + K d D
D
verror
Ki
K p + ----- + K d D
D
vdesired
+
verror
gas
10
gas
10-------MD
1-------MD
vactual
vactual
= G(D)
vdesired
verror
+
10-
K + K
------------i
p D + K d D MD
vactual
H(D) = 1
vdesired
Figure 8.17
10-
K + K
------------i
p D + K d D MD
-------------------------------------------------------------------------Ki
10
1 + K p + ----- + K d D --------- ( 1 )
MD
D
vactual
The function block is further simplified in Figure 8.18 to a final transfer function
for the whole system.
K + K
-----i
p D + K d D
-------------------------------------------------------
K
Ms
- + K p + -----i + K d D
-----
10
D
vdesired
vdesired
2
D ( Kd ) + D ( K p ) + ( Ki )
----------------------------------------------------------------------
M
D 2 ----
10- + K d + D ( K p ) + ( K i )
vactual
vactual
2
v actual
D ( K d ) + D ( Kp ) + ( Ki )
----------------- = ----------------------------------------------------------------------
v desired
M
D 2 ----
10- + K d + D ( K p ) + ( K i )
Figure 8.18
2.2K
1K
12Vdc motor
+
LM675
op-amp
shafts are coupled
+5V
-5V
5K potentiometer
desired position
voltage Vd
Figure 8.19
The feedback controller can be represented with the block diagram in Figure 8.20.
desired
position
voltage +
Vd
gain Kp
op-amp
motor
potentiometer
Figure 8.20
shaft
The transfer functions for each of the blocks are developed in Figure 8.21. Two of
the values must be provided by the system user. The op-amp is basically an inverting
amplifier with a fixed gain of -2.2 times. The potentiometer is connected as a voltage
divider and the equation relates angle to voltage. Finally the velocity of the shaft is integrated to give position.
IV+
V+ Vi V + V o
= ----------------- + ------------------ = 0
1K
2.2K
V + = V - = 0V
Vi V o
-------- + ----------- = 0
1K 2.2K
V
-----o- = 2.2
Vi
For the potentiometer assume that the potentiometer has a range of 10 turns and 0
degrees is in the center of motion. So there are 5 turns in the negative and positive
direction.
V o = 5V --------------
5 ( 2 )
V
-----o- = 0.159Vrad 1
Figure 8.21
Transfer functions for the power amplifier, potentiometer and motor shaft
The basic equation for the motor is derived in Figure 8.22 using experimental data.
In this case the motor was tested with the full inertia on the shaft, so there is no need to
calculate J.
For the motor use the differential equation and the speed curve when Vs=10V is applied:
2
1400 RPM
d-
K
K
--- + ------ = ------ V s
dt
JR
JR
1s
2s
3s
For steady-state
d-
1
--- = 0
= 1400RPM = 146.6rads
dt
2
K
K
0 + ------ 146.6 = ------ 10
JR
JR
K = 0.0682
0.8s
1s
2
1
K -
----= -- JR
1
K
0.0682 ------ = ---------JR
0.8s
K----= 18.328
JR
1
D + ------- = 18.33V s
0.8
18.33
---= --------------------Vs
D + 1.25
Figure 8.22
The individual transfer functions for the system are put into the system block diagram in Figure 8.23. The block diagram is then simplified for the entire system to a single
transfer function relating the desired voltage (setpoint) to the angular position (output).
The transfer function contains the unknown gain value Kp.
Vd
Kp
18.33
--------------------D + 1.25
-2.2
1--D
0.159
Vd
Vd
Vd
Vd
Figure 8.23
The value of Kp can be selected to tune the system performance. In Figure 8.24
the gain value is calculated to give the system an overall damping factor of 1.0, or critically damped. This is done by recognizing that the bottom (homogeneous) part of the
transfer function is second-order and then extracting the damping factor and natural frequency. The final result of Kp is negative, but this makes sense when the negative gain
on the op-amp is considered.
6.412K p = n
6.412K p = 0.625
K p = 0.0609
Figure 8.24
Calculating a gain Kp
r
G
System error,
S = cr
Feedback error,
e = cb
+
-
b
H
Figure 8.25
Controller errors
Given,
Kp
G ( D ) = -----D
H( D) = 1
Kp
r
G
-- = ------------------ = ---------------c
1 + GH
D + Kp
r + K p c = K p c
c = A
The homogeneous solution is,
rh = C 1 e
Kp t
rp = 0
0 + KpC2 = KpA
C2 = A
The solutions can be combined and the remaining unknown found for the system at
rest initially.
r = rh + r p = C 1 e
Kp t
+A
0 = C1e + A
C1 = A
r = A e
Kp t
+A
Kp t
+ A ) = Ae
Kp t
Figure 8.27
c = At
Figure 8.28
H(D) = 5
c = 4t
Transfer Function
Type
Proportional (P)
Gc = K
Proportional-Integral (PI)
1
G c = K 1 + -------
D
Proportional-Derivative (PD)
G c = K ( 1 + D )
Proportional-Integral-Derivative (PID)
1
G c = K 1 + ------- + D
D
Lead
1 + D
G c = K --------------------
1 + D
>1
Lag
1 + D
G c = K --------------------
1 + D
>1
Lead-Lag
1 + 1 D 1 + 2 D
- ----------------------G c = K --------------------- 1 + 1 D 1 + 2 D
Figure 8.29
>1
1 > 2
feedforward
function
feedback
function
Figure 8.30
process
d---X = AX + BU
dt
Y = CX + D
d---X
dt
+
B
1--D
Figure 8.31
An example is shown in Figure 8.32 that implements a second order state equation.
The system uses two integrators to integrate the angular acceleration, then the angular
velocity, to get the position.
d---
dt
1--D
d---
dt
1--D
d--- =
dt
d--- = A + B + C
dt
Figure 8.32
The previous block diagrams are useful for simulating systems. These can then be
used in feedforward control systems to estimate system performance and then predict a
useful output value.
The model,
2
K
K
+ ------ = V s ------
JR
JR
JR
V s = ------ + ( K )
K
d---
dt
JR
-----K
+
K
K
K P + -----i + DK d
D
Feedback
Plant
Figure 8.33
Gc1
Figure 8.34
Gc2
Gp1
Gp2
A cascade controller
8.4 SUMMARY
Transfer functions can be used to model the ratio of input to output.
Block diagrams can be used to describe and simplify systems.
Controllers can be designed to meet criteria, such as damping ratio and natural
frequency.
System errors can be used to determine the long term stability and accuracy of a
controlled system.
Other control types are possible for more advanced systems.
F
M
B
2. Develop a transfer function for the system below. The input is the force F and the output is the
voltage Vo. The mass is suspended by a spring and a damper. When the spring is undeflected
y=0. The height is measured with an ultrasonic proximity sensor. When y = 0, the output
Vo=0V. If y=20cm then Vo=2V and if y=-20cm then Vo=-2V. Neglect gravity.
Ks
N
K s = 10 ---m
Kd
s
K d = 5N ---m
M = 0.5Kg
Ultrasonic
Proximity
Sensor
3. Find the transfer functions for the systems below. Here the input is a torque, and the output is
the angle of the second mass.
1
J1
K s1
B1
K s2
J2
B2
4. Find the transfer functions for the system below where Vi is the input and Vo is the output.
C
Vi
R1
R2
Vo
-
5. Given the transfer function, G(s), determine the time response output Y(t) to a step input X(t).
4
Y
G = ------------- = --X ( t ) = 20 When t >= 0
D+2
X
6. Given the transfer function below, develop a mechanical system that it could represent. (Hint:
Differential Equations).
x(D)
1
------------- = ----------------------F(D )
10 + 20D
where:
x = displacement
F = force
7. Given a mass supported by a spring and damper, find the displacement of the supported mass
over time if it is released from neutral at t=0sec, and gravity pulls it downward.
a) develop a transfer function for y/F.
b) find the input function F.
c) solve the input output equation to find an explicit equation of the position as a
function of time for Ks = 10N/m, Kd = 5Ns/m, M=10kg.
d) solve part c) numerically.
8. a) What is a Setpoint, and what is it used for? b) What does feedback do in control systems?
9. The block diagram below is for a servo motor position control system. The system uses a proportional controller.
a) Draw a sketch of what the actual system might look like. Identify components.
d
V
V- d
-------2
rad +
Ve
Vm
10 - rad
-------------------D + 1 sV
Va
V
2 --------rad
1--D
A
x
B
-
+
C
11. Simplify the block diagram below. (Note: Vn is an input and cannot be combined in a transfer
function.)
+
+
A
Vi
Vo
+
Vn
12. a) Develop an equation for the system below relating the two inputs to the output. Put it in
block diagram format. (Hint: think of a summation block.)
R
Vf
Ve
R
Vd
+
R
b) Develop an equation for the system below relating the input to the output. Put
1K
1K
-
Vs
1K
Ve
+
RP
c) The equation below can be used to model a permanent magnet DC motor with
an applied torque. An equivalent block diagram is given. Prove that the block
diagram is equivalent to the equation.
2
Km
T load
Km
d-
--- + ------------- = V s ------------- ----------- dt
J R R M
JR
J R R M
Vs
K
------mRM
e
-
RM
--------------------------------2
JR RM D + Km
T load
d) Write the transfer function for the system below relating the input torque to the
output angle theta2. Then write the transfer function for the angular velocity of
mass 2.
1
2
K s1
J1
J2
K s2
B
K
------mRM
2R P
-------------------3
R P + 10
+
-
RM
---------------------------------2
JM RM D + Km
K s2
----------------------------------------------------------------------------------------------------------------------3
2
D ( J 1 J 2 ) + D ( BJ 1 ) + D ( Ks2 ( J 1 + J 2 ) ) + ( BK s2 )
KT
13. Find the system output, feedback error and system error when the input is a ramp with the
function c(t) = 0.5t. Sketch the system errors as a function of time.
C(D)
5 ---------------------D( D + 1)
R( D)
1 + 0.2D
14. Given the block diagram below, select a system gain K that will give the overall system a
damping ratio of 0.7 (for a step input). What is the resulting damped frequency of the system?
+
2
-----------------------------2
D + 3D + 2
15. The following system is a feedback controller for an elevator. It uses a desired height d provided by a user, and the actual height of the elevator h. The difference between these two is
called the error e. The PID controller will examine the value e and then control the speed of
the lift motor with a control voltage c. The elevator and controller are described with transfer
functions, as shown below. all of these equations can be combined into a single system transfer
equation as shown.
e = dh
error
2
K
+ 1 + D - PID controller
-----------------------------r- = K p + -----i + K d D = 2D
D
e
D
combine the transfer functions
2
h--10 = ---------------2
r
D +D
elevator
10
10 ( D + 1 )
2D + 1 + D - ---------------10 (D + 1)
-r- h--- = h--- = ----------------------------= --------------------- ----------------------- = -----------------------2
2
e r
D D(D + 1)
e
D
D
D +D
h 10 ( D + 1 )
eliminate e
----------= -----------------------2
dh
D
10
(
D
+
1 )
h = -----------------------(d h)
2
D
10 ( D + 1 )
10 ( D + 1 )
= -----------------------(d)
h 1 + -----------------------2
2
D
D
10 ( D + 1 )
-----------------------
2
h--D
10D + 10 system transfer function
= ---------------------------------- = ----------------------------------2
10 ( D + 1 )
d
D
+
10D
+
10
1 + -----------------------
2
D
a) Find the response of the final equation to a step input. The system starts at rest
on the ground floor, and the input (desired height) changes to 20 as a step input.
b) Calculate the damping coefficient and natural frequency of the results in part a).
16. Study the circuit below. Assume that for t<0s the circuit is discharged and off. Starting at t=0s
an input of Vi=5sin(100,000t) is applied.
1K
Vi
+
1uF
+
Vo
-
a) Write a differential equation and then a transfer function describing the circuit.
b) Find the output of the circuit using explicit integration (i.e., homogeneous and
particular solutions).
17. Write a C subroutine to implement the control system that is shown inside the dashed line. The
subroutine arguments are the setpoint, C, and the feedback value, F. the subroutine returns the
5D
+
C
1-----5D
Kp
5
F
2.
V
V
10 ----
10 ----
m
m
Vo y
Vo
------ = ------ --- = --------------------------------------------- = --------------------------------------------------------------2
F
yF
2
Ns
N
MD K d D K s
0.5kgD 5 ------ D 10 ---m
m
3.
2
K s2
----- = -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------4
3
2
D ( J 1 J 2 ) + D ( B 1 J 2 + B 2 J 1 ) + D ( K s2 ( J 1 + J 2 ) + B1 B 2 ) + D ( K s2 ( B 1 + B 2 ) )
4.
Vo
D ( R 1 R2 C )
------ = -------------------------------------------------------------------------------------------------2
Vi
D ( R 1 LC ) + D ( L + CR 1 R 2 ) + ( R 1 + R2 )
5.
y ( t ) = 40e
2t
+ 40
6.
N
K s = 10 ---m
F
Ns
K d = 20 -----m
7.
b)
1
--y- = ----------------------------------------2
F
D M + DK d + K s
F = Mg
c)
y ( t ) = 10.13e
d)
y( t + h) = y(t) + h 0 1
y(t) + 0
v( t + h)
v(t)
1 0.5 v ( t )
9.81
a)
0.25t
8.
a) A setpoint is a system input that correlates to a desired output for a system with
negative feedback control.
b) Feedback in a control system allows an out to be compared to an input. The difference, the error, is used to adjust the control variable and increase the accuracy.
9.
a)
20K
---------------------------------2
D + D + 20K
b)
10.
A+B
-----------------------------------1 + CB + A + B
11.
Vn
--1A
+
+
Vi
Vo
CA
------------1+A
12.
R
a) V + = V d ------------- = 0.5V d
R + R
IV
V- V f V - V e
= ---------------- + ----------------- = 0
R
R
V e = 2V - V f
V e = 2 ( 0.5V d ) V f
Ve = Vd Vf
Vd
Ve
Vf
V- 0 V- V s
- + ----------------- = 0
I
=
V- ------------3
3
10
10
b)
V - = 0.5V s
(1)
V + V e V+ 0
- + --------------- = 0
=
I
V+ ----------------3
RP
10
Ve
1
1
V+ -------3- + ------ = -------3
R P
10
10
substitute in (1)
V
R P + 1K
- = -------e0.5V s ------------------3
3
R P ( 10 )
10
V
2R P
-----s = -------------------3
Ve
R P + 10
Ve
Vs
c)
2R P
-------------------3
R P + 10
Km
T load
Km
d-
--- + ------------- = V s ------------- ----------- J R R M
dt
JR
J R R M
2
Km
K m T load
JR M
K m T load
J R R M D + K m R M
d)
K s2 ( 1 2 )
+
J1
= K s2 ( 1 2 ) = J 1 1 D
K s2 ( 1 2 ) = J 1 1 D
1 ( J 1 D + K s2 ) + 2 ( K s2 ) =
K s2 ( 1 2 )
+
J2
= K s2 ( 1 2 ) B2 D = J 2 2 D
K s2 ( 1 2 ) B2 D = J 2 2 D
B 2 D
(1)
2
2 ( J 2 D + K s2 + BD ) = 1 ( K s2 )
2
J 2 D + K s2 + BD
1 = 2 -----------------------------------------
K s2
J 2 D + BD + K s2
2
2 ---------------------------------------- ( J 1 D + K s2 ) + 2 ( K s2 ) =
K s2
D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + D ( BK s2 )
- =
2 -----------------------------------------------------------------------------------------------------------------------------K s2
2
K s2
----- = -----------------------------------------------------------------------------------------------------------------------------4
3
2
D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + D ( BK s2 )
2 = D 2
K s2
-----2- = ----------------------------------------------------------------------------------------------------------------------3
2
D ( J 1 J 2 ) + D ( BJ 1 ) + D ( K s2 ( J 1 + J 2 ) ) + ( BK s2 )
e)
D ( ) + D ( ) + D( ) + ( ) ----= -----------------------------------------------------------------------------------------4
3
2
Vd
D ( ) + D ( ) + D ( ) + D( ) + ( )
(2)
13.
15
1 1
t
R ( t ) = 0.224e cos ---------- t 0.464 --- + ----2
5 2t
15
1
t
S ( t ) = 0.224e cos ---------- t 0.464 --2
5
15
1
t
e ( t ) = 0.112e cos ---------- t 0.464 + -----2
10
14.
rad
d = 1.530 --------s
K = 1.30
15.
a)
h ( t ) = 2.91e
b)
= 1.58
8.873t
22.91e
1.127t
+ 20
rad
n = 3.162 --------s
16.
a)
Vo
1000
------ = ---------------------Vi
D + 1000
b)
V o = 0.050e
1000t
17.
int controller(){
}
C
-
Vo
2. Given the transfer functions and input functions, F, use a numerical method to calculate the output of the system as a function of time for 0 to 0.5 seconds in 0.05 second intervals. Record the
values in a table.
2
D
--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
1 -----------D+1
1 -----------D+2
D + 3D + 5
+
+
K
---D
1
-----------------------------2
D + 3D + 5
10
D + 3D + 5
+
+
K--D
10
1
-----------------------------2
D + 3D + 5
G2
G3
G5
G4
G6
8. Simplify the following block diagram for a crane gantry position control system. The system
also includes a negative feedback loop to control the sway of an attached load
encoder
Cp
Cd
C pe
K enc
C pc
Vs
K pwm
K pp
G motor
C Lc
K sp
CL
K AD
VL
K pot
G load
9. Write a C subroutine to implement the control system that is shown inside the dashed line. The
subroutine arguments are the setpoint, C, and the feedback value, F. the subroutine returns the
new controller output value, U.
C
U
Kp + KdD
1 -------------------2
(D + 1)
Ki
----D
5
F
10. Write a C subroutine to implement the control system that is shown inside the dashed line. The
subroutine arguments are the setpoint, C, and the feedback value, F. the subroutine returns the
Ki
K p + K d D + ----D
1 -------------------2
(D + 1)
5
F
9. PHASOR ANALYSIS
Topics:
Phasor forms for steady state analysis
Complex and polar calculation of steady state system responses
Vibration analysis
Objectives:
To be able to analyze steady state responses using the phasor transform
9.1 INTRODUCTION
When a system is stimulated by an input it will respond. Initially there is a substantial transient response, that is eventually replaced by a steady state response. Techniques
for finding the combined steady state and transient responses were covered in earlier chapters. These include the integration of differential equations, and numerical solutions. Phasor analysis can be used to find the steady state response only. These techniques involve
using the phasor transform on the system transfer function, input and output.
D = + j
where,
D = differential operator
= decay constant
= oscillation frequency
Figure 9.1
Phasor transform
D = j
An example of the phasor transform is given in Figure 9.2. We start with a transfer
function for a mass-spring-damper system. In this example numerical values are assumed
to put the equation in a numerical form. The differential operator is replaced with j, and
the equation is simplified to a complex number in the denominator. This equation then
described the overall response of the system to an input based upon the frequency of the
input. A generic form of sinusoidal input for the system is defined, and also converted to
phasor (complex) form. (Note: the frequency of the input does not show up in the complex
form of the input, but it will be used later.) The steady state response of the system is then
obtained by multiplying the transfer function by the input, to obtain the output.
The response of the steady state output x can now be found for the given input.
x( )
1
- ( A cos input + jA sin input )
x ( ) = ------------ F ( ) = ---------------------------------------------------------------------2
F( )
( 2000 1000 ) + j ( 3000 )
Figure 9.2
To continue the example in Figure 9.2 values for the sinusoidal input force are
assumed. After this the method only requires the simplification of the complex expression.
In particular having a complex denominator makes analysis difficult and is undesirable.
To simplify this expression it is multiplied by the complex conjugate. After this, the
expression is quickly reduced to a simple complex number. The complex number is then
converted to polar form, and then finally back into a function of time.
rad
= 100 --------s
input = 0.5rad
-6
x ( ) = ( 0.863 10 ) + j ( 0.505 10 )
x( ) =
0.505 10-6
-6 2
-6 2
- +
( 0.863 10 ) + ( 0.505 10 ) atan ---------------------------- 0.863 10-6
Note: the signs of the components indicate that the angle is in the bottom left quadrant of the complex plane, so the angle should be between 180 and 270 degrees.
To correct for this pi radians are added to the result of the calculation.
-6
x ( ) = 0.9999 10 3.671
This can then be converted to a function of time.
-6
Figure 9.3
Note: when dividing and multiplying complex numbers in polar form the magnitudes can
be multiplied or divided, and the angles added or subtracted.
Unfortunately when the numbers are only added or subtracted they need to be converted
back to cartesian form to perform the operations. This method eliminates the need to
multiply by the complex conjugate.
B
2
2
A + B atan ---
2
2
A
A + jBA +B
B
D
--------------= ------------------------------------------------- = ----------------------- atan --- atan ----
C + jD
A
C
2
2
D
2
2
C +D
C + D atan ----
C
A
-------------1- = A
--- ( 1 2 )
B 2
B
( A 1 ) ( B 2 ) = AB ( 1 + 2 )
For example,
2+j
( 1 + j ) --------------
3 + 4j
=
5 0.4636
2 0.7854 ------------------------------25 0.9273
2 5
= -------------- 0.7854 + 0.4636 0.9273
25
= 0.6325 0.3217
= 0.6 + j0.2
Figure 9.4
The cartesian form of complex numbers seen in the last section are well suited to
operations where complex numbers are added and subtracted. But, when complex numbers are to be multiplied and divided these become tedious and bulky. The polar form for
complex numbers simplifies many calculations. The previous example started in Figure
9.2 is redone using polar notation in Figure 9.5. In this example the input is directly converted to polar form, without the need for calculation. The input frequency is substituted
into the transfer function and it is then converted to polar form. After this the output is
found by multiplying the transfer function by the input. The calculations for magnitudes
involve simple multiplications. The angles are simply added. After this the polar form of
the result is converted directly back to a function of time.
Consider the input function from the previous example in polar form it becomes,
F ( t ) = 10 sin ( 100t + 0.5 )N
F ( ) = 10 0.5
The transfer function can also be put in polar form.
x()
1
1
------------ = --------------------------------------------------= -----------------------------------------------------------------------------------------2
F()
9998000 + j300000
( 2000 1000 ( 100 ) ) + j ( 3000 ( 100 ) )
x()
1 0
-----------= -------------------------------------------------------------------------------------------------------------------------------F()
300000
2
2
( 9998000 ) + ( 300000 ) atan ------------------------ +
9998000
x
(
)
1
0
1
------------ = ------------------------------------------ = ------------------------ ( 0 3.112 ) = 0.9998 10-7 3.112
F()
10002500 3.112
10002500
The output can now be calculated.
x( )
-7
x ( ) = ------------ F ( ) = ( 0.9998 10 3.112 ) ( 10 0.5 )
F()
-7
-6
Re
Note: recall that tan = ------ but the atan function in
Im
calculators and software only returns values between 90 to 90 degrees. To compensate for this the sign of the
real and imaginary components must be considered to
determine where the angle lies. If it lies beyond the -90
to 90 degree range the correct angle can be obtained
by adding or subtracting 180 degrees.
Figure 9.5
imaginary
+/-
+/+
-/-
-/+
real
Consider the circuit analysis example in Figure 9.6. In this example the component
values are converted to their impedances, and the input voltage is converted to phasor
form. (Note: this is a useful point to convert all magnitudes to powers of 10.) After this the
three output impedances are combined to a single impedance. In this case the calculations
were simpler in the cartesian form.
10K
1mH
+
-
100nF
10K
+
Vo
-
6
V i = 5 sin ( 10 t + 0.3 )
The impedances and input voltage can be written in phasor form.
10
5 0.3
+
-
4
10
6
10 j
1 ----------------------7 6
10 10 j
10
+
Vo
-
+
-
Z eq
+
Vo
-
1
11 - -------------------------1------- + ------= -------------------+
3
6
4
1
Z eq
10 10 j -------------------- 10
7 6
10 10 j
13
1
4
4
------= 10 j + 10 j + 10 = ( 10 ) + j ( 0.099 )
Z eq
1
Z eq = ----------------------------------------4
( 10 ) + j ( 0.099 )
Figure 9.6
The analysis continues in Figure 9.7 as the output is found using a voltage divider.
In this case a combination of cartesian and polar forms are used to simplify the calculations. The final result is then converted back from phasor form to a function of time.
( 10 ) + j ( 0.099 )
V o ( ) = ( 5 0.3 ) ---------------------------------------------------------------
4
1
-
10 + ----------------------------------------4
( 10 ) + j ( 0.099 )
1
-
V o ( ) = ( 5 0.3 ) --------------------------------------------------------------------4
4
( 10 ) ( ( 10 ) + j ( 0.099 ) ) + 1
1
5 0.3
5 0.3
V o ( ) = ( 5 0.3 ) -------------------- = ---------------------------------------------------- = --------------------------------------2 + j990
990 1.5687761
990
2
2
2 + 990 atan --------2
5
3
V o ( ) = --------- ( 0.3 1.5687761 ) = 5.05 10 1.269
990
Finally, the output voltage can be written.
6
Figure 9.7
Phasor analysis is applicable to systems that are linear. This means that the principle of superposition applies. Therefore, if an input signal has more than one frequency
component then the system can be analyzed for each component, and then the results simply added. The example considered in Figure 9.2 is extended in Figure 9.8. In this example the input has a static component, as well as frequencies at 0.5 and 20 rad/s. The
transfer function is analyzed for each of these frequencies components. The output components are found by multiplying the inputs by the response at the corresponding frequency. The results are then converted back to functions of time, and added together.
Figure 9.8
9.3 VIBRATIONS
Oscillating displacements and forces in mechanical systems will cause vibrations.
In some cases these become a nuisance, or possibly lead to premature wear and failure in
mechanisms. A common approach to dealing with these problems is to design vibration
isolators. The equations for transmissibility and isolation is shown in Figure 9.9. These
equations can compare the ratio of forces or displacements through an isolator. The calculation is easy to perform with a transfer function or Bode plot.
%I = ( 1 T )100%
Figure 9.9
Transmissibility
Given the transfer function for a vibration isolator below, find a value of K that will
give 50% isolation for a 10Hz vibration.
x out
5D + 10
--------- = ---------------------------------------2
x in
4D + 20KD + 4
Figure 9.10
9.4 SUMMARY
Phasor transforms and phasor representations can be used to find the steady state
response of a system to a given input.
Vibration analysis determines frequency components in mechanical systems.
+
-
1K
1K
1K
Vo
2. A single d.o.f. model with a weight of 1.2 kN and a stiffness of 340 N/m has a steady-state harmonic excitation force applied at 95 rpm (revolutions per minute). What damper value will
give a vibration isolation of 92%?
3. Four helical compression springs are used, one at each corner of a piece of equipment. The
spring rate is 240 N/m for each spring and the vertical static deflection of the equipment is
10mm. Calculate the mass of the equipment and determine the amount of isolation the springs
would afford if the equipment operating frequency is twice the natural frequency of the system.
4. a) For the circuit below find the transfer function and the steady state response for an input of
Vi=5sin(1000t)V.
1K
1uF
1mH
Vi
+
-
+ Vo
-
Va
+
1mH
Vi
1K
+
-
1K
1K
Vo
Va
V a Vo
Va Vi
Va
----------------------------------------------------- = 0
I
=
+
+
+
Va
1K 1K 0.001D
1K
3
1
1
1
Va ------------ + ------------------ + Vi ------------ = Vo ------------------
1K 0.001D
1K
0.001D
IVo
(1)
Vo Va
Vo
- + ----------- = 0
= ----------------0.001D 1K
1
1
1
V o ------------------ + ------------ = V a ------------------
0.001D 1K
0.001D
0.001D + 1K
V o ------------------------------------ = V a
1K
(2)
1K
1K 0.001D
1K
0.001D
1 ----------Vo
1K
------ = ------------------------------------------------------------------------------------------------------------------Vi
+ 1K ----------3 - ----------------1 - ----------------1 -
0.001D
-----------------------------------+
1K 0.001D 0.001D
1K
1K
1K
0.001 ( j10 )
0.001 ( j10 )
3
Vo
10
---------------- = ----------------------------------------------------------------------------------3
3
3
10 + 0j
( j + 1 ) ( 3 10 j10 ) + j ( 10 )
Vo
1
---------------- = --------------------------------------10 + 0j
( j + 1)(3 j) + j
Vo
1 - 4------------ 3j-
---------------- = ------------
10 + 0j
4 + 3j 4 3j
4 3j
V o = 10 -------------- = 2 0.644
25
6
Fy
= K s y K d yD F = MyD
1
--y- = ----------------------------------------2
F
D M + DK d + K s
Ks
Kd
F floor = K s y + K d yD
F floor
------------ = Ks + K d D
y
F floor
F floor y
( Ks + Kd D )
------------- = ------------ --- = ----------------------------------------2
y F
F
D M + DK d + K s
F floor
( K s + K d j )
------------- = ---------------------------------------------2
F
M + jK d + K s
2
Ks + ( K d )
F floor
------------ = -----------------------------------------------------------2
F
2
2
( K s M ) + ( K d )
(contd
For 92% isolation, there is 100-92 = 8% transmission, at 95rpm.
N
K s = 340 ---m
1200N
M = ----------------- = 122kg
N
9.81 -----kg
d
m
s
0.08 = --------------------------------------------------------------------------------------------------------------------2
2
2
N-
rad
340 ----------- 122kg + 9.95 rad
--------- K d
9.95
m
s
s
2
N- 2
rad
340 -----------
+
K
9.95
d
2
2
2
m
s
N-
rad
340 ----------- 122kg + 9.95 rad
--------- K d = --------------------------------------------------------------9.95
m
s
s
0.08
2
N
rad - 2
115600 N
2 99.0025 rad
( 1.377878 10 ) -----2- + 99.0025 ---------K d = ------------------ ------2 + K d ------------------- ---------2
2
0.0064
0.0064
m
s
m
s
8
2 2
10 - ----------------N s 2
1.197253
--------------------------------= Kd
2
15370.138 2
m rad
Ns
K d = 88.3 -----m
3.
a)
M = 0.979kg
b)
I = 67%
4.
a)
b)
V o ( t ) = 5000 + 2500e
1000t
D
--x- = ----------------------------2
F
( D + 200 )
2. For the following transfer function,
a) Draw the Bode plot on the attached semi-log graph paper.
b) Given an input of F=5sin(62.82t), find the output, x, using the Bode plot.
c) Given an input of F=5sin(62.82t), find the output, x, using phasors.
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
Topics:
Bode plots
Objectives:
To be able to describe the response of a system using Bode plots
10.1 INTRODUCTION
When a phasor transform is applied to a transfer function the result can be
expressed as a magnitude and angle that are functions of frequency. The magnitude is the
gain, and the angle is the phase shift. In the previous chapter these values were calculated
for a single frequency and then multiplied by the input values to get an output value. At
different frequencies the transfer function value will change. The transfer function gain
and phase angle can be plotted as a function of frequency to give an overall picture of system response.
Aside: Consider a graphic equalizer commonly found on home stereo equipment. The
spectrum can be adjusted so that high or low tones are emphasized or muted. The
position of the sliders adjusts the envelope that the audio signal is filtered through.
The sliders trace out a Bode gain plot. In theoretical terms the equalizer can be
described with a transfer function. As the slides are moved the transfer function is
changed, and the bode plot shifts. In the example below the slides are positioned to
pass more of the lower frequencies. The high frequencies would not be passed clearly,
and might sound somewhat muffled.
40-120
120-360
360-1K
1K-3.2K
3.2K-9.7K
9.7K-20K
gain
f(Hz)
Figure 10.1
x( )
1
-----------= ---------------------------------------------------------------------2
F( )
j ( 3000 ) + ( 2000 1000 )
x( )
1
-----------= ---------------------------------------------------------------------2
F( )
j ( 3000 ) + ( 2000 1000 )
x( )
( 2000 1000 ) j ( 3000 )-----------= ------------------------------------------------------------------------2 2
2
F( )
( 2000 1000 ) + ( 3000 )
2 2
1000 + 2000
2
for ( 2 )
3
= atan ------------2-
2
for ( > 2 )
Figure 10.2
The results in Figure 10.2 are normally left in variable form so that they may be
analyzed for a range of frequencies. An example of this type of analysis is done in Figure
10.3. A set of frequencies is used for calculations. These need to be converted from Hz to
rad/s before use. For each one of these the gain and phase angle is calculated. The gain
gives a ratio between the input sine wave and output sine wave of the system. The magnitude of the output wave can be calculated by multiplying the input wave magnitude by the
gain. (Note: recall this example was used in the previous chapter) The phase angle can be
added to the input wave to get the phase of the output wave. Gain is normally converted to
dB so that it may cover a larger range of values while still remaining similar numerically.
Also note that the frequencies are changed in multiples of tens, or magnitudes.
f(Hz)
(rad/sec)
0
0.001
0.01
0.1
1
10
100
1000
0
0.006283
0.06283
0.6283
6.283
62.83
628.3
6283
Gain
Gain (dB)
(rad.)
(deg.)
Note: The frequencies chosen should be chosen to cover the points with the greatest
amount of change.
Figure 10.3
In this example gain is defined as x/F. Therefore F is the input to the system, and x
is the resulting output. The gain means that for each unit of F input to the system, there
will be gain*F=x output. If the input and output are sinusoidal, there is a difference in
phase between the input and output wave of (the phase angle). This is shown in Figure
10.4, where an input waveform is supplied with three sinusoidal components. For each of
the frequencies a gain and phase shift are calculated. These are then used to calculate the
resulting output wave, relative to the input wave. The resulting output represents the
steady-state response to the sinusoidal output.
2
x ( t ) = 1000 ( _________ ) sin 0t + ______ ---------
360
2
+ 20 ( _________ ) sin 20t + _______ ---------
360
2
+ 10 ( __________ ) sin 0.5t ________ ---------
360
gain from
calculations
or Bode plot
Figure 10.4
5 6 789 1
5 6 789 1
5 6 789 1
5 6 789
Figure 10.5
Plot the points from Figure 10.3 in the semi-log graph paper in Figure 10.5. The general layout is pictured below.
gain
(dB)
-40dB/dec
phase
(deg)
0.1
Figure 10.6
10
f(Hz)
100
1000
Drill problem: Plot the points from Figure 10.3 on graph paper
steps_per_dec = 6;
decades = 6;
start_freq = 0.1;
// the transfer function
function foo=G(w)
D = %i * w;
foo = (D + 5) / (D^2 + 100*D + 10000);
endfunction
// this section writes the values to a datafile that may be graphed in a spreadsheet
fd = mopen("data.txt", "w");
for step = 0:(steps_per_dec * decades),
f = start_freq * 10 ^ (step / steps_per_dec); // calculate the next frequency
w = f * 2 * %pi);
// convert the frequency to radians
[gain, phase] = polar(G(w));// find the gain and convert it to mag and angle
gaindb = 20 * log10(gain);// convert magnitude to dB
phasedeg = 180 * phase / %pi;// convert to degrees
mfprint(fd, "%f, %f, %f \n", f, gaindb, phasedeg);
end
mclose(fd);
// to graph it directly the following is used
D = poly(0, D);
h = syslin(c, (D + 5) / (D^2 + 100*D + 10000) ) ;
bode(h, 0.1, 1000, Sample Transfer Function);
Figure 10.7
Figure 10.8
Drill problem: Plot the points from Figure 10.2 with a computer
Figure 10.9
D+3
----------------------------------------------------2
D + 10000D + 10000
Drill problem: Draw the Bode plot for gain and phase
Bode plots for transfer functions can be approximated with the following steps.
1. Plot the straight line pieces.
a) The gain at 0rad/sec is calculated and used to find an initial offset. For example this transfer function starts at 10(D+1)/
(D+1000)=10(0+1)/(0+1000)=0.01=-40dB.
b) Put the transfer function in root form to identify corner frequencies. For example (D+1)/(D+1000) will have corner frequencies
at 1 and 1000 rad/sec.
c) Curves that turn up or down are drawn for each corner frequency. At each corner frequency a numerator term causes the
graph to turn up, each term in the dominator causes the graph to
turn down. The slope up or down is generally +/- 20dB/decade
for each term. Also note that squared (second-order) terms would
have a slope of +/-40dB/decade.
2. The effect of each term is added up to give the resulting straight line
approximation.
3. When the smooth curve is drawn, there should commonly be a 3dB difference at the corner frequencies. In second-order systems the damping
coefficient make may the corner flatter or peaked.
Figure 10.10 The method for Bode graph straight line gain approximation
Note: Some of the straight line approximation issues are discussed below.
Why is there 3dB between a first order corner and the smooth plot, and the phase
angle is 45 degrees of the way to +/- 90 degrees.
1
G ( j ) = ----------------- c + j
the initial gain is
1
G ( 0 ) = ----------------- =
c + j0
at the corner frequency
1
G ( j ) = -------------------- c + j c
1----c
c =
1
1
= ---------------------- = ------------- --
c 2 4
c 2 --4
c > j
1- = 20 log ( 1 ) = 20 log ( )
G ( j ) = ----c
c
c
after the corner frequency,
c < j
1- = 20 log ( 1 ) = 20 log ( )
G ( j ) = ----j
each time the frequency increases by a multiple of 10, the log value
becomes 1 larger, thus resulting in a gain change of -20 dB.
Why does a pole make the phase angle move by -90deg after the corner frequency?
1
1 0
1
2
2
2
2
c
c +
c + atan ------
c
before the corner frequency,
c > j
Show that the second order transfer function below would result in a slope of +/- 40
dB/decade.
1
G ( j ) = -------------------------2( c + j )
10 ( D + 1 ) 100D + 100
- = --------------------------------------------G ( D ) = -------------------------------------------------2
( D + 10 ) ( D + 100 )
0.01D + 0.11D + 10
Step 1: Draw lines for each of the terms in the transfer function,
Gain (dB)
D+1
10 ( 1 )
GAIN 0 = ------------------------- = 10 = 20dB
( 10 ) ( 100 )
20dB/dec.
0dB
10
100
freq.(rad/sec)
(could be Hz also)
-20dB/dec.
-20dB/dec.
1 --------------D + 10
1 -----------------D + 100
Step 2: Sum the individual lines, and get the straight line approximation,
Gain (dB)
40dB
20dB
0dB
10
100
freq.(rad/sec)
3dB
40dB
20dB
0dB
10
100
freq.(rad/sec)
The process for constructing phase plots is similar to that of gain plots, as seen in
Figure 10.18. The transfer function is put into root form, and then straight line phase shifts
are drawn for each of the terms. Each term in the numerator will cause a positive shift of
90 degrees, while terms in the denominator cause negative shifts of 90 degrees. The phase
shift occurs over two decades, meaning that for a center frequency of 100, the shift would
start at 10 and end at 1000. If there are any lone D terms on the top or bottom, they will
each shift the initial value by 90 degrees, otherwise the phase should start at 0degrees.
Gain plots for transfer functions can be approximated with the following steps.
1. Plot the straight line segments.
a) Put the transfer function in root form to identify center frequencies. For example (D+1)/(D+1000) will have center frequencies
at 1 and 1000 rad/sec. This should have already been done for
the gain plot.
b) The phase at 0rad/sec is determined by looking for any individual D terms. Effectively they have a root of 0rad/sec. Each of
these in the numerator will shift the starting phase angle up by
90deg. Each in the denominator will shift the start down by 90
deg. For example the transfer function 10(D+1)/(D+1000)
would start at 0 deg while 10D(D+1)/(D+1000) would start at
+90deg.
c) Curves that turn up or down are drawn around each center frequency. Again terms in the numerator cause the curve to go up 90
deg, terms in the denominator cause the curves to go down 90
deg. Curves begin to shift one decade before the center frequency, and finish one decade after.
2. The effect of each term is added up to give the resulting straight line
approximation.
3. The smooth curve is drawn.
Figure 10.17 The method for Bode graph straight line gain approximation
The previous example started in Figure 10.14 is continued in Figure 10.18 to
develop a phase plot using the approximate technique. There are three roots for the transfer function. None of these are zero, so the phase plot starts at zero degrees. The root in the
numerator causes a shift of positive 90 deg, starting one decade before 1rad/sec and ending one decade later. The two roots in the denominator cause a shift downward.
10 ( D + 1 )
G ( D ) = ---------------------------------------------( D + 10 ) ( D + 100 )
Step 1: Draw lines for each of the terms in the transfer function
Phase angle (deg.)
D+1
+90
+45
0.1
10
100
1000
freq.(rad/sec)
-45
1 --------------D + 10
1 -----------------D + 100
-90
10
100
1000
freq.(rad/sec)
-45
-90
Phase angle (deg.)
+90
+45
0.1
10
100
1000
freq.(rad/sec)
-45
-90
Gain
Mn
dB = A
Resonant Peak
=0.1
=0.5
A
------------------------------------------------------2
2
D + ( 2 n )D + ( n )
d =
1 n
= 0.707
-40dB/decade
d
freq
Note: If is less than 1 the roots become complex, and the Bode plots get a
peak. This can be seen mathematically because the roots of the transfer
function become complex.
D+3
----------------------------------------------------2
D + 10000D + 10000
x(t)
displacement
2
2
2
x ( t ) = A sin ---------- t + 0.3A sin ------------- t + 0.01A sin ---------------- t
0.01
0.005
0.0005
t (ms)
frequency 1
freq. 3
freq. 2
Amplitude
A
log(freq.)
(Hz or rads.)
1--------0.01
1 -----------0.005
1 --------------0.0005
=100Hz
=200Hz
=2000Hz
10.4 SUMMARY
Bode plots show gain and phase angle as a function of frequency.
Bode plots can be constructed by calculating point or with straight line approximations.
A signal spectrum shows the relative strengths of components at different frequencies.
AND
5
-----2
D
Bode plot.
3. Use the straightline approximation techniques to draw the Bode plot for the transfer function
below.
F
D + 1000 G = --- = ----------------------------------2
x
D + 5D + 100
4. Given the transfer function below,
V
1000 -----o- = --------------------Vi
D + 1000
a) Find the steady state response of the circuit using phasors (i.e., phasor transforms) if the input is Vi=5sin(100,000t).
b) Draw an approximate Bode plot for the circuit.
5. For the transfer function,
D ( D + 2 )
---------------------------------------------------------2
2
D + 300D + 62500
a) Use the straight line method and the attached log paper to draw an approximate
Bode plot.
b) Verify the Bode plot by calculating values at a few points.
c) Use the Bode plot to find the response to an input of 5sin(624t) + 1sin(6.2t).
6. The applied force F is the input to the system, and the output is the displacement x. Neglect
the effects of gravity.
a) find the transfer function.
K1 = 500 N/m
K2 = 1000 N/m
M = 10kg
f) Find x(t), given F(t) = 10N for t >= 0 seconds considering the effects of gravity.
7. The following differential equation is supplied, with initial conditions.
y + y + 7y = F
y( 0) = 1
y(0) = 0
F ( t ) = 10
t>0
0.0
0.002
0.004
0.006
0.008
0.010
f) Give the expected x response of this first-order system to a step function input
for force F = 1N for t > 0 if the system starts at rest. Hint: Use the canonical
form.
-20dB
1----2
100
--------2
1000
-----------2
-40dB/dec
------52
90deg
0deg
-45deg
-180deg
2.
y----------( D )( D + 10 ) ( D + 5 -)
( D + 10 )
= --------------------------------------= --------------------2
x(D )
(D + 5)
(D + 5)
6dB
0dB
0.8
1.6
0.8
1.6
90
-90
x ( t ) = 20 cos ( 9t + 0.3 )
Aside: the numbers should be obtained from the graphs, but I have calculated them
y-( D + 10 )
9j + 10 5 9j
50 + 81 45j
= --------------------- = ----------------- -------------- = -------------------------------- = 1.236 0.425j
x
(D + 5)
9j + 5 5 9j
25 + 81
y-=
x
0.425
2
2
1.236 + 0.425 atan ---------------- = 1.307 0.3312rad
1.236
(contd
Aside: This can also be done entirely with phasors in polar notation
y-( D + 10 )
9j + 10
13.45 0.733
13.45
= --------------------- = ----------------- = ------------------------------- = ------------- 0.733 1.064 = 1.31 0.331
x
(D + 5)
9j + 5
10.30 1.064
10.30
x = 20 0.3
y ----------------= 1.31 0.331
20 0.3
y = 1.31 ( 20 ) ( 0.331 + 0.3 ) = 26.2 0.031
y ( t ) = 26.2 sin ( 9t 0.031 )
3.
for the numerator, (zero)
for the denominator, (poles)
rad
1000 --------- = 159Hz
s
2
D + 2 n D + n = D + 5D + 100
2
n = 100
rad
n = 10 --------f n = 1.59Hz
s
2 n = 5
5
= -------------- = 0.25 (underdamped)
2 ( 10 )
2
f d = f n 1 = 1.54Hz
for the initial gain
Gain
(dB)
20
f(Hz)
159
0
0.1
1.54
10
100
1000
-40dB/dec
-60
-20dB/dec
-80
phase
(deg)
0.154
0
0.1
-90
-180
15.4,15.9
1
1.54
10
1590 f(Hz)
100 159
1000
4.
5
a)
b)
16Hz
159Hz
1.6KHz
-20dB/dec
0 deg
-90 deg
5.
a)
b)
c)
6.
0.0667 - m
--x- = ------------------------2
F
D + 33.3 N
b) = 1
D = 1j
a)
m
0.0667 3
--x- = ---------------------------= 2.07 10 0rad ---2
N
F
( 1j ) + 33.3
F ( t ) = 10 cos ( t + 1 ) N
F ( ) = ( 10 1rad )N
x
------------------------------- = 2.07 10 3 0rad m
---( 10 1rad )N
N
m
3
x ( ) = 2.07 10 0rad ---- ( 10 1rad )N
N
3
x ( ) = ( 10 )2.07 10 ( 0rad + 1rad )m
x ( ) = 0.0207 1radm
x ( t ) = 0.0207 cos ( t + 1 ) m
c)
F
+ 33.3- --N--- = D
---------------------x
0.0667 m
d)
D + 2 n D + n = D + 33.3
rad
n = 5.77 --------s
= 0
f n = 0.918Hz
(undamped)
2
f d = f n 1 = 0.918Hz
0.0667
G ( 0 ) = ---------------- = 2 = 54 dB
33.3
-54
f(Hz)
0
0.1
-94
10
100
-40dB/dec
-134
-20dB/dec
phase
(deg)
f(Hz)
0
0.1
-90
-180
e)
f)
10
100
7.
a)
y = v
v = v 7y + F
b)
yi + 1
vi + 1
yi
+h
vi
given
y0
v0
0 1 yi + 0
7 1 v i
F
= 1
0
F = 10
using h=0.001s
y1
v1
= 1 + 0.001 0 1 1 + 0
0
7 1 0
10
etc.. until
y 100
v 100
c)
y ( t ) = ( 0.437e
0.5t
1.428m
=
m
0.000 ---s
d)
1
1
1
--y- = -------------------------- = ------------------------------------= ---------------------------------2
2
2
F
D +D+7
( j ) + j + 7
(7 ) + j()
1
--y- = ---------------------------------------F
2 2
2
(7 ) +
0
- = ( 0 angle ( , 7 2 ) ) = angle ( , 7 2 )
= ---------------------------------------------2
angle ( , 7 )
-20dB
-40dB
0.01Hz
1Hz
100Hz
0.01Hz
1Hz
100Hz
0deg
-180deg
8.
a)
1 --x- = -----------F
5+D
b)
1 1
1 0
1
--x- = -----------= --------------- = ------------------------------------------------ = ---------------------- atan ----
5
F
5+D
5 + j
2
2
2
2
5 +
5 + atan ----
5
c)
1
initial gain = 20 log ------------ = 14dB
5+0
5
corner freq. = ------ = 0.8Hz
2
-20dB/dec
gain(dB)
-14dB
phase(deg)
0
45
90
0.08Hz 0.8Hz
d)
8Hz
100
f = --------- = 16Hz
2
From the Bode plot,
10 0
5
2
2
5 + 100
x = ( 10 0 ) ( 0.00999 1.521 ) = 0.0999 1.521
x ( t ) = 0.0999 sin ( 100t 1.521 )
e)
0.186
0.2
Xi
0.099 0.2
0.5
i h
0.999
5t
f)
1 e
x = --- --------5
5
0.0
0.002
9.98e-4
0.004
5.92e-3
0.006
0.015
0.008
0.026
0.010
0.041
( D + 2 )
------------------------------------------------------2
2
D + 50D + 10000
2. Draw Bode plots for the following functions using straight line approximations.
1 -----------D+1
1 --------------2
D +1
1
--------------------2
(D + 1)
1
----------------------------2
D + 2D + 2
5x + 2x = 3F
7 8 9 1
7 8 9
5 6 789 1
5 6 789 1
5 6 789 1
5 6 789
Topics:
Root-locus plots
Objectives:
To be able to predict and control system stability.
11.1 INTRODUCTION
The system can also be checked for general stability when controller parameters
are varied using root-locus plots.
R = A, B
jw
x(t)
x0
-A
-B
sigma
t
Figure 11.1
R = Aj
jw
A
x(t)
sigma
-A
Figure 11.2
R = A Bj
jw
x(t)
B
-A
sigma
-B
Figure 11.3
R = A Bj
jw
x(t)
B
-A
sigma
-B
Figure 11.4
More negative real and complex roots cause a faster decaying oscillation
R = A, A
jw
x(t)
-A,-A
sigma
t
Figure 11.5
R = A, A
jw
x(t)
A,A
sigma
t
Figure 11.6
R = A Bj
jw
x(t)
B
A
-B
Figure 11.7
sigma
Complex roots with positive real parts have growing oscillations and are
unstable
Next, recall that the denominator of a transfer function is the homogeneous equation. By analyzing the function in the denominator of a transfer function the general system response can be found. An example of root-locus analysis for a mass-spring-damper
system is given in Figure 11.8. In this example the transfer function is found and the roots
of the equation are written with the quadratic equation. At this point there are three
unspecified values that can be manipulated to change the roots. The mass and damper values are fixed, and the spring value will be varied. The range of values for the spring coefficient should be determined by practical and design limitations. For example, the spring
coefficient should not be zero or negative.
x ( D )1
-----------= ----------------------------------------2
F(D)
MD + K d D + K s
Ks
Note: We want
the form below
A
-------------------------------------(D + B)(D + C)
Kd
1---x
(
D
)
M
------------- = ------------------------------------F(D)
Ks
2 Kd
D + ------ D + ----M
M
Aside:
2
ax + bx + c = ( x + A ) ( x + B )
Kd
K
K
--------- ------d- 4 -----s
2
M
M
M
-----------------------------------------B, C =
2
b b 4ac
A, B = -------------------------------------2a
Kd
Ks
0
100
100
100
0
100
100
100
0
100
1000
10000
Figure 11.8
The roots of the equation can then be plotted to provide a root locus diagram.
These will show how the values of the roots change as the design parameter is varied. If
any of these roots pass into the right hand plane we will know that the system is unstable.
In addition complex roots will indicate oscillation.
Imaginary
Real
Figure 11.9
1--D
K
1
K
G ( D ) = ---D
H(D ) = 1
First, we must develop a transfer function for the entire control system.
K-
--
D
G( D)
K
G S ( D ) = ------------------------------------- = --------------------------- = -------------1 + G ( D )H ( D )
D+K
K
1 + ---- ( 1 )
D
Next, we use the characteristic equation of the denominator to find the roots as
the value of K varies. These can then be plotted on a complex plane. Note:
the value of gain K is normally found from 0 to +infinity.
D+K = 0
root
0
1
2
3
etc...
-0
-1
-2
-3
j
K
K = 0
Note: This system will always be stable because all of the roots for all values of K are
negative real, and it will always have a damped response. Also, larger values of K,
make the system more stable.
Aside: Scilab can be used to draw root locus plots for systems of the form below, where
there is a simple gain, K, multiplying the openloop gain, G(s).
+
1--D
K
1
1
G ( D ) = ---D
H(D ) = 1
D + 3D + 2
2
D + 3D + 2 + K = 0
3 9 4(2 + K)
1 4K
roots = ------------------------------------------------ = 1.5 -------------------2
2
Next, find values for the roots and plot the values,
K
roots
0
1
2
3
K(D + 5 )
G ( D )H ( D ) = --------------------------------------2
D ( D + 4D + 8 )
11.3 SUMMARY
Root-locus plots show the roots of a transfer function denominator to determine
stability
1
--------------------2
(D + 1)
3.0
-
KdD
2. The block diagram below is for a motor position control system. The system has a proportional
controller with a variable gain K.
d
Vd +
Ve
-
Vs
100 -----------D+2
1--D
Va
2
4. A feedback control system is shown below. The system incorporates a PID controller. The
closed loop transfer function is given.
Y
X +
3 Ki
-----------K p + ----- + K d D
D+9
D
4
D ( 3K d ) + D ( 3K p ) + ( 3K i )
--Y- = ---------------------------------------------------------------------------------------------2
X
D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i )
1
K 5 + ---- + D
D
+
-
0.1
-------------------------------------2
D + 10D + 100
+
-
10
-----D
0.01
6. The feedback loop below is for controlling a DC motor with a PID controller.
Vd
D
+ PD + 1
-----------------------------D
Vs
Va
1--D
3.0
1 -------------------2
(D + 1)
KdD
3.0
------------------------------------2
( D + 1 ) + Kd D
+
-
3.0
--------------------------------------------------2
D + D ( K d + 2 ) + 4.0
2
K d 2 ( K d + 2 ) 4 ( 4.0 )
D = --------------------------------------------------------------------------2
D + D ( K d + 2 ) + 4.0 = 0
K d 2 K d + 4K d 12
D = ----------------------------------------------------------------2
Kd
roots
0
1
2
5
10
100
1000
-1 +/- 1.732j
-1.5 +/- 1.323j
-2.000, -2.000
-0.628, -6.372
-0.343, -11.657
-0.039, -102.0
-0.004, -1000
0 > K d 2 K d + 4K d 12
2
2 + K d > K d + 4K d 12
16 > 0
0 > Kd 2
Kd > 2
The roots becomes complex when
2
0 > K d + 4K d 12
4 16 4 ( 12 )
K d = ----------------------------------------------2
K d = 6, 2
2.
a)
200K
--------------------------------------2
D + 2D + 200K
2 4 4 ( 200K )
roots = ------------------------------------------------ = 1 1 200K
2
Im
K
roots
b)
0
0.001
0.005
0.1
1
5
10
c)
0,-2
-0.1,-1.9
-1,-1
etc.
K=0.005
Re
-2
D + 2D + 200K = D + 2 n D + n
-1
n = 1
K = 0.005
3.
a)
D + D + 20K = 0
1 1 4 ( 20K )
D = --------------------------------------------2
b)
roots
0
1/80
1
10
1000
0.000, -1.000
-0.500, -0.500
-0.5 +/- 4.444j
-0.5 +/- 14.13j
-0.5 +/- 141.4j
1
K > -----80
n = 20K
K>0
c)
o
20 ( 10 )
----- = ---------------------------------------2
d
D + D + 20 ( 10 )
o + d + d 200 = 200d
Homogeneous:
2
A + A + 200 = 0
1 1 4 ( 200 )
A = -------------------------------------------2
o ( t ) = C1 e
0.5t
A = 0.5 14.1j
sin ( 14.1t + C 2 )
Particular:
= A
0 + 0 + A200 = 200 ( 1rad )
A = 1rad
o ( t ) = 1rad
Initial Conditions (assume at rest):
o ( t ) = C1 e
0.5t
0.5t
0.5t
0.5t
(1)
cos ( 14.1t + C 2 )
C 2 = 1.54
d)
o
20 ( 0.01 )
----- = -------------------------------------------2
d
D + D + 20 ( 0.01 )
o + d + d 0.2 = 0.2d
Homogeneous:
2
A + A + 0.2 = 0
1 1 4 ( 0.2 )
A = -----------------------------------------2
o ( t ) = C1 e
0.724t
A = 0.7236068, 0.2763932
+ C2 e
0.276t
Particular:
= A
0 + 0 + A0.2 = 0.2 ( 1rad )
A = 1rad
o ( t ) = 1rad
Initial Conditions (assume at rest):
o ( t ) = C1 e
0.724t
o ( 0 ) = C1 e
0.724t
+ C2e
0.276t
+ C2 e
0.276t
+ 1rad
+ 1rad = 0
(1)
C 1 + C 2 = 1rad
' o ( t ) = 0.724 ( C 1 e
0.724t
) 0.276 ( C 2 e
0.276t
C 1 = 0.381C 2
0.381C 2 + C 2 = 1rad
C 2 = 1.616rad
0.724t
+ ( 1.616 )e
0.276t
+ 1rad
4.
(ans.
X +
-
Kp D + K i + Kd D
------------------------------------------D
3 -----------D+9
4
X +
-
3K p D + 3K i + 3K d D
----------------------------------------------------D(D + 9)
4
3K p D + 3K i + 3K d D
------------------------------------------------------------------------------------------2
D ( D + 9 ) + 12K p D + 12K i + 12K d D
2
3K p D + 3K i + 3K d D
---------------------------------------------------------------------------------------------2
D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i )
b)
roots
-100
-10
-1
-0.1
0
1
10
100
-0.092, 0.109
-0.241, 0.418
-0.46, 2.369
-0.57, 105.6
-0.588, -20.41
-0.808 +/- 0.52j
-0.087 +/- 0.303j
-0.0087 +/- 0.1j
Stable for,
( 9 + 12K p ) 48K i
K d > ----------------------------------------------576K d K i
K d > 0.682
c)
Kp = 1
Ki = 1
Kd = 1
2
3K p D + 3K i + 3K d D
--Y- = ---------------------------------------------------------------------------------------------2
X
D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i )
2
2
3D + 3D + 3 3- -------------------------------------------------D +D+1
--Y- = ---------------------------------------------
=
2
13 2
X
D 13 + D21 + 12
D + D1.615 + 0.923
3
final gain = 20 log ------ = 12.7
13
3
initial gain = 20 log ------ = 12.0
12
for the numerator,
n =
1
= --------- = 0.5
2 n
1 = 1
2
d = n 1 =
1 0.5 = 0.866
0.923 = 0.961
2
1.615
= ------------- = 0.840
2 n
2
-12dB
3K p D + 3K i + 3K d D
--Y- = ---------------------------------------------------------------------------------------------2
X
D ( 12K d + 1 ) + D ( 9 + 12K p ) + ( 12K i )
12K i
--------------------- = 2
12K d + 1
12K i = 48K d + 4
9 + 12K p
2 n = ---------------------- = 20.5 ( 2 )
12K d + 1
24K d = 7 + 12K p
n =
At this point there are two equations and two unknowns, one value
must be selected to continue, therefore,
K p = 10
24K d = 7 + 12K p = 7 + 12 ( 10 ) = 127
K d = 5.292
K i = 21.5
D ( 12 ( 5.292 ) + 1 ) + D ( 9 + 12 ( 10 ) ) + ( 12 ( 21.5 ) ) = 0
2
3 ( 0 )D + 3 ( 0 ) + 3 ( 1 )D
--Y- = --------------------------------------------------------------------------------------------------2
X
D ( 12 ( 1 ) + 1 ) + D ( 9 + 12 ( 0 ) ) + ( 12 ( 0 ) )
2
3D
--Y- = -------------------------2
X
13D + 9D
2
Y ( 13D + 9D ) = X ( 3D )
Y 13 + Y 9 = X 3
X = t
X = 1
X = 0
9
Y + Y ------ = 0
13
It is a first order system,
Y ( t ) = C1 e
9
------ t
13
+ C2
Y( 0) = 0
Y' ( 0 ) = 0
0 = C11 + C2
starts at rest/undeflected
C1 = C2
9
------
t
913
----Y' ( t ) = C e
13 1
9
0 = ------ C 1
13 1
C1 = 0
C2 = 0
no response
5.
X
1
K 5 + ---- + D
0.1
-------------------------------------2
D + 10D + 100
10
-----D
0.01
5D + 1 + D
K ------------------------------
+
-
0.1
---------------------------------------------------------------10
2
D + 10D + 100 + 0.1 ------
D
0.01
X
0.01
X
6.
2
a)
b)
c)
Kd = 1 Ns/m
Kd1 = 1 Ns/m
Ks1
M = 1 kg
M = 1kg
y
F
F
Ks
b) If the input force is a step function of magnitude 1N, calculate the time response
for y by solving a differential equation for a Ks value of 10N/m.
c) Draw the poles for the transfer function on a real-complex plane.
d) Draw a Bode plot for Ks = 1N/m.
2. Draw a root locus diagram for the feedback system below given the variable parameter P.
Vd
+
-
Vs
1--D
Va
2
3. For the transfer functions below, draw the root locus plots assuming there is unity feedback,
i.e., H(D) = 1. Draw an approximate time response for each for a step input.
G(s) =
1 -----------D+1
1 --------------2
D +1
1 -------------------2
(D + 1)
1
-----------------------------2
D + 2D + 2
+
K
1
--------------------2
(D + 1)
1 -------------------2
(D + 1)
2
-----2
D
+
K
2D
+
K
2
-----2
D
(D + 1)
Topics:
Objectives:
12.1 INTRODUCTION
how they are different from linear
- no transfer functions
- new elements needed in block diagrams
---
Q ( ) = 8.5 e 1
In this case the relationship between pressure drop and flow are non-linear. We
need to develop an equation that approximates the local operating point.
q
p R ( q q )
( p )
R --------------q
q
q
R = 2 -----2K
q
p
( p )
Figure 12.1
Kd
Ks
F
x + x ----------------------------- + x ----------------------------- = ---------------------------- M 0.0001t
M 0.0001t
M 0.0001t
12.3.2 Switching
- system components turned on/off
- cables in tension/compression
- show an example where input conditions change
y + 4y = f ( t )
f(t < 0) = 0
f ( 0 t < 5s ) = 2
f ( t 5s ) = 0
- give PWM (Pulse Width Modulation) example with ripple showing equivalent
voltage. PWM is used to generate analog voltage equivalents. Show for a system with first
order response with tau = 0.1s for a frequency of 1KHz, 10Hz and 1Hz. Point out the ripple and effective voltage.
- important to consider when doing system analysis
12.3.3 Deadband
- Friction in all components
- costs money to reduce friction, so it is better to compensate in software
- small actuation signals not large enough to overcome friction
- This effect is normally known as stiction, a combination of the words static and
friction.
- Friction is common in less expensive motors, and when a motor is driving a
mechanical system.
- In systems there are two type of friction that must be considered.
- The static friction, stiction, will prevent initial motion. If the systems breaks
free and starts turning, the kinetic friction will provide a roughly constant friction resistance.
- relationship in figure below.
- the region where the applied voltage has no effect is called the deadband.
The kinetic
friction results
in a different curve
while slowing down
Vapplied
deadband
Figure 12.2
motor starts
to turn as
friction is
overcome
Vadjusted
Vstick
Vwanted
-Vstick
Figure 12.3
Cadjusted
C max = 255
Cstick
Cwanted
C min = 255
C max = 255
-Cstick
if(Cwanted > 0)
C min = 255
C wanted
C adjusted = C stick + ------------------ ( C max C stick )
C max
if(Cwanted < 0)
C wanted
C adjusted = C stick + ------------------ ( C min C stick )
C min
if(Cwanted = 0)
C adjusted = 0
Figure 12.4
- c-code below
#define
#define
#define
#define
c_stick_pos
c_stick_neg
c_max
c_min
100
-110 /* make the value positive */
255
-255 /* make the value positive */
Figure 12.5
- correct by tracking the previous motion direction and taking extra steps when
reversing direction
12.4 SUMMARY
analog IO - 13.1
Objectives:
To understand the basics of conversion to and from analog values.
13.1 INTRODUCTION
An analog value is continuous, not discrete, as shown in Figure 13.1. In the previous chapters, techniques were discussed for designing continuos control systems. In this
chapter we will examine analog inputs and outputs so that we may design continuous control systems that use computers.
Voltage
logical
continuous
t
Figure 13.1
Typical analog inputs and outputs for computers are listed below. Actuators and
sensors that can be used with analog inputs and outputs will be discussed in later chapters.
Inputs:
oven temperature
fluid pressure
fluid flow rate
Outputs:
analog IO - 13.2
physical
phenomenon
Figure 13.2
transducer
(ie., sensor)
signal
conditioning
analog
input
integer
Analog inputs
Analog to digital and digital to analog conversion uses integers within the computer. Integers limit the resolution of the numbers to a discrete, or quantized range. The
effect of using integers is shown in Figure 13.3 where the desired or actual analog value is
continuous, but the possible integer values are quantified with a staircase set of values.
Consider when a continuous analog voltage is being read, it must be quantized into an
available integer value. Likewise, a desired analog output value is limited to available
quantized values. In general the difference between the analog and quantized integer value
is an error based upon the resolution of the analog I/O.
analog
continuous
quantized
integer
analog IO - 13.3
Figure 13.3
Quantization error
voltage
time
T = (Sampling Frequency)-1
Figure 13.4
Sampling time
A more realistic drawing of sampled data is shown in Figure 13.5. This data is
noisier, and even between the start and end of the data sample there is a significant change
in the voltage value. The data value sampled will be somewhere between the voltage at the
start and end of the sample. The maximum (Vmax) and minimum (Vmin) voltages are a
function of the control hardware. These are often specified when purchasing hardware, but
reasonable ranges are;
analog IO - 13.4
0V to 5V
0V to 10V
-5V to 5V
-10V to 10V
The number of bits of the A/D converter is the number of bits in the result word. If
the A/D converter is 8 bit then the result can read up to 256 different voltage levels. Most
A/D converters have 12 bits, 16 bit converters are used for precision measurements.
V(t)
V max
V ( t2 )
V ( t1 )
V min
t
t1 t2
where,
V ( t ) = the actual voltage over time
Figure 13.5
The parameters defined in Figure 13.5 can be used to calculate values for A/D converters. These equations are summarized in Figure 13.6. Equation 1 relates the number of
analog IO - 13.5
bits of an A/D converter to the resolution. Equation 2 gives the error that can be expected
with an A/D converter given the range between the minimum and maximum voltages, and
the resolution (this is commonly called the quantization error). Equation 3 relates the voltage range and resolution to the voltage input to estimate the integer that the A/D converter
will record. Finally, equation 4 allows a conversion between the integer value from the A/
D converter, and a voltage in the computer.
R = 2
(1)
V max Vmin
V ERROR = -----------------------------
2R
(2)
V in V min
- (R 1)
V I = INT --------------------------- V max V min
(3)
VI
V C = ------------ ( V max V min ) + Vmin
R1
(4)
where,
R = resolution of A/D converter
V I = the integer value representing the input voltage
V C = the voltage calculated from the integer value
V ERROR = the maximum quantization error
Figure 13.6
Consider a simple example, a 10 bit A/D converter can read voltages between 10V and 10V. This gives a resolution of 1024, where 0 is -10V and 1023 is +10V. Because
there are only 1024 steps there is a maximum error of 9.8mV. If a voltage of 4.564V is
input into the PLC, the A/D converter converts the voltage to an integer value of 746.
When we convert this back to a voltage the result is 4.570V. The resulting quantization
error is 4.570V-4.564V=+0.006V. This error can be reduced by selecting an A/D converter with more bits. Each bit halves the quantization error.
analog IO - 13.6
Given,
N = 10
V max = 10V
V min = 10V
V in = 4.564V
Calculate,
R = 2
= 1024
V max V min
V ERROR = ----------------------------- = 0.0098V
2R
V in V min
- R = 746
V I = INT --------------------------- V max V min
VI
V C = ----- ( V max Vmin ) + V min = 4.570V
R
Figure 13.7
If the voltage being sampled is changing too fast we may get false readings, as
shown in Figure 13.8. In the upper graph the waveform completes seven cycles, and 9
samples are taken. The bottom graph plots out the values read. The sampling frequency
was too low, so the signal read appears to be different that it actually is, this is called aliasing.
analog IO - 13.7
Figure 13.8
Figure 13.9
The Nyquist criterion specifies that sampling frequencies should be at least twice
the frequency of the signal being measured, otherwise aliasing will occur. The example in
analog IO - 13.8
Figure 13.8 violated this principle, so the signal was aliased. If this happens in real applications the process will appear to operate erratically. In practice the sample frequency
should be 4 or more times faster than the system frequency.
f AD > 2f signal
where,
f AD = sampling frequency
f signal = maximum frequency of the input
There are other practical details that should be considered when designing applications with analog inputs;
Noise - Since the sampling window for a signal is short, noise will have added
effect on the signal read. For example, a momentary voltage spike might result
in a higher than normal reading. Shielded data cables are commonly used to
reduce the noise levels.
Delay - When the sample is requested, a short period of time passes before the
final sample value is obtained.
Multiplexing - Most analog input cards allow multiple inputs. These may share
the A/D converter using a technique called multiplexing. If there are 4 channels
using an A/D converter with a maximum sampling rate of 100Hz, the maximum
sampling rate per channel is 25Hz.
Signal Conditioners - Signal conditioners are used to amplify, or filter signals
coming from transducers, before they are read by the A/D converter.
Resistance - A/D converters normally have high input impedance (resistance), so
they affect circuits they are measuring.
Single Ended Inputs - Voltage inputs to a PLC can use a single common for multiple inputs, these types of inputs are called single ended inputs. These tend to
be more prone to noise.
Double Ended Inputs - Each double ended input has its own common. This
reduces problems with electrical noise, but also tends to reduce the number of
inputs by half.
Sampling Rates - The maximum number of samples that can be read each second. If reading multiple channels with a multiplexer, this may be reduced.
Quantization Error - Analog IO is limited by the binary resolution of the converter. This means that the output is at discrete levels, instead of continuous values.
Triggers - often external digital signals are used to signal the start of data collection.
Range - the typical voltages that the card can read. Typical voltage ranges are 10V to 10V, 0V to 10V, 0V to 5V, 1V to 5V, -5V to 5V, 4mA to 20mA.
DMA - a method to write large blocks of memory directly to computer memory.
This is normally used for high speed data captures.
analog IO - 13.9
Filters - some A/D input cards will provide built in functions to filter the incoming data to remove high frequency noise components.
Input impedance - most analog inputs have very high input resistances, in the
range of Mohms.
ASIDE: This device is an 8 bit A/D converter. The main concept behind this is the successive approximation logic. Once the reset is toggled the converter will start by setting
the most significant bit of the 8 bit number. This will be converted to a voltage Ve that is
a function of the +/-Vref values. The value of Ve is compared to Vin and a simple logic
check determines which is larger. If the value of Ve is larger the bit is turned off. The
logic then repeats similar steps from the most to least significant bits. Once the last bit
has been set on/off and checked the conversion will be complete, and a done bit can be
set to indicate a valid conversion value.
Vin above (+ve) or below (-ve) Ve
Vin
+
-
+Vref
clock
successive
approximation
logic
D to A
converter
Ve
reset
-Vref
done
data out
Quite often an A/D converter will multiplex between various inputs. As it switches the voltage will be sampled by a sample and hold circuit. This will then be converted to a digital value. The sample and hold circuits can be used before the multiplexer to collect
data values at the same instant in time.
analog IO - 13.10
R = 2
V max Vmin
V ERROR = -----------------------------
2R
(5)
(6)
V desired V min
VI = INT ----------------------------------R
Vmax V min
(7)
VI
V output = ----- ( V max V min ) + V min
R
(8)
where,
R = resolution of A/D converter
V ERROR = the maximum quantization error
V I = the integer value representing the desired voltage
V output = the voltage output using the integer value
V desired = the desired output voltage
analog IO - 13.11
Given,
N = 8
V max = 10V
V min = 0V
V desired = 6.234V
Calculate,
R = 2
= 256
V max V min
V ERROR = ----------------------------- = 0.020V
2R
V in V min
- R = 160
V I = INT --------------------------- V max V min
VI
V C = ----- ( V max Vmin ) + V min = 6.250V
R
The current output from a D/A converter is normally limited to a small value, typically less than 20mA. This is enough for instrumentation, but for high current loads, such
as motors, a current amplifier is needed. This type of interface will be discussed later. If
the current limit is exceeded for 5V output, the voltage will decrease (so dont exceed the
rated voltage). If the current limit is exceeded for long periods of time the D/A output may
be damaged.
analog IO - 13.12
ASIDE:
5K
MSB bit 3
bit 2
10K
20K
V +
+
0
Computer
bit 1
V ss
40K
Vo
-
LSB bit 0
80K
= 0 = V
Next, sum the currents into the inverting input as a function of the output voltage and the
input voltages from the computer,
Vb3
V b2
Vb1
V b0
Vo
-------------- + -------------- + -------------- + -------------- = ----------10K 20K 40K 80K
5K
V o = 0.5V b3 + 0.25V b2 + 0.125V b1 + 0.0625Vb 0
Consider an example where the binary output is 1110, with 5V for on,
V o = 0.5 ( 5V ) + 0.25 ( 5V ) + 0.125 ( 5V ) + 0.625 ( 0V ) = 4.375V
13.4.1 Shielding
When a changing magnetic field cuts across a conductor, it will induce a current
flow. The resistance in the circuits will convert this to a voltage. These unwanted voltages
analog IO - 13.13
result in erroneous readings from sensors, and signal to outputs. Shielding will reduce the
effects of the interference. When shielding and grounding are done properly, the effects of
electrical noise will be negligible. Shielding is normally used for; all logical signals in
noisy environments, high speed counters or high speed circuitry, and all analog signals.
There are two major approaches to reducing noise; shielding and twisted pairs.
Shielding involves encasing conductors and electrical equipment with metal. As a result
electrical equipment is normally housed in metal cases. Wires are normally put in cables
with a metal sheath surrounding both wires. The metal sheath may be a thin film, or a
woven metal mesh. Shielded wires are connected at one end to "drain" the unwanted signals into the cases of the instruments. Figure 13.13 shows a thermocouple connected with
a thermocouple. The cross section of the wire contains two insulated conductors. Both of
the wires are covered with a metal foil, and final covering of insulation finishes the cable.
The wires are connected to the thermocouple as expected, but the shield is only connected
on the amplifier end to the case. The case is then connected to the shielding ground, shown
here as three diagonal lines.
Two conductor
shielded cable
cross section
Insulated wires
Metal sheath
Insulating cover
analog IO - 13.14
13.4.2 Grounding
- ground voltages are based upon the natural voltage level in the physical ground
(the earth under your feet). This will vary over a distance. Most buildings and electrical
systems use a ground reference for the building. Between different points on the same
building ground voltage levels may vary as much as a few hundred millivolts. This can
lead to significant problems with voltage readings and system safety.
- A signal can be floating, or connected to a ground
- if floating a system normally has a self contained power source, or self reference
such as a battery, strain gauge or thermocouple. These are usually read with double ended
outputs. The potential for floating voltage levels can be minimized by connecting larger
resistors (up to 100K) from the input to ground.
- a grounded system uses a single common (ground) for all signals. These are normally connected to a single ended inputs.
- the analog common can also be connected to the ground with a large resistor to
drain off induced voltages.
analog IO - 13.15
- cable shields or grounds are normally only connected at one side to prevent
ground loops.
13.6 SUMMARY
A/D conversion will convert a continuous value to an integer value.
D/A conversion is easier and faster and will convert a digital value to an analog
value.
Resolution limits the accuracy of A/D and D/A converters.
Sampling too slowly will alias the real signal.
Analog inputs are sensitive to noise.
Analog shielding should be used to improve the quality of electrical signals.
analog IO - 13.16
2.
10V ( 10V )
R = ---------------------------------- = 200
0.1V
7 bits, R = 128
8 bits, R = 256
The minimum number of bits is 8.
3.
N = 12
R = 4096
V min = 10V
V max = 10V
V in = 2.735V
V in V min
- ( R 1 ) = 2607
V I = INT --------------------------- V max V min
VI
V C = ------------ ( V max V min ) + Vmin = 2.733V
R1
error = V c V in = ( 0.002 )V
14.1 INTRODUCTION
Continuous sensors convert physical phenomena to measurable signals, typically
voltages or currents. Consider a simple temperature measuring device, there will be an
increase in output voltage proportional to a temperature rise. A computer could measure
the voltage, and convert it to a temperature. The basic physical phenomena typically measured with sensors include;
- angular or linear position
- acceleration
- temperature
- pressure or flow rates
- stress, strain or force
- light intensity
- sound
Most of these sensors are based on subtle electrical properties of materials and
devices. As a result the signals often require signal conditioners. These are often amplifiers that boost currents and voltages to larger voltages.
Sensors are also called transducers. This is because they convert an input phenomena to an output in a different form. This transformation relies upon a manufactured
device with limitations and imperfection. As a result sensor limitations are often charac-
terized with;
Accuracy - This is the maximum difference between the indicated and actual reading. For example, if a sensor reads a force of 100N with a 1% accuracy, then
the force could be anywhere from 99N to 101N.
Resolution - Used for systems that step through readings. This is the smallest
increment that the sensor can detect, this may also be incorporated into the
accuracy value. For example if a sensor measures up to 10 inches of linear displacements, and it outputs a number between 0 and 100, then the resolution of
the device is 0.1 inches.
Repeatability - When a single sensor condition is made and repeated, there will be
a small variation for that particular reading. If we take a statistical range for
repeated readings (e.g., 3 standard deviations) this will be the repeatability.
For example, if a flow rate sensor has a repeatability of 0.5cfm, readings for an
actual flow of 100cfm should rarely be outside 99.5cfm to 100.5cfm.
Linearity - In a linear sensor the input phenomenon has a linear relationship with
the output signal. In most sensors this is a desirable feature. When the relationship is not linear, the conversion from the sensor output (e.g., voltage) to a calculated quantity (e.g., force) becomes more complex.
Precision - This considers accuracy, resolution and repeatability or one device relative to another.
Range - Natural limits for the sensor. For example, a sensor for reading angular
rotation may only rotate 200 degrees.
Dynamic Response - The frequency range for regular operation of the sensor. Typically sensors will have an upper operation frequency, occasionally there will be
lower frequency limits. For example, our ears hear best between 10Hz and
16KHz.
Environmental - Sensors all have some limitations over factors such as temperature, humidity, dirt/oil, corrosives and pressures. For example many sensors
will work in relative humidities (RH) from 10% to 80%.
Calibration - When manufactured or installed, many sensors will need some calibration to determine or set the relationship between the input phenomena, and
output. For example, a temperature reading sensor may need to be zeroed or
adjusted so that the measured temperature matches the actual temperature. This
may require special equipment, and need to be performed frequently.
Cost - Generally more precision costs more. Some sensors are very inexpensive,
but the signal conditioning equipment costs are significant.
V1
resistive
film
V1
Vw
Vw
V2
wiper
V2
physical
schematic
Figure 14.1
A Potentiometer
The potentiometer in Figure 14.2 is being used as a voltage divider. As the wiper
rotates the output voltage will be proportional to the angle of rotation.
V1
max
V out
w
V out = ( V 2 V 1 ) ----------- + V 1
max
V2
Figure 14.2
Potentiometers are popular because they are inexpensive, and dont require special
signal conditioners. But, they have limited accuracy, normally in the range of 1% and they
are subject to mechanical wear.
Potentiometers measure absolute position, and they are calibrated by rotating them
in their mounting brackets, and then tightening them in place. The range of rotation is normally limited to less than 360 degrees or multiples of 360 degrees. Some potentiometers
can rotate without limits, and the wiper will jump from one end of the resistor to the other.
Faults in potentiometers can be detected by designing the potentiometer to never
reach the ends of the range of motion. If an output voltage from the potentiometer ever
reaches either end of the range, then a problem has occurred, and the machine can be shut
down. Two examples of problems that might cause this are wires that fall off, or the potentiometer rotates in its mounting.
14.2.2 Encoders
Encoders use rotating disks with optical windows, as shown in Figure 14.3. The
encoder contains an optical disk with fine windows etched into it. Light from emitters
passes through the openings in the disk to detectors. As the encoder shaft is rotated, the
light beams are broken. The encoder shown here is a quadrature encode, and it will be discussed later.
light
emitters
light
detectors
Shaft rotates
Figure 14.3
An Encoder Disk
There are two fundamental types of encoders; absolute and incremental. An absolute encoder will measure the position of the shaft for a single rotation. The same shaft
angle will always produce the same reading. The output is normally a binary or grey code
number. An incremental (or relative) encoder will output two pulses that can be used to
determine displacement. Logic circuits or software is used to determine the direction of
rotation, and count pulses to determine the displacement. The velocity can be determined
by measuring the time between pulses.
Encoder disks are shown in Figure 14.4. The absolute encoder has two rings, the
outer ring is the most significant digit of the encoder, the inner ring is the least significant
digit. The relative encoder has two rings, with one ring rotated a few degrees ahead of the
other, but otherwise the same. Both rings detect position to a quarter of the disk. To add
accuracy to the absolute encoder more rings must be added to the disk, and more emitters
and detectors. To add accuracy to the relative encoder we only need to add more windows
to the existing two rings. Typical encoders will have from 2 to thousands of windows per
ring.
relative encoder
(quadrature)
Figure 14.4
absolute encoder
Encoder Disks
When using absolute encoders, the position during a single rotation is measured
directly. If the encoder rotates multiple times then the total number of rotations must be
counted separately.
When using a relative encoder, the distance of rotation is determined by counting
the pulses from one of the rings. If the encoder only rotates in one direction then a simple
count of pulses from one ring will determine the total distance. If the encoder can rotate
both directions a second ring must be used to determine when to subtract pulses. The
quadrature scheme, using two rings, is shown in Figure 14.5. The signals are set up so that
one is out of phase with the other. Notice that for different directions of rotation, input B
either leads or lags A.
Quad input A
clockwise rotation
Quad Input B
Quad input A
counterclockwise rotation
Quad Input B
Note: To determine direction we can do a simple check. If both are off or on, the first to
change state determines direction. Consider a point in the graphs above where both
A and B are off. If A is the first input to turn on the encoder is rotating clockwise. If
B is the first to turn on the rotation is counterclockwise.
Aside: A circuit (or program) can be built for this circuit using an up/down counter. If
the positive edge of input A is used to trigger the clock, and input B is used to drive
the up/down count, the counter will keep track of the encoder position.
Figure 14.5
Quadrature Encoders
Interfaces for encoders are commonly available for PLCs and as purchased units.
Newer PLCs will also allow two normal inputs to be used to decode encoder inputs.
Normally absolute and relative encoders require a calibration phase when a controller is turned on. This normally involves moving an axis until it reaches a logical sensor
that marks the end of the range. The end of range is then used as the zero position.
Machines using encoders, and other relative sensors, are noticeable in that they normally
move to some extreme position before use.
14.2.2.1 - Tachometers
Tachometers measure the velocity of a rotating shaft. A common technique is to
mount a magnet to a rotating shaft. When the magnetic moves past a stationary pick-up
coil, current is induced. For each rotation of the shaft there is a pulse in the coil, as shown
in Figure 14.6. When the time between the pulses is measured the period for one rotation
can be found, and the frequency calculated. This technique often requires some signal
conditioning circuitry.
pickup
coil
rotating
shaft
Vout
Vout
t
magnet
1/f
Figure 14.6
A Magnetic Tachometer
V2
a
Vout = V 1 + ( V 2 V1 ) ---
L
V out
L
a
V1
Figure 14.7
Linear Potentiometer
A rod drives
the sliding core
x
V = Kx
where,
AC input
V = output voltage
K = constant for device
x = core displacement
signal out
Figure 14.8
An LVDT
Aside: The circuit below can be used to produce a voltage that is proportional to position.
The two diodes convert the AC wave to a half wave DC wave. The capacitor and resistor values can be selected to act as a low pass filter. The final capacitor should be large
enough to smooth out the voltage ripple on the output.
Vac out
Vac in
Vdc out
LVDT
Figure 14.9
These devices are more accurate than linear potentiometers, and have less friction.
Typical applications for these devices include measuring dimensions on parts for quality
control. They are often used for pressure measurements with Bourdon tubes and bellows/
diaphragms. A major disadvantage of these sensors is the high cost, often in the thousands.
Moving
Stationary
Note: you can recreate this effect with the strips below. Photocopy the pattern twice,
overlay the sheets and hold them up to the light. You will notice that shifting one sheet
will cause the stripes to move up or down.
on
off
on
off
14.2.3.4 - Accelerometers
Accelerometers measure acceleration using a mass suspended on a force sensor, as
shown in Figure 14.12. When the sensor accelerates, the inertial resistance of the mass
will cause the force sensor to deflect. By measuring the deflection the acceleration can be
determined. In this case the mass is cantilevered on the force sensor. A base and housing
enclose the sensor. A small mounting stud (a threaded shaft) is used to mount the accelerometer.
Base
Mounting
Stud
Force
Sensor
Mass
Housing
between 10Hz to 10KHz. Temperature variations will affect the accuracy of the sensors.
Standard accelerometers can be linear up to 100,000 m/s**2: high shock designs can be
used up to 1,000,000 m/s**2. There is often a trade-off between a wide frequency range
and device sensitivity (note: higher sensitivity requires a larger mass). Figure 14.13 shows
the sensitivity of two accelerometers with different resonant frequencies. A smaller resonant frequency limits the maximum frequency for the reading. The smaller frequency
results in a smaller sensitivity. The units for sensitivity is charge per m/s**2.
sensitivity
22 KHz
180KHz
4.5 pC/(m/s**2)
.004
isolated
stud
accelerometer
isolated
wafer
surface
Sensor
preamp
signal processor/
recorder
Source of vibrations,
or site for vibration
measurement
control system
L
V
+
I
V
L
L
R = --- = --- = -----I
A
wt
where,
R = resistance of wire
V, I = voltage and current
L = length of wire
w, t = width and thickness
A = cross sectional area of conductor
= resistivity of material
After the wire in Figure 14.16 has been deformed it will take on the new dimensions and resistance shown in Figure 14.17. If a force is applied as shown, the wire will
become longer, as predicted by Youngs modulus. But, the cross sectional area will
decrease, as predicted by Poisons ratio. The new length and cross sectional area can then
be used to find a new resistance.
F
F
= --- = ------ = E
A
wt
F
= ---------Ewt
L'
L(1 + )
R' = -------- = ----------------------------------------------
w't'
w ( 1 )t ( 1 )
(1 + )
R = R' R = R ---------------------------------------- 1
( 1 ) ( 1 )
where,
Aside: Changes in strain gauge resistance are typically small (large values would require
strains that would cause the gauges to plastically deform). As a result, Wheatstone
bridges are used to amplify the small change. In this circuit the variable resistor R2
would be tuned until Vo = 0V. Then the resistance of the strain gage can be calculated
using the given equation.
R2 R1
R strain = ------------- when Vo = 0V
R3
V+
R2
R4
R1
Rstrain
Vo
R3
R5
stress
direction
uniaxial
rosette
14.2.4.2 - Piezoelectric
When a crystal undergoes strain it displaces a small amount of charge. In other
words, when the distance between atoms in the crystal lattice changes some electrons are
forced out or drawn in. This also changes the capacitance of the crystal. This is known as
the Piezoelectric effect. Figure 14.21 shows the relationships for a crystal undergoing a
linear deformation. The charge generated is a function of the force applied, the strain in
the material, and a constant specific to the material. The change in capacitance is proportional to the change in the thickness.
+
q
a
F
ab
d---C = --------i = g F
c
dt
where,
C = capacitance change
a, b, c = geometry of material
= dielectric constant (quartz typ. 4.06*10**-11 F/m)
i = current generated
F = force applied
g = constant for material (quartz typ. 50*10**-3 Vm/N)
E = Youngs modulus (quartz typ. 8.6*10**10 N/m**2)
PV = nRT
where,
P = the gas pressure
V = the volume of the gas
n = the number of moles of the gas
R = the ideal gas constant =
T = the gas temperature
When flowing, the flow may be smooth, or laminar. In case of high flow rates or
unrestricted flow, turbulence may result. The Reynolds number is used to determine the
transition to turbulence. The equation below is for calculation the Reynolds number for
fluid flow in a pipe. A value below 2000 will result in laminar flow. At a value of about
3000 the fluid flow will become uneven. At a value between 7000 and 8000 the flow will
become turbulent.
VD
R = -----------u
where,
R = Reynolds number
V = velocity
D = pipe diameter
= fluid density
u = viscosity
14.2.5.1 - Pressure
Figure 14.22 shows different two mechanisms for pressure measurement. The
Bourdon tube uses a circular pressure tube. When the pressure inside is higher than the
surrounding air pressure (14.7psi approx.) the tube will straighten. A position sensor, connected to the end of the tube, will be elongated when the pressure increases.
pressure
increase
pressure
increase
position sensor
position sensor
pressure
pressure
a) Bourdon Tube
Figure 14.22 Pressure Transducers
b) Baffle
These sensors are very common and have typical accuracies of 0.5%.
differential
pressure
transducer
fluid flow
Aside: Bernoullis equation can be used to relate the pressure drop in a venturi valve.
2
v--p- + ---+ gz = C
2
where,
p = pressure
= density
v = velocity
g = gravitational constant
z = height above a reference
C = constant
Consider the centerline of the fluid flow through the valve. Assume the fluid is incompressible, so the density does not change. And, assume that the center line of the valve does
not change. This gives us a simpler equation, as shown below, that relates the velocity
and pressure before and after it is compressed.
2
2
p before v before
p after v after
---------------- + ------------------ + gz = C = ------------ + -------------- + gz
2
2
p before v before
p after v after
--------------- + ------------------ = ----------- + -------------
2
2
2
p before p after
v after v before
= -------------- ------------------
2
2
The flow velocity v in the valve will be larger than the velocity in the larger pipe section before. So, the right hand side of the expression will be positive. This will mean
that the pressure before will always be higher than the pressure after, and the difference will be proportional to the velocity squared.
result is a distributed force that causes a bending moment, and hence twisting of the pipe.
The amount of bending is proportional to the velocity of the fluid flow. These devices typically have a large constriction on the flow, and result is significant loses. Some of the
devices also use bent tubes to increase the sensitivity, but this also increases the flow resistance. The typical accuracy for a Coriolis flowmeter is 0.1%.
gas flow
pitot
tube
connecting hose
pressure
sensor
14.2.6 Temperature
Temperature measurements are very common with control systems. The temperature ranges are normally described with the following classifications.
very low temperatures <-60 deg C - e.g. superconductors in MRI units
low temperature measurement -60 to 0 deg C - e.g. freezer controls
fine temperature measurements 0 to 100 deg C - e.g. environmental controls
high temperature measurements <3000 deg F - e.g. metal refining/processing
Material
Temperature
Range C (F)
Platinum
-200 - 850 (-328 - 1562)
Nickel
-80 - 300 (-112 - 572)
Copper
-200 - 260 (-328 - 500)
Figure 14.26 RTD Properties
Typical
Resistance
(ohms)
100
120
10
14.2.6.2 - Thermocouples
Each metal has a natural potential level, and when two different metals touch there
is a small potential difference, a voltage. (Note: when designing assemblies, dissimilar
metals should not touch, this will lead to corrosion.) Thermocouples use a junction of dissimilar metals to generate a voltage proportional to temperature. This principle was discovered by T.J. Seebeck.
The basic calculations for thermocouples are shown in Figure 14.27. This calculation provides the measured voltage using a reference temperature and a constant specific
to the device. The equation can also be rearranged to provide a temperature given a voltage.
+
- V out
measuring
device
V out = ( T T ref )
V out
T = ---------- + T ref
where,
V
50
------- (typical)
C
T, T ref = current and reference temperatures
= constant (V/C)
Materials
Temperature
Range
(F)
Voltage Range
(mV)
copper/constantan
-200 to 400
-5.60 to 17.82
iron/constantan
0 to 870
0 to 42.28
chromel/constantan
-200 to 900
-8.82 to 68.78
chromel/aluminum
-200 to 1250
-5.97 to 50.63
platinum-13%rhodium/platinum
0 to 1450
0 to 16.74
platinum-10%rhodium/platinum
0 to 1450
0 to 14.97
tungsten-5%rhenium/tungsten-26%rhenium
0 to 2760
0 to 37.07
mV
80
60
K
J
40
20
0
0
500
1000
1500
2000
2500
(F)
14.2.6.3 - Thermistors
Thermistors are non-linear devices, their resistance will decrease with an increase
in temperature. (Note: this is because the extra heat reduces electron mobility in the semiconductor.) The resistance can change by more than 1000 times. The basic calculation is
shown in Figure 14.29.
often metal oxide semiconductors The calculation uses a reference temperature
and resistance, with a constant for the device, to predict the resistance at another temperature. The expression can be rearranged to calculate the temperature given the resistance.
Rt = Ro e
1 1
--- -----
T T o
T o
T = --------------------------------Rt
T o ln ------ +
R o
where,
Ro, R t = resistances at reference and measured temps.
T o, T = reference and actual temperatures
Aside: The circuit below can be used to convert the resistance of the thermistor to a voltage using a Wheatstone bridge and an inverting amplifier.
+V
R1
R3
R5
+
R2
R4
Vout
Thermistors are small, inexpensive devices that are often made as beads, or metallized surfaces. The devices respond quickly to temperature changes, and they have a
higher resistance, so junction effects are not an issue. Typical accuracies are 1%, but the
devices are not linear, have a limited temperature/resistance range and can be self heating.
14.2.7 Light
14.2.7.1 - Light Dependant Resistors (LDR)
Light dependant resistors (LDRs) change from high resistance (>Mohms) in bright
light to low resistance (<Kohms) in the dark. The change in resistance is non-linear, and is
also relatively slow (ms).
Aside: an LDR can be used in a voltage divider to convert the change in resistance to a
measurable voltage.
V high
These are common in low
cost night lights.
V out
V low
14.2.8 Chemical
14.2.8.1 - pH
The pH of an ionic fluid can be measured over the range from a strong base (alkaline) with pH=14, to a neutral value, pH=7, to a strong acid, pH=0. These measurements
are normally made with electrodes that are in direct contact with the fluids.
14.2.8.2 - Conductivity
Conductivity of a material, often a liquid is often used to detect impurities. This
can be measured directly be applying a voltage across two plates submerged in the liquid
and measuring the current. High frequency inductive fields is another alternative.
14.2.9 Others
A number of other detectors/sensors are listed below,
Combustion - gases such as CO2 can be an indicator of combustion
Humidity - normally in gases
Dew Point - to determine when condensation will form
Analog Input
IN1
REF1
SHLD
device +
#1
-
device +
#1
-
Ain 0
Ain 1
device +
#1
-
Ain 0
Ain 0
device +
#1
-
Ain 2
Ain 1
Ain 3
Ain 1
Ain 4
Ain 2
Ain 5
Ain 2
Ain 6
Ain 3
Ain 7
Ain 3
COM
Single ended - with this arrangement the
signal quality can be poorer, but more
inputs are available.
Note: op-amps are used in this section to implement the amplifiers because they are
inexpensive, common, and well suited to simple design and construction projects.
When purchasing a commercial signal conditioner, the circuitry will be more complex, and include other circuitry for other factors such as temperature compensation.
+V
offset
Ro
Rf
Rg
-V
Vin
gain
+
Ri
Vout
Rf + Rg
V out = ----------------- V in + offset
Ri
Figure 14.34 A Single Ended Signal Amplifier
A differential amplifier with a current input is shown in Figure 14.35. Note that Rc
converts a current to a voltage. The voltage is then amplified to a larger voltage.
R1
Iin
Rc
R2
Rf
Vout
R3
R4
The circuit in Figure 14.36 will convert a differential (double ended) signal to a
single ended signal. The two input op-amps are used as unity gain followers, to create a
high input impedance. The following amplifier amplifies the voltage difference.
Vin
Vout
+
CMRR
adjust
+V
R1
R3
R5
+
R2
Vout
R4
R2
1
1
1
1
V out = V ( R 5 ) ------------------ ------ + ------ + ------ ------
R 1 + R 2 R 3 R 4 R 5 R 3
or if R = R 1 = R 2 = R 4 = R 5
R
V out = V ---------
2R 3
Figure 14.37 A Resistance to Voltage Amplifier
14.5 SUMMARY
Selection of continuous sensors must include issues such as accuracy and resolution.
Angular positions can be measured with potentiometers and encoders (more
accurate).
Tachometers are useful for measuring angular velocity.
14.6 REFERENCES
Bryan, L.A. and Bryan, E.A., Programmable Controllers; Theory and Implementation, Industrial
Text Co., 1988.
Swainston, F., A Systems Approach to Programmable Controllers, Delmar Publishers Inc., 1992.
ratio. If the position of the geared down shaft needs to be positioned to 0.1 degrees, what is the
minimum resolution of the incremental encoder?
8. What is the difference between a strain gauge and an accelerometer? How do they work?
9. Use the equations for a permanent magnet DC motor to explain how it can be used as a tachometer.
10. What are the trade-offs between encoders and potentiometers?
11. A potentiometer is connected to a PLC analog input card. The potentiometer can rotate 300
degrees, and the voltage supply for the potentiometer is +/-10V. Write a ladder logic program
to read the voltage from the potentiometer and convert it to an angle in radians stored in F8:0.
1000 T ref
800 T ref
1
--= ------------------------ = --------------------------0.030
0.040
T ref = 200F
0.040
50V
= --------------------------- = ------------1000 200
F
6.
a)
w
42deg
V out = ( V2 V 1 ) ----------- + V 1 = ( 5V 0V ) ------------------ + 0V = 0.7V
max
300deg
b)
w
2.765V = ( 5V 0V ) ------------------ + 0V
300deg
w
2.765V = ( 5V 0V ) ------------------ + 0V
300deg
w = 165.9deg
7.
deg
output = 0.1 -------------count
input
--------------- = 50
-----1
output
deg
deg
input = 50 0.1 -------------- = 5 ------------- count
count
deg
360 --------rot
count
R = ------------------ = 72 -------------deg
rot
5 -------------count
8.
strain gauge measures strain in a material using a stretching wire that increases resistance - accelerometers measure acceleration with a cantilevered mass on a piezoelectric element.
9.
R
V = Ks
DMM
K
K
+ ------ = V s ------
JR
JR
JR
V s = ( K ) + ------
K
+
-
10.
encoders cost more but can have higher resolutions. Potentiometers have limited
ranges of motion
11.
FS
BT9:0/EN
BTW
Rack: 0
Group: 0
Module: 0
BT Array: BT9:0
Data File: N7:0
Length: 37
Continuous: no
BT9:1/EN
BT9:1/DN
BTR
Rack: 0
Group: 0
Module: 0
BT Array: BT9:1
Data File: N7:37
Length: 20
Continuous: no
CPT
Dest F8:0
Expression
"20.0 * N7:41 / 4095.0 - 10"
CPT
Dest F8:0
Expression
"300.0 * (F8:0 + 10) / 20"
RAD
Source F8:0
Dest F8:1
2. A high precision potentiometer has an accuracy of +/- 0.1% and can rotate 300degrees and is
used as a voltage divider with a of 0V and 5V. The output voltage is being read by an A/D converter with a 0V to 10V input range. How many bits does the A/D converter need to accommodate
the accuracy of the potentiometer?
3. The table of position and voltage values below were measured for an inexpensive potentiometer. Write a C subroutine that will accept a voltage value and interpolate the position value.
theta (deg) V
0
67
145
195
213
296
315
0.1
0.6
1.6
2.4
3.4
4.2
5.0
Topics:
Servo Motors; AC and DC
Stepper motors
Single axis motion control
Hydraulic actuators
Objectives:
To understand the main differences between continuous actuators
Be able to select a continuous actuator
To be able to plan a motion for a single servo actuator
15.1 INTRODUCTION
Continuous actuators allow a system to position or adjust outputs over a wide
range of values. Even in their simplest form, continuous actuators tend to be mechanically
complex devices. For example, a linear slide system might be composed of a motor with
an electronic controller driving a mechanical slide with a ball screw. The cost for such
actuators can easily be higher than for the control system itself. These actuators also
require sophisticated control techniques that will be discussed in later chapters. In general,
when there is a choice, it is better to use discrete actuators to reduce costs and complexity.
Motor Categories
AC motors - rotate with relatively constant speeds proportional to the frequency of the supply power
induction motors - squirrel cage, wound rotor - inexpensive, efficient.
synchronous - fixed speed, efficient
DC motors - have large torque and speed ranges
permanent magnet - variable speed
wound rotor and stator - series, shunt and compound (universal)
Hybrid
brushless permanent magnet stepper motors
Contactors are used to switch motor power on/off
Drives can be used to vary motor speeds electrically. This can also be done with
mechanical or hydraulic machines.
Popular drive categories
Variable Frequency Drives (VFD) - vary the frequency of the power
delivered to the motor to vary speed.
DC motor controllers - variable voltage or current to vary the motor speed
Eddy Current Clutches for AC motors - low efficiency, uses a moving
iron drum and windings
Wound rotor AC motor controllers - low efficiency, uses variable resistors
to adjust the winding currents
A control system is required when a motor is used for an application that requires
continuous position or velocity. A typical controller is shown in Figure 15.1. In any controlled system a command generator is required to specify a desired position. The controller will compare the feedback from the encoder to the desired position or velocity to
determine the system error. The controller with then generate an output, based on the system error. The output is then passed through a power amplifier, which in turn drives the
motor. The encoder is connected directly to the motor shaft to provide feedback of position.
command
generator
(e.g., PLC)
desired position
or velocity
voltage/
current
controller
Figure 15.1
amplified
voltage/
current
power
amp
encoder
motor
I
I
magnetic
field
axis of
rotation
Figure 15.2
A Simplified Rotor
The power is delivered to the rotor using a commutator and brushes, as shown in
Figure 15.3. In the figure the power is supplied to the rotor through graphite brushes rubbing against the commutator. The commutator is split so that every half revolution the
polarity of the voltage on the rotor, and the induced magnetic field reverses to push against
the permanent magnets.
motor
shaft
brushes
split commutator
Top
split commutator
Front
motor
shaft
brushes
V+
Figure 15.3
power
supply
V-
The direction of rotation will be determined by the polarity of the applied voltage,
and the speed is proportional to the voltage. A feedback controller is used with these
motors to provide motor positioning and velocity control.
These motors are losing popularity to brushless motors. The brushes are subject to
wear, which increases maintenance costs. In addition, the use of brushes increases resistance, and lowers the motors efficiency.
ASIDE: The controller to drive a servo motor normally uses a Pulse Width Modulated
(PWM) signal. As shown below the signal produces an effective voltage that is relative to the time that the signal is on. The percentage of time that the signal is on is
called the duty cycle. When the voltage is on all the time the effective voltage delivered is the maximum voltage. So, if the voltage is only on half the time, the effective
voltage is half the maximum voltage. This method is popular because it can produce a variable effective voltage efficiently. The frequency of these waves is normally above 20KHz, above the range of human hearing.
V max
t
V max
t
V max
t
V max
0% duty cycle
0
Veff = --------- V max
100
Figure 15.4
ASIDE: A PWM signal can be used to drive a motor with the circuit shown below. The
PWM signal switches the NPN transistor, thus switching power to the motor. In this
case the voltage polarity on the motor will always be the same direction, so the
motor may only turn in one direction.
V+
V-
V+
signal
source
DC motor
com
Figure 15.5
power
supply
Figure 15.6
+Vs
Va
Vb
Vb
Va
-Vs
15.2.2 AC Motors
Power is normally generated as 3-phase AC, so using this increases the efficiency
of electrical drives.
In AC motors the AC current is used to create changing fields in the motor.
Typically AC motors have windings on the stator with multiple poles. Each pole
is a pair of windings. As the AC current reverses, the magnetic field in the rotor appears to
rotate.
L1
stator windings
rotor
Neut.
Figure 15.7
L2
L1
L3
Neut.
Neut.
Neut.
Figure 15.8
An induction motor has the windings on the stator. The rotor is normally a squirrel
cage design. The squirrel cage is a cast aluminum core that when exposed to a changing
magnetic field will set up an opposing field. When an AC voltage is applied to the stator
coils an AC magnetic field is created, the squirrel cage sets up an opposing magnetic field
and the resulting torque causes the motor to turn.
The motor will turn at a frequency close to that of the applied voltage, but there is
always some slip. It is possible to control the speed of the motor by controlling the frequency of the AC voltage. Synchronous motor drives control the speed of the motors by
synthesizing a variable frequency AC waveform, as shown in Figure 15.9.
Controller
Figure 15.9
These drives should be used for applications that only require a single rotational
direction. The torque speed curve for a typical induction motor is shown in Figure 15.10.
When the motor is used with a fixed frequency AC source the synchronous speed of the
motor will be the frequency of AC voltage divided by the number of poles in the motor.
The motor actually has the maximum torque below the synchronous speed. For example a
motor 2 pole motor might have a synchronous speed of (2*60*60/2) 3600 RPM, but be
rated for 3520 RPM. When a feedback controller is used the issue of slip becomes insignificant.
torque
operating range
synchronous speed
speed
Class A
torque
speed
torque
Class B
speed
torque
Class C
speed
torque
Class D
speed
Wound rotor induction motors use external resistors. varying the resistance
allows the motors torque speed curve to vary. As the resistance value is increased the
motor torque speed curve shifts from the Class A to Class D shapes.
The figure below shows the relationship between the motor speed and applied
power, slip, and number of poles. An ideal motor with no load would have a slip of 0%.
f120
S
RPM = ----------- 1 --------------
p
100%
where,
Single phase AC motors can run in either direction. To compensate for this a
shading pole is used on the stator windings. It basically acts as an inductor to one side of
the field which slows the filed buildup and collapse. The result is that the field strength
seems to naturally rotate.
Thermal protection is normally used in motors to prevent overheating.
Universal motors were presented earlier for DC applications, but they can also be
used for AC power sources. This is because the field polarity in the rotor and stator both
reverse as the AC current reverses.
Synchronous motors are different from induction motors in that they are designed
to rotate at the frequency of the fields, in other words there is no slip.
Synchronous motors use generated fields in the rotor to oppose the stators field.
Starting AC motors can be hard because of the low torque at low speeds. To deal
with this a switching arrangement is often used. At low speeds other coils or capacitors are
connected into the circuits. At higher speeds centrifugal switches disconnect these and the
motor behavior switches.
Single phase induction motors are typically used for loads under 1HP. Various
types (based upon their starting and running modes) are,
- split phase - there are two windings on the motor. A starting winding is
used to provide torque at lower speeds.
- capacitor run - capacitor start
- capacitor start and run
- shaded pole - these motors use a small offset coil (such as a single copper
winding) to encourage the field buildup to occur asymmetrically. These
motors are for low torque applications much less than 1HP.
- universal motors (also used with DC) have a wound rotor and stator that
are connected in series.
Split Winding
Vin
running
winding
squirrel cage
rotor
starting winding
starting capacitor
Capacitor Start
Vin
running
winding
squirrel cage
rotor
starting winding
capacitor
Capacitor Run
Vin
running
winding
squirrel cage
rotor
capacitor winding
running capacitor
Capacitor Start and Capacitor Run Motor
Vin
running
winding
starting capacitor
squirrel cage
rotor
starting winding
V1
V2
V3
torque
speed
1
a
1a
1b
2a
2b
unipolar
bipolar
1a
controller
2a
1b
2b
stepper
motor
Step
1a
2a
1b
2b
1
2
3
4
1
0
0
1
0
1
1
0
1
1
0
0
0
0
1
1
To turn the motor the phases are stepped through 1, 2, 3, 4, and then back to 1.
To reverse the direction of the motor the sequence of steps can be reversed,
eg. 4, 3, 2, 1, 4, ..... If a set of outputs is kept on constantly the motor will be
held in position.
torque
speed
Va
I a = -----Ra
T = Kt Ia
where,
I a, V a, R a = Armature current, voltage and resistance
T = Torque on motor shaft
K t = Motor speed constant
operating
range
Series motors\
- have the rotor and stator coils connected in series.
- as the motor speed increases the current increases, the motor can theoreti-
Va
I a = ----------------Ra + Rf
2
T = Kt Ia = Kt Ia
where,
I a, V a = Armature current, voltage
R a, R f = Armature and field coil resistance
T = Torque on motor shaft
K t = Motor speed constant
stall torque
The XXXXXXX
e f = r a i a + Dla i a + e m
e m = K e D
T = KT ia
e a = ( r a + l a D )i a + K e D
T
e a = ( r a + l a D ) ------ + K e D
K T
Figure 15.19 Equations for an armature controlled DC motor
Compound motors\
- have the rotor and stator coils connected in series.
- differential compound motors have the shunt and series winding field
aligned so that they oppose each other.
- cumulative compound motors have the shunt and series winding fields
aligned so that they add
cumulative
differential
ef = rf if + l f if D
T = KT if
T
--- = JD 2 + BD
1 --- = ----------------------2
T
JD + BD
KT
--- = --- T
--- = ----------------------2
if
T if
JD + BD
KT 1
i
---- = --- ----f = ----------------------- -----------------ef
if ef
JD 2 + BD r f + l f D
i
T
1 -
---- = T
--- ----f = K T ---------------- r f + l f D
ef
if ef
15.3 HYDRAULICS
Hydraulic systems are used in applications requiring a large amount of force and
slow speeds. When used for continuous actuation they are mainly used with position feedback. An example system is shown in Figure 15.21. The controller examines the position
of the hydraulic system, and drivers a servo valve. This controls the flow of fluid to the
actuator. The remainder of the provides the hydraulic power to drive the system.
valve
controller
position
sensor
hydraulic
power
supply
position
hydraulic
actuator
sump
15.5 SUMMARY
AC motors work at higher speeds
DC motors work over a range of speeds
Motion control introduces velocity and acceleration limits to servo control
Hydraulics make positioning easy
in
P = 0.125 -------
rot
in
R = 0.025 ---------step
in
0.025 ---------R
step
rot
= --- = --------------------------- = 0.2 ---------P
step
in
0.125 -------
rot
b)
c)
Thus
1
step
-----------------= 5 ---------rot
rot
0.2 ---------step
rot
deg
= 0.2 ---------- = 72 ---------step
step
in
3 ----s
steps
PPS = ------------------------ = 120 ------------in
s
0.025 ---------step
2.
a)
step
pulses = ( 3rot ) 5 ---------- = 15steps
rot
b)
pulses
rot- step
steps
1min steps
step
---------------- = 25 -------5 ---------- = 125 ------------- = 125 ------------- ------------- = 2.08 ---------
s
min
rot
min
60s min
s
3.
stepper motor
speed
very low speeds
torque
low torque
vfd
dc motor
wide range
How many pulses are required to rotate the motor through three complete revolutions? If it is desired to rotate the motor at a speed of 25 rev/min, what pulse
rate must be generated by the robot controller?
2. Describe the voltage ripple that would occur when using a permanent magnet DC motor as a
tachometer. Hint: consider the use of the commutator to switch the polarity of the coil.
3. Compare the advantages/disadvantages of DC permanent magnet motors and AC induction
motors.
Topics:
Motion controllers
Motion profiles, trapezoidal and smooth
Gain schedulers
Objectives:
To understand single and multi axis motion control systems.
16.1 INTRODUCTION
A system with a feedback controller will attempt to drive the system to a state
described by the desired input, such as a velocity. In earlier chapters we simply chose step
inputs, ramp inputs and other simple inputs to determine the system response. In practical
applications this setpoint needs to be generated automatically. A simple motion control
system is used to generate setpoints over time.
An example of a motion control system is shown in Figure 16.1. The motion controller will accept commands or other inputs to generate a motion profile using parameters
such as distance to move, maximum acceleration and maximum velocity. The motion profile is then used to generate a set of setpoints, and times they should be output. The setpoint scheduler will then use a real-time clock to output these setpoints to the motor drive.
motion controller
Motor
motion
commands
Setpoint
Generator
Servo
Drive
Setpoint
Scheduler
setpoint schedule
motion profile
0.4
t
0
0.2
Figure 16.1
0.6
0.8
t (s)
setpoint
0
0.1
0.2
0.3
0.4
.....
0.0
0.2
0.4
0.4
0.4
....
A motion controller
The combination of a motion controller, drive and actuator is called an axis. When
there is more than one drive and actuator the system is said to have multiple axes. Complex motion control systems such as computer controlled milling machines (CNC) and
robots have 3 to 6 axes which must be moved in coordination.
( deg )
100
20
t (s)
Figure 16.2
A trapezoidal velocity profile is shown in Figure 16.3. The area under the curve is
the total distance moved. The slope of the initial and final ramp are the maximum acceleration and deceleration. The top level of the trapezoid is the maximum velocity. Some controllers allow the user to use the acceleration and deceleration times instead of the
maximum acceleration and deceleration. This profile gives a continuous acceleration, but
there will be a jerk (third order derivative) at the four sharp corners.
( deg )
max
max
max
t (s)
0
0
where,
t acc
t max
t dec
t total
The basic relationships for these variables are shown in Figure 16.4. The equations
can be used to find the acceleration and deceleration times. These equations can also be
used to find the time at the maximum velocity. If this time is negative it indicates that the
axis will not reach the maximum velocity, and the acceleration and deceleration times
must be decreased. The resulting velocity profile will be a triangle.
max
t acc = t dec = ----------- max
(1)
(2)
t acc
t dec
1
1
= --- t acc max + t max max + --- t dec max = max -------- + t max + --------
2
2
2
2
t acc
t dec
(3)
(4)
Note: if the time calculated in equation 4 is negative then the axis never reaches
maximum velocity, and the velocity profile becomes a triangle.
Figure 16.4
For the example in Figure 16.5 the move starts at 100deg and ends at 20 deg. The
acceleration and decelerations are completed in half a second. The system moves for 7.5
seconds at the maximum velocity.
Given,
start = 100deg
end = 20deg
deg
max = 10 --------s
deg
10 -------- max
s
= ------------ = --------------- = 0.5s
max
deg20 -------2
s
80deg 0.5s
0.5s
- ---------- = --------------t max = ------------ --------- ---------- ---------- = 7.5s
max
2
2
deg
2
2
10 --------s
t total = t acc + t max + t dec = 0.5s + 7.5s + 0.5s = 8.5s
Figure 16.5
The motion example in Figure 16.6 is so short the axis never reaches the maximum
velocity. This is made obvious by the negative time at maximum velocity. In place of this
the acceleration and deceleration times can be calculated by using the basic acceleration
position relationship. The result in this example is a motion that accelerates for 0.316s and
then decelerates for the same time.
Given,
start = 20deg
end = 22deg
deg
max = 10 --------s
- =
----------max
2deg =
--------------deg20 -------2
s
0.1s = 0.316s
t max = 0s
Figure 16.6
Given the parameters calculated for the motion, the setpoints for motion can be
calculated with the equations in Figure 16.7.
Figure 16.7
A subroutine that implements these is shown in Figure 16.8. In this subroutine the
time is looped with fixed time steps. The position setpoint values are put into the setpoint
array, which will then be used elsewhere to guide the mechanism.
void generate_setpoint_table(
double t_acc, double t_max, double t_step,
double vel_max, double acc_max,
double theta_start, double theta_end,
double setpoint[], int *count){
double t, t_1, t_2, t_total;
t_1 = t_acc;
t_2 = t_acc + t_max;
t_total = t_acc + t_max + t_acc;
*count = 0;
for(t = 0.0; t <= t_total; t += t_step){
if( t < t_1){
setpoint[*count] = 0.5*acc_max*t*t + theta_start;
} else if ( (t >= t_1) && (t < t_2)){
setpoint[*count] = 0.5*acc_max*t_acc*t_acc
+ vel_max*(t - t_1) + theta_start;
} else if ( (t >= t_2) && (t < t_total)){
setpoint[*count] = 0.5*acc_max*t_acc*t_acc
+ vel_max*(t_max)
+ 0.5*acc_max*(t-t_2)*(t-t_2) + theta_start;
} else {
setpoint[*count] = theta_end;
}
*count++;
}
setpoint[*count] = theta_end;
*count++;
}
Figure 16.8
In some cases the jerk should be minimized. This can be achieved by replacing the
acceleration ramps with a smooth polynomial, as shown in Figure 16.9. In this case two
quadratic polynomials will be used for the acceleration, and another two for the deceleration.
( t ) = At + Bt + C
( deg )
max
max
max
t (s)
0
0
where,
t acc
t max
t dec
t total
Given,
start
end
max
max
dt
dt
2
These can be used to calculate the polynomial coefficients,
2
C = 0
B = 0
0 = A0 + B0 + C
0 = 2A0 + B
max = At acc
max
A = -----------2t acc
max
A = -----------2t acc
2 max
t acc = -------------max
2
max
max
max
A = ------------ = ------------------------ = --------------2 max
2t acc
4 max
2 ---------------
max
The equation for the first segment is,
2
max 2
( t ) = --------------- t
4 max
t acc
0 t < -------2
The equation for the second segment can be found using the first segment,
2
max
2
( t ) = max --------------- ( t acc t )
4 max
2
max 2
2
( t ) = max --------------- ( t 2t acc t + t acc )
4 max
t acc
------- t < t acc
2
The distance covered during acceleration, the area under the curves, is,
acc =
t acc
-------2
t acc
max 2
max 2
--------
acc
acc
max 3
= ------------------ t
12 max
2
t acc
-------2
max t 3
2
2
+ max t --------------- ---- t acc t + t acc t
4 max 3
t acc
t acc
------2
2
3
3
max t acc t acc
max t acc
max t acc 3
t acc
t acc
- -------- + max t acc -------------- -------- t acc + t 3acc max ------- + --------------- -------- -------- + -------
= ----------------12 max 8
4 max 3
2
4 max 24
4
2
acc
7 max 3
max 3
max t acc
max 3
= ------------------ t acc + --------------------- ------------------ t acc ------------------ t acc
96 max
2
12 max
96 max
acc
14max 3
max t acc
= --------------------- t acc + --------------------96 max
2
(1)
where,
start, end = start and end positions of motion
t start, t end = start and end times for the motion
3
p ( u ) = Au + Bu + Cu + D
(2)
p(1 ) = 1
d---p(0) = 0
dt
d---p(1 ) = 0
dt
0 = A0 + B0 + C0 + D
D = 0
C = 0
3
B = --- A
2
0 = 3A0 + 2B0 + C
0 = 3A1 + 2B1
3
1 = A1 + B1 + ( 0 )0 + 0
A = 2
B = 3
p ( u ) = 2 u + 3u
(3)
where,
p ( u ) = A sin ( Bt + C ) + D
The coefficients can be calculated using the conditions used previously,
d---p ( 0 ) = AB cos ( B ( 0 ) + C ) = 0
dt
C = -- cos ( C ) = 0
2
d---p ( 1 ) = AB cos ( B ( 1 ) + C ) = 0
dt
cos ( B + C ) = 0
B + C = --2
p ( 0 ) = A sin B ( 0 ) --- + D = 0
p ( 1 ) = A sin ( 1 ) --- + D = 1
2
A ( 1 ) + A = 1
B =
A = D
1
A = --2
1
p ( u ) = --- sin t --- + --2
2
2
#define
int
int
int
int
int
TABLE_SIZE
11
point_master[TABLE_SIZE] = {0, 24, 95, 206, 345, 500, 655, 794, 905, 976, 1000};
point_position[TABLE_SIZE];
point_time[TABLE_SIZE];
point_start_time;
point_index;
int
int
max
t acc
t max
t dec
The area under the velocity curve is the distance (angle in this case) travelled. First
we can determine the distance covered during acceleration, and deceleration and
the time during acceleration, and deceleration.
max
20 40 50
t acc = t dec = ------------ = ------, ---------, --------- = ( 0.4, 0.4, 0.333 )sec.
50 100 150
max
t acc max.vel.
0.4 ( 20 ) 0.4 ( 40 ) 0.333 ( 50 )
acc. = dec. = ---------------------------- = ------------------, ------------------, ------------------------ = ( 4, 8, 8.33 )deg.
2
2
2
2
The next step is to examine the moves specified,
move = end start = ( 120 40, 0 80, 0 ( 40 ) ) = ( 80, 80, 40 )deg.
Remove the angles covered during accel./deccel., and find the travel time at maximum
velocity.
move 2 acc
80 2 ( 4 ) 80 2 ( 8 ) 40 2 ( 8.333 )
t max = ----------------------------------- = -----------------------, -----------------------, ---------------------------------
max
20
40
50
t max = ( 3.6, 1.6, 0.46668 )sec.
t total = t acc + t max + t dec = ( 4.4, 2.4, 1.13 )s
The longest motion time is 4.4s for joint 1, and this can be used to prolong the other
motions. The calculation begins by rewriting the velocity/position relationship using
a new maximum velocity.
max'
1 max'
1 max'
= --- ------------- max' + t total 2 ------------- max' + --- ------------- max'
2 max
max
2 max
max'
= max' t total -------------
max
2
401.99502
max' = ---------------------------------------- = 421, 19.0
2
19.0
t acc' = t acc' = ---------- = 0.19s
100
A new maximum velocity can be calculated for joint 3 using this equation.
2
should be updated. A diagram of a scheduler is shown in Figure 16.18. In this system the
setpoint scheduler is an interrupt driven subroutine that compares the system clock to the
total motion time. When enough time has elapsed the routine will move to the next value
in the setpoint table. The frequency of the interrupt clock should be smaller than or equal
to the time steps used to calculate the setpoints. The servo drive is implemented with an
algorithm such a PID control.
Choose new
point from
trajectory table
desired
Read desired
Compute
error
Interrupt
Clock
Output actuator
signal
Return
speed
actual position
trajectory table
time
trajectory table time step
position
required
actual
time
For complex surfaces we want to contour appropriately. These surfaces will almost always
be represented with spline patches.
(u=0,v=1)
(u=1,v=1)
xp
p ( u, v ) = y p
zp
(u=0,v=0)
(u=1,v=0)
t t start
( t ) = start + ( end start )p --------------------------
t end t start
3
p ( u ) = 2 u + 3u
t t start
( t ) = ( end start )p --------------------------
t end t start
2
p ( u ) = 6 u + 6u
JRD + K
-----------------------K
physical system
P
+
1---------K enc
K
-----------------------2
JRD + K
K enc
16.6 SUMMARY
Axis limits can be used to calculate motion profiles.
Trapezoidal and smooth motion profiles were presented.
Motion profiles can be used to generate setpoint tables.
Values from the setpoints can then be output by a scheduler to drive an axis.
t max
t total
t (s)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
10.0
10.5
11.0
deg
200deg 20 --------- 0.2s
max t acc
s
= ---------------------------------- = --------------------------------------------------- = 9.8s
max
deg
20 --------s
= t acc + t max + t dec = 0.2s + 9.8s + 0.2s = 10.2s
angle
(deg)
-100
-92
-82
-72
-62
-52
-42
-32
-22
-12
-2
8
18
28
38
48
58
68
78
88
98
100
100
1
degdeg
2
( 0.2s ) + 20 --------- ( 0.5s 0.2s ) 100deg
0.5s = --- 100 -------2
2
s
s
0.5s = 92deg
deg
1.0s = 0.5s + 20 --------- ( 0.5s ) = 92deg + 10deg
s
2.
t
(s)
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
counts
100
283
483
683
883
1083
1283
1483
1683
1883
2083
2283
2483
2683
2883
3000
moves to -1000 inches. The maximum velocity is 100 in/s and the maximum acceleration is
50in/s/s.
Topics:
Laplace transforms
Using tables to do Laplace transforms
Using the s-domain to find outputs
Solving Partial Fractions
Objectives:
To be able to find time responses of linear systems using Laplace transforms.
17.1 INTRODUCTION
Laplace transforms provide a method for representing and analyzing linear systems using algebraic methods. In systems that begin undeflected and at rest the Laplace s
can directly replace the d/dt operator in differential equations. It is a superset of the phasor
representation in that it has both a complex part, for the steady state response, but also a
real part, representing the transient part. As with the other representations the Laplace s is
related to the rate of change in the system.
D = s
s = + j
Figure 17.1
The Laplace s
The basic definition of the Laplace transform is shown in Figure 17.2. The normal
convention is to show the function of time with a lower case letter, while the same function in the s-domain is shown in upper case. Another useful observation is that the transform starts at t=0s. Examples of the application of the transform are shown in Figure 17.3
for a step function and in Figure 17.4 for a first order derivative.
F(s) =
0 f ( t )e
st
dt
where,
f ( t ) = the function in terms of time t
F ( s ) = the function in terms of the Laplace s
Figure 17.2
Figure 17.3
f ( t )e
st
dt =
5e
st
5 st
dt = --- e
s
s0
5 s
5
5e
= --- e
------------ = --s
s
s
d
st
G ( s ) = L [ g ( t ) ] = L ----- f ( t ) = (d/dt)f ( t )e dt
dt
0
we can use integration by parts to go backwards,
b
u dv = uv a v du
b
0 (d/dt)f ( t )e
st
dt
therefore,
st
du = df ( t )
v = e
u = f(t)
dv = se
f ( t ) ( s )e
st
dt = f ( t )e
(d/dt)f ( t )e
st
st
0
dt = [ f ( t )e
st
dt
(d/dt)f ( t )e
st
dt
f ( t )e
0s
] + s f ( t )e
st
dt
d
L ----- f ( t ) = f ( 0 ) + sL [ f ( t ) ]
dt
Figure 17.4
The previous proofs were presented to establish the theoretical basis for this
method, however tables of values will be presented in a later section for the most popular
transforms.
FREQUENCY DOMAIN
TIME DOMAIN
f(t )
f(s )
Kf ( t )
KL [ f ( t ) ]
f1 ( t ) + f2 ( t ) f3 ( t ) +
df ( t )
----------dt
f 1 ( s ) + f 2 ( s ) f 3 ( s ) +
sL [ f ( t ) ] f ( 0 )
d f(t )
-------------2
dt
n
d------------f ( t )n
dt
n1
s L[f(t ) ] s
0 f ( t ) dt
L
[ f ( t ) -]
---------------s
f ( t a )u ( t a ), a > 0
at
f(t)
f ( at ), a > 0
tf ( t )
n
t f(t)
as
L[ f( t )]
f(s a)
1--- --s-
f
a a
-------------df ( s -)
ds
n
n d f(s)
( 1 ) -------------n
ds
f------( t )t
f ( u ) du
s
Figure 17.5
df ( 0 )
s L [ f ( t ) ] sf ( 0 ) -----------------dt
2
f(0 ) s
n 2 df ( 0
n1
d
f(0
------------------) ------------------------n
dt
dt
L [ x + 7x + 8x = 9 ] =
Figure 17.6
where,
x(0) = 1
x(0) = 2
x(0) = 3
The Laplace transform tables shown in Figure 17.7 and Figure 17.8 are normally
used for converting to/from the time/s-domain.
TIME DOMAIN
(t)
unit impulse
step
ramp
FREQUENCY DOMAIN
A
--s
1
---2
s
2
----
s
n!
-----------
t ,n>0
s
e
at
exponential decay
sin ( t )
cos ( t )
n+1
1---------s+a
----------------2
2
s +
s
----------------2
s +
at
1
------------------2
(s + a)
2 at
2! -----------------3
(s + a)
te
t e
Figure 17.7
TIME DOMAIN
e
e
at
sin ( t )
at
cos ( t )
at
FREQUENCY DOMAIN
------------------------------2
2
(s + a) +
s+a ------------------------------2
(s + a) +
------------------------------2
2
(s + a) +
sin ( t )
at
2Ae
C aB
B cos t + ----------------- sin t
t
2t A e
complex conjugate
complex conjugate
A
A
------------------------------ + ------------------------------------2
2
( s + j )
( s + + j )
cos ( t + )
at
bt
( c a )e ( c b )e
--------------------------------------------------------ba
at
bt
e e
-----------------------ba
Figure 17.8
Bs + C
-------------------------------2
2
(s + a) +
A
A
----------------------+ -------------------------------------s + j
s + + j
cos ( t + )
s+c
-------------------------------( s + a)(s + b)
1
-------------------------------( s + a)(s + b)
f ( t ) = 5 sin ( 5t + 8 )
f(s ) = L[ f( t) ] =
Figure 17.9
5
6
f ( s ) = --- + ----------s s+7
1
f(t ) = L [ f(s )] =
Figure 17.10 Drill Problem: Converting from the s-domain to time domain
shown in the mass-spring-damper example in Figure 17.11. In this case we assume the
system starts undeflected and at rest, so the D operator may be directly replaced with the
Laplace s. If the system did not start at rest and undeflected, the D operator would be
replaced with a more complex expression that includes the initial conditions.
F = MD x + K d Dx + K s x
Ks
Kd
F
( t )2
--------= MD + K d D + K s
x( t)
F(t)
F( s)
2
L ---------- = ----------- = Ms + K d s + K s
x( t)
x(s )
d f ( t -)
d f ( 0 )n
n1
n 2 df ( 0 )
------------------ ------------------L ------------= s L[f(t )] s
f(0 ) s
n
n
dt
dt
dt
Device
Time domain
Resistor
V ( t ) = RI ( t )
V ( s ) = RI ( s )
Z = R
1
V ( t ) = ---- I ( t ) dt
C
d
V ( t ) = L ----- I ( t )
dt
1 I(s)
V ( s ) = ---- -------- C s
1
Z = -----sC
Z = Ls
Capacitor
Inductor
s-domain
V ( s ) = LsI ( s )
Impedance
Figure 17.13 shows an example of circuit analysis using Laplace transforms. The
circuit is analyzed as a voltage divider, using the impedances of the devices. The switch
that closes at t=0s ensures that the circuit starts at rest. The calculation result is a transfer
function.
L
t=0sec
Vi
+
-
+
Vo
-
1
Vi ------------------
R
1
V i ----------------------
DC + ---
1
+
DCR
R
R
R
D R LC + DLR + R
----------------------
DLR
+
1
1 + DCR
DL + ------------------
1
DC + ---
R
Vo
R
------ = -------------------------------------------- 2 2
Vi
s R LC + sLR + R
= 0 for t < 0
= A for t >= 0
A
F ( s ) = L [ F ( t ) ] = --s
Figure 17.14 An input function
In the previous section we converted differential equations, for systems, to transfer
functions in the s-domain. These transfer functions are a ratio of output divided by input.
If the transfer function is multiplied by the input function, both in the s-domain, the result
is the system output in the s-domain.
Given,
x ( s )1
---------= ------------------------------------2
F( s)
Ms + K d s + K s
A
F ( s ) = --s
Therefore,
x( s)
1
- A
--x ( s ) = ----------- F ( s ) = ------------------------------------2
F ( s )
Ms + K d s + Ks s
Assume,
Ns
K d = 3000 -----m
N
K s = 2000 ---m
M = 1000kg
A = 1000N
1
x ( s ) = --------------------------------2
( s + 3s + 2 )s
smaller, simpler components. The previous example in Figure 17.15 is continued in Figure
17.16 using a partial fraction expansion. In this example the roots of the third order
denominator polynomial, are calculated. These provide three partial fraction terms. The
residues (numerators) of the partial fraction terms must still be calculated. The example
shows a method for finding resides by multiplying the output function by a root term, and
then finding the limit as s approaches the root.
1
A
B
C
1
= ------------------------------------ = --- + ----------- + ----------x ( s ) = --------------------------------2
( s + 1 ) ( s + 2 )s
s s+1 s+2
( s + 3s + 2 )s
1
A = lim s ------------------------------------
( s + 1 ) ( s + 2 )s
s0
1
= --2
1
B = lim ( s + 1 ) ------------------------------------
( s + 1 ) ( s + 2 )s
s 1
= 1
1
C = lim ( s + 2 ) ------------------------------------
( s + 1 ) ( s + 2 )s
s 2
1
= --2
Aside: the short cut above can reduce time for simple partial fraction
expansions. A simple proof for finding B above is given in this box.
1
B - ---------C----------------------------------- = A
--- + ---------+
( s + 1 ) ( s + 2 )s
s s+1 s+2
C
1
A
B
( s + 1 ) ------------------------------------ = ( s + 1 ) --- + ( s + 1 ) ----------- + ( s + 1 ) ----------( s + 1 ) ( s + 2 )s
s
s+1
s+2
1 A
C
-----------------= ( s + 1 ) --- + B + ( s + 1 ) ----------( s + 2 )s
s
s+2
A
1
lim ------------------- = lim ( s + 1 ) --s
(
s
+
2
)s
s 1
s 1
C
+ lim B + lim ( s + 1 ) ----------s
+2
s 1
s 1
1
lim ------------------- = lim B = B
s 1 ( s + 2 )s
s 1
1
0.5
1
0.5
= ------- + ----------- + ----------x ( s ) = --------------------------------2
s
s
+
1
s
+2
( s + 3s + 2 )s
x(t) = L [ x( s)] = L
x(t) = L
0.5
1 - ---------0.5------- + ---------+
s
s+1 s+2
0.5
1 1 0.5
------- + L 1 ---------+ L ----------s
s+1
s+2
t
x ( t ) = [ 0.5 ] + [ ( 1 )e ] + [ ( 0.5 )e
t
x ( t ) = 0.5 e + 0.5e
2t
2t
yes
is the
function in
the transform
tables?
no
can the
function be
simplified?
simplify the
function
yes
no
Use partial fractions to
break the function into
smaller parts
Figure 17.18 The methodology for doing an inverse transform of an output function
Figure 17.19 shows the basic procedure for partial fraction expansion. In cases
where the numerator is greater than the denominator, the overall order of the expression
can be reduced by long division. After this the denominator can be reduced from a polynomial to multiplied roots. Calculators or computers are normally used when the order of the
polynomial is greater than second order. This results in a number of terms with unknown
resides that can be found using a limit or algebra based technique.
is the
order of the
numerator >=
denominator?
yes
no
Find roots of the denominator
and break the equation into
partial fraction form with
unknown values
OR
use limits technique.
If there are higher order
roots (repeated terms)
then derivatives will be
required to find solutions
Done
5s + 3s + 8s + 6x ( s ) = ------------------------------------------2
s +4
This cannot be solved using partial fractions because the numerator is 3rd order
and the denominator is only 2nd order. Therefore long division can be used to
reduce the order of the equation.
5s + 3
2
s +4
5s + 3s + 8s + 6
3
5s + 20s
2
3s 12s + 6
2
3s + 12
12s 6
This can now be used to write a new function that has a reduced portion that can be
solved with partial fractions.
12s 6x ( s ) = 5s + 3 + --------------------2
s +4
solve
--------------------12s 6A
B
= ------------- + ------------2
s + 2j s 2j
s +4
Figure 17.20 Partial fractions when the numerator is larger than the denominator
Partial fraction expansion of a third order polynomial is shown in Figure 17.21.
The s-squared term requires special treatment. Here it produces partial two partial fraction
terms divided by s and s-squared. This pattern is used whenever there is a root to an exponent.
A B
1 C
= ---2- + --- + ----------x ( s ) = -------------------2
s s+1
s
s (s + 1)
1 -
C = lim ( s + 1 ) -------------------2
s 1
s (s + 1)
1 -
2
A = lim s -------------------2
s0
s (s + 1)
= 1
1
= lim ----------- = 1
s
+
1
s0
d 2
1 -
B = lim ----- s -------------------2
ds
s0
s (s + 1)
d
1
= lim ----- -----------
ds
s
+
1
s0
= lim [ ( s + 1 ) ] = 1
s0
5
F ( s ) = -----------------------32
s (s + 1)
5 A B
C
D
E
----------------------= ----2 + --- + ------------------3- + ------------------2- + ---------------3
2
s (s + 1)
(s + 1)
s
(s + 1)
s (s + 1)
A B
C
D
E
5 ----------------------= ----2 + --- + ------------------3- + ------------------2- + ---------------3
2
s
(
s
+
1)
s
(s + 1)
(s + 1)
s (s + 1)
3
A ( s + 1 ) + Bs ( s + 1 ) + Cs + Ds ( s + 1 ) + Es ( s + 1 ) = ----------------------------------------------------------------------------------------------------------------------------------------3
2
s (s + 1)
4
s ( B + E ) + s ( A + 3B + D + 2E ) + s ( 3A + 3B + C + D + E ) + s ( 3A + B ) + ( A )= -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------3
2
s (s + 1)
0
1
3
3
1
1
3
3
1
0
0
0
1
0
0
0
1
1
0
0
1
2
1
0
0
A
0
B
0
=
C
0
D
0
E
5
A
0 1 0
B
1 3 0
=
C
3 3 1
D
3 1 0
E
1 0 0
0
1
1
0
0
1
2
1
0
0
0
5
0
15
=
0
5
0
10
5
15
5 5 15
5
10
15
----------------------= ---2- + --------- + ------------------3- + ------------------2- + ---------------3
2
s
(
s
+ 1)
s
(s + 1)
(s + 1)
s (s + 1)
A B
C
D
E
5 ----------------------= ---2- + --- + ------------------3- + ------------------2- + ---------------3
2
s
(
s
+
1)
s
(s + 1)
(s + 1)
s (s + 1)
5
5
2
A = lim -----------------------3- s = lim ------------------3- = 5
2
s0
s 0 (s + 1)
s (s + 1)
d
5
d
5
2
B = lim ----- -----------------------3- s = lim ----- ------------------3-
2
ds
ds
s0
s0
(s + 1)
s (s + 1)
5( 3 )
= lim ------------------4- = 15
s 0 (s + 1)
5
5
3
C = lim -----------------------3- ( s + 1 ) = lim ----2 = 5
2
s 1
s 1 s
s (s + 1)
5
1 d
1 d 5
1 2 ( 5 -)
3
D = lim ----- ----- -----------------------3- ( s + 1 ) = lim ----- ----- ---2- = lim ----- ------------= 10
s 1 1! ds 2
s 1 1! ds s
s 1 1! s 3
s (s + 1)
1 d2
5
1 d25
1 30
3
E = lim ----- ----- -----------------------3- ( s + 1 ) = lim ----- ----- ----2 = lim ----- -----4- = 15
2
2!
ds
2!
ds
2!
s 1
s 1
s 1
s
s
s (s + 1)
5 5 15
5
10
15
----------------------= ---2- + --------- + ------------------3- + ------------------2- + ---------------3
2
s
(s + 1)
s
(s + 1)
(s + 1)
s (s + 1)
5 A B
C
D
E
----------------------= ---2- + --- + ------------------3- + ------------------2- + ---------------3
2
s (s + 1)
(s + 1)
s
(s + 1)
s (s + 1)
5
A B
C
D
E
lim -----------------------3- = ---2- + --- + ------------------3- + ------------------2- + ---------------s (s + 1)
(s + 1)
s 1 2
s
(s + 1)
s (s + 1)
5
A B
C
D
E
3
lim ( s + 1 ) -----------------------3- = ---2- + --- + ------------------3- + ------------------2- + ----------------
2
s (s + 1)
( s + 1 )
s 1
s
(s + 1)
s (s + 1)
3
5 A(s + 1)
(s + 1)2
-+B
---------------------lim ---2- = ---------------------+ C + D(s + 1) + E(s + 1)
2
s
s 1 s
s
5
A ( 1 + 1 ) - --------------------------B( 1 + 1 ) 2
------------- = --------------------------+
+
+
+
C
D
(
1
+
1
)
E
(
1
+
1
)
2
2
1
( 1 )
( 1 )
3
5
A ( 0 ) - ------------B( 0) 2
------------- = ------------+
+ C + D(0) + E(0)
2
2
1
( 1 )
( 1 )
C = 5
2
s
s 1 dt s
2 ( s + 1 ) 3 3 ( s + 1 ) 2
( s + 1 ) 3 3 ( s + 1 ) 2
------------2 ( 5 )----------------------------------------- + ---------------------- + D + 2E ( s + 1 )
+
lim
= A
+ B -----------------3
3
2
2
s
s 1
s
s
s
s
------------2 ( 5 )= D = 10
3
( 1 )
For E, differentiate twice, then evaluate (the terms for A and B will be ignored to save
space, but these will drop out anyway).
3
3
2 5
d
A
(
s
+
1
)
B
(
s
+
1
)
- + ----------------------- + C + D ( s + 1 ) + E ( s + 1 )2
lim ----- ----2 = ---------------------2
s
s 1 dt s
s
d 2 ( 5 -)
= A ( ) + B ( ) + D + 2E ( s + 1 )
lim ----- ------------
s 1 dt s 3
3 ( 2 ( 5 ) )
lim ------------------------= A ( ) + B ( ) + 2E
4
s 1
s
------------------------3 ( 2 ( 5 ) )
= A ( 0 ) + B ( 0 ) + 2E
4
( 1 )
E = 15
Figure 17.25 A proof of the need for differentiation for repeated roots
17.6 EXAMPLES
Given,
1---x(s)
M
----------- = --------------------------------F( s)
K
2 K
s + -----d- s + -----s
M
M
Component values are,
M = 1kg
N
K s = 2 ---m
Ns
K d = 0.5 -----m
x(s)
5s - ----------------------------1
-
x ( s ) = F ( s ) ----------- = --------------2 2
F ( s )
2
s + 6 s + 0.5s + 2
5s
x ( s ) = --------------------------------------------------------2
2
( s + 36 ) ( s + 0.5s + 2 )
A
B
C
D
x ( s ) = ------------- + ------------- + ------------------------------------- + ------------------------------------s + 6j s 6j s 0.25 + 1.39j s 0.25 1.39j
A =
30j
( s + 6j ) ( 5s )
------------------------------------------------------------------------ = ---------------------------------------------2
( 12j ) ( 36 3j + 2 )
( s 6j ) ( s + 6j ) ( s + 0.5s + 2 )
lim
s 6j
A = 73.2 10 3.05
*
B = A = 73.2 10 3.05
Continue on to find C, D same way
3
x ( t ) = 2 ( 73.2 10 )e
0t
cos ( 6t 3.05 ) +
17.6.2 Circuits
It is not necessary to develop a transfer functions for a system. The equation for the
voltage divider is shown in Figure 17.28. Impedance values and the input voltage are converted to the s-domain and written in the equation. The resulting output function is manipulated into partial fraction form and the residues calculated. An inverse Laplace transform
is used to convert the equation into a function of time using the tables.
R = 10
t=0
V s = 3 cos t
+
-
+
Vo
-
C = 10 F
As normal, relate the source voltage to the output voltage using component values in
the s-domain.
ZC
V o = Vs -------------------
ZR + ZC
1
3s
Z C = -----ZR = R
V s ( s ) = ------------sC
2
s +1
Next, equations are combined. The numerator of resulting output function must be
reduced by long division.
1-
---- sC
3s
3s
3s
- ---------------- = ------------------------------------------- = ------------------------------------------------------V o = ------------2
2
2
3 3
1-
s + 1 R + ----( s + 1 ) ( 1 + sRC )
( s + 1 ) ( s10 10 + 1 )
sC
The output function can be converted to a partial fraction form and the residues calculated.
2
2
3s
As + B- ---------CAs
+ As + Bs + B + Cs + C----------------------------------------------------------------------------------=
+
=
V o = ----------------------------------2
2
2
(s + 1 )(s + 1)
s +1 s+1
(s + 1)( s + 1)
2
3s
s ( A + C ) + s ( A + B ) + ( B + C )V o = ----------------------------------= ---------------------------------------------------------------------------2
2
(s + 1)(s + 1)
(s + 1)(s + 1)
B+C = 0
A+C = 0
A+B = 3
B = C
A = C
C C = 3
C = 1.5
A = 1.5
B = 1.5
The output function can be converted to a function of time using the transform tables,
as shown below.
1
1.5s
+ 1.5- --------- 1.5 1 1.5s + 1.5
1.5----------------------- + L 1 ---------+
= L ----------------------2
2
s
+1
s
+
1
s +1
s +1
s 1 1
t
------------+ 1.5L ------------ 1.5e
2
2
s +1
s +1
Vo ( t ) = L [ V o ( s ) ] = L
V o ( t ) = 1.5L
1.5
2
2
t
1.5 + 1.5 sin t + atan ------- 1.5e
1.5
t
V o ( t ) = 2.121 sin t + --- 1.5e
4
f(t)
5
t
seconds
0
f ( t ) = 5tu ( t ) 5 ( t 1 )u ( t 1 ) 5 ( t 3 )u ( t 3 ) + 5 ( t 4 )u ( t 4 )
s
3s
4s
5 5e - 5e
5e
-----------+ -----------f ( s ) = ---2- --------2
2
2
s
s
s
s
Figure 17.30 Switching on and off function parts
x ( t ) = lim [ sx ( s ) ]
s0
1s
1
1
- = ------------------------------------- = 1--x ( t ) = lim --------------------------------= lim ------------------------2
2
2
2
s 0 ( s + 3s + 2 )s
s 0 s + 3s + 2
(0 ) + 3(0) + 2
x ( t 0 ) = lim [ sx ( s ) ]
1(s)
1
1- = --x ( t 0 ) = lim --------------------------------= ------------------------------------------= 0
2
2
s ( s + 3s + 2 )s
( ( ) + 3( ) + 2 )
Figure 17.31 Final and initial values theorems
System Model
(differential equations)
Some input
disturbance in
terms of time
Laplace Transform
Transfer
Function
We can figure
out from plots
Experiments
Root-Locus
for stability
Substitute
Output Function
(Laplace form)
Bode and Phase Plots
- Straight line
- Exact plot
Phasor Transform
Output Function
(Laplace terms)
Partial Fraction
Inverse Laplace
Time based
response
equation
17.9 SUMMARY
Transfer and input functions can be converted to the s-domain
Output functions can be calculated using input and transfer functions
Output functions can be converted back to the time domain using partial fractions.
L[5]
b)
L[ e
c)
L [ 5e
d)
L [ 5te
e)
L [ 5t ]
f)
g)
3t
3t
3t
o)
L [ x + 5x + 3x ], x ( 0 ) = 8, x ( 0 ) = 7
p)
d
L ----- sin ( 6t )
dt
d 32
L ----- t
dt
q)
]
0 y dt
r)
s)
L [ 3t ( t 1 ) + e
L [ 4t ]
t)
L [ cos ( 5t ) ]
u)
L[e
2t
v)
L[e
( t 3 )
cos ( 5t ) ]
w)
L [ 5e
cos ( 5t + 1 ) ]
x)
L [ cos ( 7t + 2 ) + e
( t + 1)
L[ 3( t 1 ) + e
i)
L [ 5e
3t
j)
L [ 5e
3t
k)
L [ sin ( 5t ) ]
y)
L [ cos ( 5t + 1 ) ]
l)
L [ sinh ( 3t ) ]
z)
L [ 6e
m)
L [ t sin ( 2t ) ]
n)
d 2 3t
L ----- t e
dt
aa)
u( t 2) ]
h)
5t
3t
u( t 1) ]
2.7t
t3
cos ( 9.2t + 3 ) ]
2. Convert the following functions below from the laplace to time domains using the tables.
a)
b)
c)
d)
1---------s+1
g)
5---------s+1
h)
6--2
s
6---
i)
j)
s
e)
f)
L
L
k)
5-- 4.5s
(1 e
)
s
4 + 3j - ---------------------4 3j --------------------+
s + 1 2j s + 1 + 2j
6- ------------6
--+ 2 4
s
s +9
6
-------------------------
s + 5s + 6
6
---------------------------------2
4s + 20s + 24
s +6
3. Convert the following functions below from the laplace to time domains using partial fractions
and the tables.
1
a)
b)
L [ ]
c)
L [ ]
d)
L [ ]
e)
f)
L
L
6
---------------2
s + 5s
2
9s + 6s + 3
---------------------------------------3
s + 5s + 4s + 6
g)
h)
i)
j)
k)
l)
L
L
s---------------------------------------+ 9s + 6s + 3
3
2
s + 5s + 4s + 6
9s + 4-----------------3
(s + 3)
9s + 4 ----------------------3
s (s + 3)
2
s------------------------+ 2s + 12
s + 3s + 2
2
s------------------------+ 3s + 52
6s + 6
2
s------------------------+ 2s + 32
s + 2s + 1
4. Convert the output function below Y(s) to the time domain Y(t) using the tables.
5
12 - --------------------3 - ---------------------3 Y ( s ) = --- + ------------+
+
s s 2 + 4 s + 2 3j s + 2 + 3j
5x + 6x + 2x = 5F
a)
y + 8y = 3x
b)
y y + 5x = 0
c)
6. Given the transfer function, G(s), determine the time response output Y(t) to a step input X(t).
4
Y( s)
G ( s ) = ----------- = ----------X ( t ) = 20 When t >= 0
s+2
X(s)
7. Given the following input functions and transfer functions, find the response in time.
Transfer Function
Input
a)
x ( s )s+2
m
---------= --------------------------------- ----
F(s)
( s + 3 ) ( s + 4 ) N
F ( t ) = 5N
b)
x ( s )s+2
m
---------= --------------------------------- ----
F(s)
(s + 3)(s + 4) N
x ( t ) = 5m
b)
L[ e
c)
d 3
d
L ----- y + 2 ----- y + y =
dt
dt
d)
e)
+ 5t ( u ( t 2 ) u ( t ) ) ] =
1 + j - --------------------1j ---------------------+
=
s + 3 + 4j s + 3 4j
1
3
- =
s + ----------- + ---------------------------2
s + 2 s + 4s + 40
where at t=0
y0 = 1
y 0' = 2
y 0'' = 3
y 0''' = 4
+ 40 + 20 + 2 = 4
5--s
l)
0.5- ---------0.5---------
s3 s+3
b)
1---------s+3
m)
4
16s 12s 16
--------------------- + -------------------= ----------------------32
3
2
2
2
(s + 4)
(s + 4)
(s + 4)
c)
5---------s+3
5 ------------------
n)
2s -----------------3
(s + 3)
o)
( s x 7s 8 ) + 5 ( sx 7 ) + 3x
p)
6s
----------------2
s + 36
q)
r)
y-s
72
1 ------ 18
------ + ---------5
4
s+5
s
s
d)
e)
f)
g)
h)
i)
j)
k)
(s + 3)
5
---2
s
8
---3
s
s
----------------2
s + 25
3
e ---- 3--- + ---------2 s
s+1
s
5(s + 3)
-----------------------------2
2
(s + 3) + 5
2.5 1 - ---------------------2.5 1 --------------------+
s + 3 5j s + 3 + 5j
s( ) +
= ---------------------------2
s + 6s + 34
5 ---------------2
s)
y)
z)
2s
t)
e------- e--------
s
s
u)
e
------------s+2
v)
e ---------s+1
w)
5 - e------- e-----------------+
s
s
s+3
4 2s
2s
s + 25
x)
2s
0.416s 6.37 e
cos ( 2 ) s sin ( 2 ) 7 e
-------------------------------------------- + ----------- = ------------------------------------- + ----------2
2
s1
s1
s + 49
s + 49
s---------------------------------cos 1 5 sin 1
2
s + 25
3 3 - ------------------------------3 3 s( ) +
-----------------------------+
= ---------------------------------------2
s + 2.7 9.2j s + 2.7 + 9.2j
s + 5.4s + 91.93
2.
a) e
b) 5e
g)
5 5u ( t 4.5 )
h)
2 ( 5 )e
i)
t + 2 sin ( 3t )
j)
6e
k)
1.5e
c) 6t
d) 3t
e) e
f)
3t
+ 2e
( 1 )t
3
cos 2t + atan ---
4
4t
2t
6e
2t
3t
1.5e
3t
6 sin ( 6t )
3.
a)
3t
+ 2e
4t
g)
b)
h)
c)
i)
d)
j)
e)
1.2 1.2e
f)
8.34e
4.4t
0.3t
l)
( t ) + 2te
4.
y ( t ) = 5 + 6 sin ( 2t ) + 2 ( 3 )e
2t
5.
a)
b)
c)
2t
( t)
k) --------- + 0.834 cos ( t + 0.927 )
6
5t
+ 2 ( 0.99 )e
(t) e
5
--x- = ---------------------------2
F
5s + 6s + 2
y-3
= ----------x
s+8
y-5
= ----------x
s1
cos ( 3t 0 )
6.
y ( t ) = 40 40e
2t
7.
a)
b)
( 5 ( t ) + 30 5e
2t
)N
8.
a)
L [ 5e
4 t
cos ( 3t + 2 ) ] = L [ 2 A e
cos ( t + ) ]
= 4
A = 2.5
= 3
= 2
A
A
1.040 + 2.273j 1.040 2.273j
----------------------+ -------------------------------------- = --------------------------------------- + --------------------------------------s + j
s + + j
s + 4 3j
s + 4 + 3j
b)
L[e
2t
+ 5t ( u ( t 2 ) u ( t ) ) ] = L [ e
2t
] + L [ 5tu ( t 2 ) ] L [ 5tu ( t ) ]
1
5
1
5
= ----------- + 5L [ tu ( t 2 ) ] ----2 = ----------- + 5L [ ( t 2 )u ( t 2 ) + 2u ( t 2 ) ] ----2
s+2
s+2
s
s
5
1
= ----------- + 5L [ ( t 2 )u ( t 2 ) ] + 10L [ u ( t 2 ) ] ---2s+2
s
2s
2s
5
1
= ----------- + 5e L [ t ] + 10e L [ 1 ] ----2
s+2
s
2s
2s
5
1
5e
10e
= ----------- + -----------+ --------------- ----2
2
s
s+2
s
s
c)
d
3
2
1
0
---- y = s y + 1s + 2s + 3s
dt
d-
1
0
---= s y+s 1
dt y
d 3
d
3
2
L ----- y + 2 ----- y + y = ( s y + 1s + 2s + 3 ) + ( sy + 1 ) + ( y )
dt
dt
3
= y ( s + s + 1 ) + ( s + 2s + 4 )
d)
L
complex conjugate
1 + j - --------------------1j A
A
1
---------------------+
= L ----------------------- + -------------------------------------s + 3 + 4j s + 3 4j
s + j
s + + j
2
A =
= 2Ae
e)
1 + 1 = 1.141
t
cos ( t + ) = 2.282e
3t
= 3
= 4
cos 4t ---
4
13
1
3
- = L [ s ] + L ---------+ L ---------------------------s + ----------- + ---------------------------2
2
s + 2 s + 4s + 40
s+2
s + 4s + 40
d
2t
3
d 2t
6
- = ---= ----- ( t ) + e + L ----------------------------- ( t ) + e + 0.5L -----------------------------2
2
dt
dt
( s + 2 ) + 36
( s + 2 ) + 36
d
2t
2t
= ----- ( t ) + e + 0.5e sin ( 6t )
dt
9.
a)
s + 4s + 4s + 4
---------------------------------------3
s + 4s
1
3
2
s + 4s + 4s + 4
3
( s + 4s )
s + 4s
4s + 4
2
4s + 4A Bs + Cs ( A + B ) + s ( C ) + ( 4A )
= 1 + ---------------= 1 + --- + --------------= 1 + ----------------------------------------------------------3
2
3
s s +4
s + 4s
s + 4s
1
3s = 1 + --- + ------------s s2 + 4
= ( t ) + 1 + 3 cos ( 2t )
A = 1
C = 0
B = 3
b)
s +4
5
A = lim -------------------------------------------------- = --6
s 1 ( s + 2 ) ( s + 3 ) ( s + 4 )
2
s +4
8
B = lim -------------------------------------------------- = -----2
s 2 ( s + 1 ) ( s + 3 ) ( s + 4 )
2
s +4
13
C = lim -------------------------------------------------- = -----2
s 3 ( s + 1 ) ( s + 2 ) ( s + 4 )
2
s +4
20
D = lim -------------------------------------------------- = -----6
s 4 ( s + 1 ) ( s + 2 ) ( s + 3 )
5--- t
2t 13 3t 10 4t
e 4e + ------ e ------ e
3
6
2
10.
6 39.50t
( t ) = 66 10 e
3.216e
0.1383t
+ 1.216e
0.3368t
+ 2.00
t
L f ---
a
= aF ( as )
d)
b)
1 s
L [ f ( at ) ] = --- F ---
a a
e)
c)
L[ e
at
f(t )] = F(s + a)
f)
lim f ( t ) = lim sF ( s )
s0
lim f ( t ) = lim sF ( s )
s0
d
L [ tf ( t ) ] = ----- F ( s )
dt
2. The applied force F is the input to the system, and the output is the displacement x.
a) find the transfer function.
K1 = 500 N/m
K2 = 1000 N/m
M = 10kg
F
b) What is the steady state response for an applied force F(t) = 10cos(t + 1) N ?
c) Give the transfer function if x is the input.
d) Find x(t), given F(t) = 10N for t >= 0 seconds using Laplace methods.
3. The following differential equation is supplied, with initial conditions.
y + y + 7y = F
y( 0) = 1
y(0) = 0
F ( t ) = 10
t>0
D
--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
D ( D + 2 )--x- = ----------------------------2
F
( D + 200 )
F = 5 sin ( 62.82t )
17.13 REFERENCES
Irwin, J.D., and Graf, E.R., Industrial Noise and Vibration Control, Prentice Hall Publishers,
1979.
Close, C.M. and Frederick, D.K., Modeling and Analysis of Dynamic Systems, second edition,
John Wiley and Sons, Inc., 1995.
Topics:
Objectives:
18.1 INTRODUCTION
Control variable
INPUT
(e.g. gas)
vdesired
verror
+
OUTPUT
(e.g. velocity)
SYSTEM
(e.g. a car)
Driver or
cruise control
gas
car
vactual
The control system is in the box and could be a driver or a cruise control
(this type is known as a feedback control system)
Figure 18.1
Figure 18.2
Some of the things we do naturally (like the rules above) can be done with mathematics
de
u = K c e + K i edt + K d ------
dt
Figure 18.3
The figure below shows a basic PID controller in block diagram form.
proportional
PID Controller
Kp ( e )
integral
e
Figure 18.4
Ki ( e )
derivative
d
K d ----- e
dt
+V
+
u
amp
+
+
-V
motor
e.g.
dv error
gas = K c v error + K i v error dt + K d ----------------
dt
Rules 2 & 3
(general difference)
Kc
Ki
Kd
Rule 4
(Immediate error)
Rule 1
(Long term error)
For a PI Controller
gas = K c v error + K i v error dt
For a P Controller
gas = K c v error
For a PD Controller
dv error
gas = K c v error + K d ----------------
dt
dt
1
= --s
K
gas
- = K c + -----i + K d s
L -----------v error
s
dx
------ sx
dt
xdt
x
= -s
d x
dv
F = Ma = M -------2- = M -----dt
dt
F
d--- = M ---v
dt
v
1
L --- = ------F
Ms
gas
10
1-----Ms
vactual
gas
10-----Ms
vactual
5. Finally, knowing the error is verror, and we can control gas (the control variable),
we can select a control system.
verror
gas
Controller
gas
Ki
L ------------- = K c + ----- + K d s
v error
s
vdesired
verror
+
Ki
K c + ----- + K d s
s
gas
10
1-----Ms
vactual
vdesired
100
step
vdesired(t) = 100 for t >= 0 sec
ramp
vdesired(t) = 50t for t >= 0 sec
t(sec)
0
Input type
Time function
STEP
f ( t ) = Au ( t )
RAMP
f ( t ) = Atu ( t )
SINUSOID
f ( t ) = A sin ( t )u ( t )
PULSE
f ( t ) = A ( u ( t ) u ( t t1 ) )
etc......
Laplace function
A
f ( s ) = --s
A
f ( s ) = ----2
s
2
A
f ( s ) = ----------------2
2
s +
f( s) =
Therefore to continue the car example, lets assume the input below,
v desired ( t ) = 100
t 0 sec
100
v desired ( s ) = L [ v desired ( t ) ] = --------s
Next, lets use the input, and transfer function to find the output of the system.
v actual
v actual = ----------------- v desired
v desired
v actual
2
s ( K d ) + s ( K c ) + K i 100
= ------------------------------------------------------------- ---------
M
s 2 ---- s
10- + K d + s ( K c ) + K i
At this point we have the output function, but not in terms of time yet. To do this
we break up the function into partial fractions, and then find inverse Laplace
transforms for each.
2
s + s + 0.1
2
-
v actual = 10 ------------------------------------------------2
s ( s ( 1.01 ) + s + 0.1 )
ax + bx + c = 0
b b 4ac
x = -------------------------------------2a
1 1 4 ( 1.01 ) ( 0.1 )
x = ------------------------------------------------------------ = 0.113, 0.877
2 ( 1.01 )
10
s + s + 0.1
v actual = ---------- ----------------------------------------------------
1.01 s ( s + 0.113 )s + 0.877
A
B
C
v actual = --- + ---------------------- + ---------------------s s + 0.114 s + 0.795
10
s + s + 0.1
A = lim s ---------- ---------------------------------------------------------
1.01 s ( s + 0.113 ) ( s + 0.877 )
s0
10
0.1
= ---------- -------------------------------------
1.01 ( 0.113 ) ( 0.877 )
A = 99.9
2
B =
lim
s 0.113
10 - -------------------------------------------------------s + s + 0.1
--------- ( s + 0.113 )
1.01 s ( s + 0.113 ) ( s + 0.877 )
C =
lim
s 0.877
2
10
s + s + 0.1
--------- -------------------------------------------------------- ( s + 0.877 )
1.01 s ( s + 0.113 ) ( s + 0.877 )
2
f(s )
A
--s
At
A
---2
s
Ae
A ----------s+
A
----------------2
2
s +
A sin ( t )
n t
n 1
--------------------------------------2
2
s + 2 n s + n
sin ( n t 1 )
for ( < 1 )
etc.
To finish the problem, we simply convert each term of the partial fraction back to
the time domain.
99.9
0.264
1.16
v actual = ---------- + ---------------------- ---------------------s
s + 0.113 s + 0.877
0.113t
1.16e
0.877t
Type
Transfer Function
Proportional (P)
Gc = K
Proportional-Integral (PI)
1
G c = K 1 + -----
Proportional-Derivative (PD)
G c = K ( 1 + s )
Proportional-Integral-Derivative (PID)
1
G c = K 1 + ----- + s
Lead
Lag
Lead-Lag
1 + s
G c = K ------------------
1 + s
>1
1 + s
G c = K ------------------
1 + s
>1
1 + 1 s 1 + 2 s
G c = K --------------------- ---------------------
1 + 1 s 1 + 2 s
>1
1 > 2
Note: this procedure can take some time to do, but the results are very important
when designing a control system.
Consider the example below,
1--s
K
1
K
G ( s ) = ---s
H(s) = 1
First, we must develop a transfer function for the entire control system.
K
----
s
G
(
s
)
K
G S ( s ) = --------------------------------- = --------------------------- = -----------1 + G ( s )H ( s )
s+K
K
1 + ---- ( 1 )
s
Next, we use the characteristic equation of the denominator to find the roots as
the value of K varies. These can then be plotted on a complex plane. Note:
the value of gain K is normally found from 0 to +infinity.
s+K = 0
K
0
1
2
3
etc..
root
j
K
K = 0
Note: because all of the roots for all values of K are real negative this system will
always be stable, and it will always tend to have a damped response. The large the
value of K, the more stable the system becomes.
Consider the previous example, the transfer function for the whole system was
found, but then only the denominator was used to determine stability. So in general we do
not need to find the transfer function for the whole system.
H(s) = 1
First, find the characteristic equation,. and an equation for the roots,
K
- ( 1 ) = 0
1 + ------------------------2
s + 3s + 2
2
s + 3s + 2 + K = 0
3 9 4(2 + K)
1 4K
roots = ------------------------------------------------ = 1.5 -------------------2
2
Next, find values for the roots and plot the values,
K
root
0
1
2
3
0
-1
-2
-3
equation.
( s + z 1 ) ( s + z 2 ) ( s + z m )
1 + G ( s )H ( s ) = 1 + K ---------------------------------------------------------------- = 0
( s + p1 ) ( s + p 2 ) ( s + pn )
2. count the number of poles and zeros. The difference (n-m) will indicate how
many root loci lines end at infinity (used later).
3. plot the root loci that lie on the real axis. Points will be on a root locus line if
they have an odd number of poles and zeros to the right. Draw these lines in.
4. determine the asymptotes for the loci that go to infinity using the formula below.
Next, determine where the asymptotes intersect the real axis using the second
formula. Finally, draw the asymptotes on the graph.
180 ( 2k + 1 )
( k ) = ----------------------------------nm
k [ 0, n m 1 ]
( p1 + p2 + + pn ) ( z 1 + z2 + + z m )
= -------------------------------------------------------------------------------------------nm
5. the breakaway and breakin points are found next. Breakaway points exist
between two poles on the real axis. Breakin points exist between zeros. to calculate these the following polynomial must be solved. The resulting roots are
the breakin/breakout points.
A = ( s + p 1 ) ( s + p 2 ) ( s + p n )
B = ( s + z 1 ) ( s + z 2 ) ( s + z m )
d-
d
---A B A ----- B = 0
ds
ds
6. Find the points where the loci lines intersect the imaginary axis. To do this substitute the phasor for the laplace variable, and solve for the frequencies. Plot the
asymptotic curves to pass through the imaginary axis at this point.
( j + z 1 ) ( j + z 2 ) ( j + z m )
1 + K --------------------------------------------------------------------------- = 0
( j + p 1 ) ( j + p2 ) ( j + p n )
Consider the example in the previous section,
H(s) = 1
(
s
+
1
)
(s + 2)
s + 3s + 2
Step 2: (find loci ending at infinity)
m = 0
n = 2
nm = 2
-2
-1
k I [ 0, 1 ]
180 ( 2 ( 0 ) + 1 )
( 0 ) = ------------------------------------- = 90
2
180 ( 2 ( 1 ) + 1 )
( 1 ) = ------------------------------------- = 270
2
-2
-1
( 0 )( 1 2)
= ----------------------------- = 0
2
asymptotes
A = 1
B = s + 3s + 2
d---A = 0
ds
d---B = 2s + 3
ds
d
d
A ----- B B ----- A = 0
ds
ds
2
1 ( 2s + 3 ) ( s + 3s + 2 ) ( 0 ) = 0
-2
-1
-1.5
2s + 3 = 0
s = 1.5
Note: because the loci do not intersect the imaginary axis, we know the system will be
stable, so step 6 is not necessary, but we it will be done for illustrative purposes.
s + 3s + 2 + K = 0
2
( j ) + 3 ( j ) + 2 + K = 0
2
+ 3j + 2 + K = 0
2
+ ( 3j ) + ( 2 K ) = 0
2
3j ( 3j ) 4 ( 2 K )
3j 9 + 8 + 4K
3j 4K 1
= -------------------------------------------------------------- = ---------------------------------------------- = ------------------------------2
2
2
In this case the frequency has an imaginary value. This means that there will be no
frequency that will intercept the imaginary axis.
K( s + 5 )
G ( s )H ( s ) = --------------------------------2
s ( s + 4s + 8 )
18.5 SUMMARY
x( t) = 5
t 0 sec
8. Draw a detailed root locus diagram for the transfer function below. Be careful to specify angles
of departure, ranges for breakout/breakin points, and gains and frequency at stability limits.
2
2K ( s + 0.5 ) ( s + 2s + 2 )G ( s ) = ----------------------------------------------------------3
s (s + 1)(s + 2)
10. Draw the root locus diagram for the transfer function below,
2
11. Draw the root locus diagram for the transfer function below,
K(s + 1)(s + 2)
G ( s ) = ------------------------------------3
s
12. The block diagram below is for a motor position control system. The system has a proportional controller with a variable gain K.
d
Vd +
Ve
-
Vs
100---------s+2
1--s
Va
2
ans.
200K
----------------------------------2
s + 2s + 200K
b) Draw the Root-Locus diagram for the system (as K varies). Use either the
approximate or exact techniques.
2 4 4 ( 200K )
roots = ------------------------------------------------ = 1 1 200K
2
Im
K
r
o
o
K=0.005
ts
ans.
-2
Re
-1
c) Select a K value that will result in an overall damping coefficient of 1. State if the Root-Locus
diagram shows that the system is stable for the chosen K.
(ans.
K = 0.005
n = 1
s + 2s + 200K = s + 2n s + n
OR
5--2
s
V' ( 0 ) = 2
s ( 3K d ) + s ( 3K p ) + ( 3K i )
--Y- = -----------------------------------------------------------------------------------2
X
s ( 4 + 3K d ) + s ( 36 + 3K p ) + ( 3K i )
equation as shown.
e = dh
error
2
Ki
c-2s + 1 + s
= K p + ----- + K d s = -------------------------s
s
e
combine the transfer functions
2
PID controller
h--10 = -----------2
c
s +s
elevator
10
10 ( s + 1 )
2s + 1 + s - -----------10 (s + 1)
c-- h--- = h--- = ------------------------= ------------------- ------------------- = ---------------------2
2
e c
s s(s + 1 )
e
s
s
s +s
h 10 ( s + 1 )
eliminate e
----------= ---------------------2
dh
s
10
(
s
+
1 )
h = ---------------------(d h)
2
s
10 ( s + 1 )
10 ( s + 1 )
= ---------------------(d)
h 1 + ---------------------2
2
s
s
10 ( s + 1 )
---------------------
2
h--s
10s + 10 system transfer function
= -------------------------------- = ------------------------------2
10 ( s + 1 )
d
s
+
10s
+
10
1 + ---------------------
2
s
a) Find the response of the final equation to a step input. The system starts at rest
on the ground floor, and the input (desired height) changes to 20 as a step input.
b) Write find the damping coefficient and natural frequency of the results in part
a).
c) verify the solution using the initial and final value theorems.
(ans. a) h
10s + 10 --- = ------------------------------2
d
s + 10s + 10
20
d ( s ) = -----s
d ( t ) = 20u ( t )
B
C
10s + 10 - 20
------ = A
--- + -------------------------------- + -------------------------------h = ------------------------------2
s s + 5 3.873j s + 5 + 3.873j
s + 10s + 10 s
200s + 200 -
A = lim ------------------------------= 20
2
s 0 s + 10s + 10
B =
200s + 200 -
---------------------------------------= 2.5 22.6j
s 5 + 3.873j s ( s + 5 + 3.873j )
lim
C = 2.5 + 22.6j
h ( t ) = 20 + L
h ( t ) = 20 + 2 ( 22.73 )e
b)
5t
25
1 -----2- n
n
25 2
2
15 = 1 -----2- n = n 25
c)
= 4.602
= 5
= 3.873
5
= -----n
1 n =
n =
35 = 5.916
5
= ---------- = 0.845
35
10s + 10 - 20
------ = lim 10s
-------- 20 = 0
h ( 0 ) = lim s ------------------------------s s 2 + 10s + 10 s
s s2
10s + 10 - 20
10 )20------ = (---------------h ( ) = lim s ------------------------------= 20
2
s
10
s0
s + 10s + 10
5 = n
3.873 =
A =
1--------0.25
Vd
e
+
Ki
K p + ----- + sK d
s
Motor
Vs
1--s
Potentiometer
Va
1--------0.25
b) Given the transfer function below, select values for Kp, Ki and Kd that will result in a second
order response that has a damping coefficient of 0.125 and a natural frequency of 10rad/s.
(Hint: eliminate Ki).
2
c) The function below has a step input of magnitude 1.0. Find the output as a function of time
using numerical methods. Give the results in a table OR graph.
a
s ( 0.9 ) + ( 4 ) ----- = ----------------------------------2
d
s + 2.5s + 100
d) The function below has a step input of magnitude 1. Find the output as a function of time by
integrating the differential equation (i.e., using the homogeneous and particular solutions).
a
s ( 0.9 ) + ( 4 )
----- = -----------------------------------2
d
s + 2.5s + 100
e) The function below has a step input of magnitude 1. Find the output as a function of time using
Laplace transforms.
2. Select a controller transfer function, Gc, that will reduce the system to a first order system with
a time constant of 0.5s, as shown below.
Controller
d
1--------0.25
Vd
Motor
Vs
Gc
+
-
Potentiometer
Va
1--------0.25
a
1
----- = ----------d
s+2
1--s
convolution - 19.1
19. CONVOLUTION
Topics:
Objectives:
19.1 INTRODUCTION
Can be used to find normal input reponses for linear systems
It is most useful for finding an output response for a system given an arbitrary
input function
It is also the basis for other methods that come later for system analysis.
convolution - 19.2
Unit impulse
P(t)
1
Area = --- = 1
1--
Area =
( t ) dt
= 1
t
0
u(t < 0) = 0
u(t 0) = 1
d---u( t) = ( t )
dt
If we look at an input signal (force here) we can break it into very small segments
in time. As the time becomes small we can approximate it with a set of impulses.
convolution - 19.3
F(t)
t 0
An impulse
t
Figure 19.1
F(t)
x( t)
g ( t )F ( t ) dt
t
Figure 19.2
If we add all of the impulse responses together we will get a total system
response. This operation is called convolution.
convolution - 19.4
F( t)
x(t)
sum of responses
t
g ( t )F ( ) d
0
t
impulse responses
c(t) =
g ( t )r ( ) d
0
Figure 19.3
A set of pulses for a system gives summed responses to give the output
Consider the unit impulse of a system with the given differential equation. Note:
This method is only valid for trivial differential equations with only one homogeneous
term. The preferred method is shown later.
convolution - 19.5
The following first order differential equation has an input F. The input can be
replaced with a unit impulse function.
x + 0.5x = 2F
X + 0.5X = 2 ( t )
The homogeneous solution can be found.
X + 0.5X = 0
guess,
Xh = e
At
At
X h = Ae
A + 0.5 = 0
X h = Ce
0.5 t
X + 0.5X = 2 ( t )
guess,
Xp = A
Xp = 0
0 + 0.5A = 2 ( 0 )
Xp = A = 0
The initial condition caused by the impulse function found, assuming a zero initial condition.
1-
1-
--- --- dt X0 + 0.5 ( 0 ) = 2 dt
X0 = 2
The initial condition caused by the impulse function found, assuming a zero initial condition.
X ( t ) = Ce
0.5 t
X ( 0 ) = 2 = Ce
X ( t ) = 2e
0.5 t
Note that the derivation of the unit impulse function assumed zero initial conditions, so the process of convolution must also assume systems start at rest and undeflected.
19.4 CONVOLUTION
The following example shows the use of the convolution integral to find a num-
convolution - 19.6
ber of responses.
The unit impulse response to a step input can be calculated using the convolution integral.
t
x(t) =
X ( t )F ( ) d
0
x ( t ) = 2e
0.5t e
0.5t
( 2e
0.5 ( t )
) ( 1 ) d = 2e
0.5t
0
0.5 ( 0 )
0.5
d = 2e
0.5t e
0.5
---------0.5
The unit impulse response to a unit ramp input can be calculated using the convolution
integral.
t
x(t) =
X ( t )F ( ) d
x(t) =
2e
t
0.5 ( t )
d = 2e
0.5t
0.5
The unit impulse response to a sinusoidal input can be calculated using the convolution
integral.
t
x(t) =
X ( t )F ( ) d
0
x(t) =
2e
t
0.5 ( t )
sin ( ) d = 2e
0.5t
0.5
sin ( ) d
convolution - 19.7
n1
x ( t ) = h X ( t ( i + 0.5 )h )F ( ( i + 0.5 )h )
i=0
where,
t
n = --- = number of steps
h
h = step size (s)
X ( t ) = the unit impulse response function
F ( t ) = the input function
This can be applied to the previous example for a unit step input to find the system position at 10 seconds, with a 2 second time step.
convolution - 19.8
n1
x ( t ) = h 2e
0.5 ( t ( i + 0.5 )h )
F ( ( i + 0.5 )h )
i=0
10
n = ------ = 5
2
51
x ( 10 ) = 2 2e
0.5 ( 10 2 ( i + 0.5 ) )
(1)
i=0
4
x ( 10 ) = 4 e
4.5 + i
i=0
x ( 10 ) = 4 ( e
4.5 + 0
x ( 10 ) = 4 ( e
4.5
+e
+e
4.5 + 1
3.5
+e
2.5
+e
4.5 + 2
+e
1.5
+e
+e
4.5 + 3
0.5
+e
4.5 + 4
0.5 ( 10 )
) = 4 ( 1 0.006738 ) = 3.973048
convolution - 19.9
convolution - 19.10
c(t) =
g ( t )r ( ) d
0
C ( s ) = G ( s )R ( s )
Figure 19.4
19.7 SUMMARY
Topics:
Objectives:
20.1 INTRODUCTION
- state equations can be converted to transfer functions. The derivation follows.
x = Ax + Bu
y = Cx + Du
In the s-domain
sX X 0 = AX + BU
X ( sI A ) = BU + X 0
1
X = ( sI A ) BU + ( sI A ) X 0
Y = CX + DU
1
Y = C ( ( sI A ) BU + ( sI A ) X 0 ) + DU
1
Y = ( C ( sI A ) B + D )U + C ( sI A ) X 0
Assuming the system starts at rest,
1
Y = ( C ( sI A ) B + D )U
Y1
--= ( C ( sI A ) B + D )
U
x = Ax + Bu
y = Cx + Du
Map one state vector to another one
x = Tz
This can be used to calculate new coefficient matrices
1
z = T ( Ax + Bu )
1
z = T ( ATz + Bu )
1
1
z = T ATz + T Bu
1
1
z = T ATz + T Bu
A = T AT
B = T B
C = CT
D = D
z = Az + Bu
The output equation becomes
y = Cx + Du
y = CTz + Du
y = CTz + Du
y = Cz + Du
The equivalent set of state equations is,
z = Az + Bu
A = T AT
B = T B
y = Cz + Du
C = CT
D = D
G = C ( sI A ) B + D
1
G = C ( sI A ) B + D
This can be shown by applying the transformation matrix
1
G = CTT ( sI A ) TT B + TT D
1
G = ( CT ) ( T ( sI A ) T ) ( T B ) + D
1
G = ( CT ) ( ( sI A ) ) ( T B ) + D
1
G = C ( ( sI A ) )B + D = G
The transfer function form can be put into a matrix form. In this case the denominator is the characteristic equation.
The transfer function can be said to be equivalant to the determinants of the matrix form.
( sI A ) B
C
D
sI A )D ( B )C------------------------------------ = (--------------------------------------------= ( sI A )D + BC ( sI A )
sI A
sI A
( sI A ) B
C
D
1
poles
G = C ( sI A ) B + D = ------------------------------------ = -------------sI A
zeros
sI A = characteristic equation = homogeneous
The free (homogeneous) response of a system can be used to find the state transition matrix.
The homogeneous equation can be written in the s-domain, and then converted to time.
X h = sI A X0
1
x h ( t ) = L [ sI A X 0 ]
xh ( t ) = L
-I + A
--- + A
------ + A
------ + X 0
s s s2 s3
At
xh ( t ) = e x0
At
= transition matrix
At
1 22
1 33
= I + At + ----- A t + ----- A t +
2!
3!
The forced response (particular) response of the system can be found using convolution,
The homogeneous equation can be written in the s-domain, and then converted to time.
x = Ax + Bu
t
x(t) = e x( 0) + e
At
A(t )
Bu ( ) d
y = Cx + Du
y ( t ) = Ce x ( 0 ) + Ce
At
A(t )
Bu ( ) d + Du ( t )
initial
response
impulse
response
Given a second order differential equation, the state matrix can be found,
F = Mx
x = v
0
d- x
---= 01 x + 1 F
A = 0 1
F
dt
---v
00 v
0 0
v = ----M
M
This can be used to find the inverse matrix,
s 1
1s --1--- 1
------2 2
0
s
s s2
1
s
1
0
0
1
s
1
s
------------( sI A ) = s
= 2
=
=
=
0 1
0
s
0 0
0 s
s 0
1
---- ---0 --2 2
s
s s
The function of time can be found assuming an initial position of 10 and velocity of 5.
1
At
1
1
= L [ ( sI A ) ] = 1 t
0 1
At
x h ( t ) = x = e 10 = 1 t 10 = 10 + 5t
v
5
0 1 5
5
Given a second order differential equation, the state matrix can be found,
F = Mx
If a matrix is diagonizable, the diagonal matrix can be found with the following
technique. This can be used for more advanced analysis techniques to create diagonal (and
T AT =
A = TT
=
n
At
t 1
= Te T
e 1
=
e
n t
Av = v
The Eigenvectors are then combined into a single matrix, this is the transition matrix.
example,
+ 1 2
3 4
2
= ( + 1 ) ( 4 ) ( 6 ) = 3 + 2 = ( 2 ) ( 1 )
1 = 2
2 = 1
2 u 21 = 3u 11
v1 =
u 11
u 21
= 2
3
u 22 = u 12
v2 =
u 12
u 22
= 1
1
Av i = i v i
u11
u 21
,A
u 12
u 22
= 2 2 ,1 1
3
1
1 2 2 , 1 2 1 = 4 , 1
3 4 3 3 4 1
6 1
4, 1 = 4, 1
6 1
6 1
verified
If there are repeated Eigenvalues in the system the Jordan Form can be used.
J =
1 1 0
0 2 1
0 0 3
0 0 0 n 1 1
0 0 0
Jordan block
2 1 1
1 2 1
1 1 2
= ( + 3) ( 0)
2 T 11 + T 21 + T 31
2 1 1 T 11
= 1 2 1 T 21 = T 11 2T 21 + T 31
1 1 2 T 31
T 11 + T 21 2T 31
T 11 = 2T 11 + T 21 + T 31
T 21 = T 11 + 2T 21 + T 31
T 31 = T 11 + T 21 + 2T 31
T 12
1 T 22
T 32
2 T 12 + T 22 + T 32
2 1 1 T 12
= 1 2 1 T 22 = T 12 2T 22 + T 32
1 1 2 T 32
T 12 + T 22 2T 32
T 12 = 2T 12 + T 22 + T 32
T 22 = T 12 + 2T 22 + T 32
T 32 = T 12 + T 22 + 2T 32
T 13
0 T 23
T 33
2 T 13 + T 23 + T 33
2 1 1 T 13
= 1 2 1 T 23 = T 13 2T 23 + T 33
1 1 2 T 33
T 13 + T 23 2T 33
4T 13 = 2T 13 + T 23 + T 33
4T 23 = T 13 + 2T 23 + T 33
4T 33 = T 13 + T 23 + 2T 33
FIX
T 21 =
T 21 =
T 31 =
Given the following Eigen values and vectors the free response is,
1 = 2
2 = 1
n
xh ( t ) =
i=1
vi e
i t
v1 = 2
3
v2 = 1
1
2t
2t
1t
= 2 e + 1 e = 2e + e
2t
t
3
1
3e + e
zeros of state space functions can be found using the state matrices.
= 0
x = v 5F
v = 3x 4v + 5F
d- x
---= 0 1 x + 5 F
dt v
3 4 v
5
A =
0 1
3 4
x = 1 0 x + 0F
v
B = 5
5
C = 1 0
D = 0
0 1 5
s0
0 1
3 4
5
1 0
s0
= 0
3 s0 + 4 5 = 0
1 0
0
s 0 ( 0 ) ( 1 ) ( 4.5 ) + 5 ( ( s 0 + 4 ) ) = 0
24.5 = 5s 0
s 0 = 4.9
There is one zero at -4.9 rad/sec. The order of the polynomial determines the
number of zeros. If the polynomial has no s terms, then there are no zeros.
20.2 OBSERVABILITY
- a system is observable iff the system state x(t) can be found by observing the
input u and output y over a period of time from x(t) to x(t+h).
- If an input is to be observable it must be detectable in the output. For example
consider the following state equations.
d- x
---= a b x + e u
dt v
c d v
f
y = g h x + j u
v
k
if the parameters c, d or f were not in the final expression for y, the second state equation would be said to be not observable. This will happen when an output parameter
is zero in the C matrix and an equation is decoupled in the A matrix, i.e. the same
variable column is zero.
For example,
d- x
---= 1 0 x + 3 u
dt v
0 a v
b
y = 5 0 x + 6 u
v
G = C ( sI A ) B + D = 5 0 s 0 1 0
0 s
0 a
1
G = 5 0 s1 0
0 s+a
G = 5 0
1---------s1
0
3 +
6
b
s+a 0
0 s 1 3
3 +
-------------------------------=
+ 6
6
5 0 ( s 1 ) ( s + a -)
b
b
3
----------15
+ 6 = 5 0 s 1 + 6 = ----------- + 6
s1
b ---------s+a
3
1- b
---------s+a
Note: the variables a and b both dissapeared because this system is not
observable.
This often happens when a system has elements that are decoupled, or when a
pole and zero cancel each other.
Observability can be verified formally for an LTI system with the following relationship.
The system is unobservable if the equation below is true for any state,
At *
1
0
*
x = 0 , 1 ,
0
0
t0
Ce x = 0
0
0
One example,
A = 1 2
3 4
C = 5 6
At *
e e
Ce x = 5 6
3t
2t
4t
1 0
1 0
,
= ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t ) ,
0 1
0 1
At *
Ce x = ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t ) 1 , ( 5e t + 6e 3t ) ( 5e 2t + 6e 4t ) 0
0
1
At *
3t
2t
4t
Ce x = { ( 5e + 6e ), ( 5e + 6e ) }
Both values are non-zero, so both states are observable.
One example,
A = 1 0
0 2
C = 3 0
At *
t
Ce x = 3 0 e 0 1 , 0 = 3e t 0 1 , 0 = { 3e , 0 }
2t
0 1
0 e 0 1
Another theorem for testing observability is given below. If any of the states satisfies the equation it is unobservable.
C
CA
CA
x = 0
n1
rank ( M o ) = rank
C
CA
CA
= n
n1
20.3 CONTROLLABILITY
- a system is controllable iff there is an input u(t) that will cause the system to go
from any initial state to any final state in a finite time.
- stabilizable if it is controllable or if the uncontrollable nodes are stable.
- If an input is to be observable it must be detectable in the output. For example
consider the following state equations.
d- x
---= a b x + e u
dt v
c d v
f
y = g h x + j u
v
k
if the parameters c, d or f were not in the final expression for y, the second state equation would be said to be not uncontrollable. This will happen when an output
parameter is zero in the C matrix and an equation is decoupled in the A matrix, i.e.
the same variable column is zero.
For example,
d- x
---= 1 0 x + 2 u
dt v
a b v
0
y = 3 4 x + 5 u
v
G = C ( sI A ) B + D = 3 4 s 0 1 0
0 s
a b
1
G = 3 4 s1 0
0 s+b
G = 3 4
1---------s1
0
2 +
5
0
s+B 0
0 s1 2
2 +
+ 5
5 = 3 4 -------------------------------(s 1)( s + b) 0
0
6
2 +
2 +
3
4
-+5
5 = ----------- ----------5 = ---------s
1
s
1
s
+
b
1- 0
0
---------s+b
Note: the variables a and b both dissapeared because this system is not controllable.
(x ) e B = 0
t0
An example is,
A = 1 2
3 4
B = 5
6
T
1 0
,
e
0 1
1 2t
3 4
e 1t e 2t 5
5 =
,
1 0 0 1
6
e 3t e 4t 6
5e 1t + 6e 2t
1t
2t
3t
4t
= { 5e + 6e , 5e + 6e }
1 0, 0 1
5e 3t + 6e 4t
The results are both non-zero, so both states are controllable.
An example is,
A = 1 0
0 2
B = 0
3
T
1 0
,
e
0 1
1 0t
0 2
e 1t 0 0
0 =
,
1 0 0 1
3
0 e 2t 3
0
2t
1 0 , 0 1 2t = { 0, 3e }
3e
The first state has a zero, so it in not controllable.
* T
( x ) B AB A n 1 B
rank ( M c ) = rank B AB A n 1 B = n
20.4 OBSERVERS
Observers estimate system state variables when not all of the variables are
directly observable.
Observers use a limited set of system states that are available to identify other
system states that are not observable.
The separation principle ensures that the observer cannot effect the stability of
the system it is observing.
20.5 SUMMARY
20.9 BIBLIOGRAPHY
How, J, "16.31 Feedback Control Course Notes", MIT Opencourseware website.
Topics:
Objectives:
21.1 INTRODUCTION
There are a few terms to be defined,
identification - the process of determining a state space (or other) model of
a process.
parameter identification - used to estimate model parameter values.
estimatior - this tries to determine state variable values given available
data.
observer - a special type of estimator that uses available input and output
data to estimate a system state given unknown initial conditions. This is
simalar to running a system simulation in parallel with the real system.
compensator - a feedback control system designed to drive a system to an
arbitrary input.
regulator - a feedback control system designed to drive a system to a given
operating point. A special case of a compensator with a zero input.
x = Ax + Bu
y = Cx + 0u
The Eigenvalues of A give the system poles
u = r Kx
r = setpoint
u = control output/plant input
K = proportional gain values
x = Ax + B ( r Kx )
x = ( A BK )x + Br
x = A c x + Br
y = Cx
x =
y = 1 x
0
0 1 x+ 0 u
2 3
4
0
4
0 1 0
2 3 4
= 0 4
4 12
The rank is 2, and the order of the matrix A is 2, so the system is controllable.
Check for the stability of the system without a controller,
sI A =
s 0 0 1
0 s
2 3
s 1
2 s 3
=
2
sI A = s ( s 3 ) 2 = s 3s 2 = ( s ) ( s )
Therefore, the uncompensated system is.......
Select an input vector with variables,
u = Kx = k 1 k 2 x
Ac =
0
1
0 1 0
0 1 0 0 =
k1 k2 =
4k 1 4k 2
( 2 4k 1 ) ( 3 4k 2 )
2 3
4
2 3
sI A c =
0
1
s 0
( 2 4k 1 ) ( 3 4k 2 )
0 s
s
1
s + 2 + 4k 1 s + 3 + 4k 2
sI A c = s ( s + 3 + 4k 2 ) + ( s + 2 + 4k 1 ) = s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 )
The system is clearly second order, so the second order methods may be used to select
parameters. However for variety, why dont we select a system that has poles at,
p = 1 + 2j, 1 2j
2
s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = ( s + 1 + 2j ) ( s + 1 2j )
2
s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = s + s ( 1 + 2j + 1 2j ) + ( ( 1 + 2j ) ( 1 2j ) )
s + s ( 4 + 4k 2 ) + ( 2 + 4k 1 ) = s + s ( 2 ) + ( 5 )
4 + 4k 2 = 2
k 2 = 0.5
2 + 4k 1 = 5
k 1 = 0.75
u
B
+
-
x
1--s
+
A
K
Figure 21.1
The value of K in the previous example could also be picked using Ackermanns
formula.
K = 0 0 1 Mc d ( A )
where,
d ( s ) = the desired response
The state equations and the desired response of the system are,
x =
y = 1 x
0
0 1 x+ 0 u
2 3
4
2
d ( s ) = s + s ( 2 ) + ( 5 )
The Markov (?) parameter can be calculated first.
Mc =
0
4
0 1 0
2 3 4
K = 0 1 0 4
4 12
= 0 4
4 12
0 1
2 3
+ 0 1 ( 2 ) + ( 5 )I
2 3
12 4
4 0
K = 0 1 ---------------------- 2 3 + 0 2 + 5 0
0 16 6 7
4 6
0 5
K = 0 1 0.75 0.25 3 1 = 0.25 0 3 1 = 0.75 0.25
0.25 0
2 6
2 6
NOTE: DOESNT MATCH, TRACK DOWN ALGEBRA PROB
21.3 OBSERVERS
Observers are used to estimate the next system state. They can also be used to
predict future state values.
As an example, an observer might be used for a targeting system that is tracking a
moving object. The estimator can be used to direct a missle to an estimated location so that
the missle and target arrive at the same time.
An observer is shown in block diagram form
Actual Plant
(and assumed state equations)
U
d---X = AX + BU
dt
Y = CX
Observer
d-
---X = AX + BU + K( Y Y )
dt
Y = CX
Figure 21.2
Aside: For an open loop estimator a simulation is run, while the system is running. The
simulation values are compared to the actual system responses, and the model is
adjusted. It is assumed that the initial value of the state is unknown. If A is stable then
the estimation will eventually converge (this can be checked using the Eigenvalues).
However, there is no guarantee that it will converge quickly.
y = Cx
CI
d-
---x = Ax
dt
At
x ( t ) = e x ( 0 )
A closed loop estimatior is shown below. It uses a learning matrix L to adjust the
convergence rate. A larger L value will result in faster convergence, but the exponent must
y = Cx
The estimator error
y = y y = Cx
The estimator
d-
---x = Ax ( t ) + Bu ( t ) + L ( y ( t ) y ( t ) )
dt
y ( t ) = Cx ( t )
where,
L = a gain matrix
The convergence rate,
d-
---x = ( A LC )x
dt
( A LC )t
x ( t ) = e
x(0)
C
CA
CA
= n
n1
Consider the example below, for a second order system. The learning rate values
are selected to determine how quickly the model converges.
y = 1 0
x1
x2
a1 1
a2 0
sI ( A LC ) =
l1
l2
1 0 =
s + a1 + l1 1
a2 + l2
a1 l1 1
a2 l2 0
= s ( s + a 1 + l 1 ) ( 1 ) ( a 2 + l2 )
= s ( 1 ) + s ( a1 + l 1 ) + ( a2 + l2 )
L = K e = e ( s )M o
0
1
Consider the state equations for a mass spring damper system (y is the position).
0
d
x
----= K
dt v
---------s
M
0
A = K
---------s
M
1
0
x +
K d
1- F
------------ v
M
M
1
Kd
--------M
B =
y = 1 0 x + 0 F
v
0
1
----M
C = 1 0
D = 0
0
1 0 Ks
--------M
1
K d
--------M
= rank
d
x
----= K
dt
---------s
v
M
1
Kd
--------M
0
x +
1- F
---v
M
1 0
= rank
0 1
1 0
= 2
0 1
y = Cx
y = 1 0 x
v
The open loop estimator will normally ignore the response to initial conditions
and will have long term errors resulting from modeling errors and random disturbances in
the system.
y ( t ) = Cx ( t )
d
x
----= K
dt
---------s
v
M
y ( t ) = 1 0 x
v
1
Kd
--------M
0
l
x +
F + 1 y
1
----l2
v
M
W = KX
X = C
---K
Y- = M 1 ---Y---W
W
M = C
---K
Y- = MX
---W
d- Y
d
1 Y
-------- = ----- MX = M ( AX + BU ) = MAX + MBU = MAM ---- + MBU
dt
dt W
W
The matrices can be partitioned to isolate the unmeasurable state variables. Note that the
Eigenvalues of Hw must be relatively small.
MAM
Hy Hw
MB =
Bw
d-
---W = H y Y + H w W + Bw U
dt
system state.
u = Kx = K y
CL
+
d-
- ---x
dt
1--s
+
d---x
dt
u
K
A
system
B
estimator and regulator
1--s
x
C
dt x
LC A BK LC x
1
T = I 0 = T
I I
A cl
The estimators and regulators can be designed separately and then combined.
This is the separation principle.
The simplified block diagram for the system is,
d-
---x
dt
x
1--s
d---x
dt
1--s
BK
+
A BK LC
A
LC
The compensator,
The system can also be set up for a compensator by assuming the setpoint input is
not zero,
1
0
x +
Kd
1- F
------------ v
M
M
1
Kd
--------M
0
B = 1
----M
y = 1 0 x + 0 F
v
C = 1 0
D = 0
The regulator will be designed to have a response time of 0.5s and damped frequency of 2
rad/s.
1
s = ------- + j ( 2 ) = 2 2j
0.5
i ( A BK ) = 2 2j
0
1
0
1
0
= i K K 1 k1 k2 = i K k K k
s
1
d
2
---- ------------------- ---------s --------d
- --------------------M
M
M
M M
0
1
s 0
Ks k1 K d k2
0 s
--------------------- ---------------------M
M
s
1
K s + k 1
K d + k 2
---------------- s + ---------------- M
M
K d + k 2 K s + k 1
K d + k 2 K s + k 1
- + ----------------- = s s + ----------------- + ----------------= s s + ----------------
M
M
M M
Kd + k2
Ks + k1
Kd + k2 2
------------------ ------------------ 4 -----------------
M
M
M
s = ------------------------------------------------------------------------------------------------2
Kd + k2
-----------------M
k 2 = 4M Kd
2 = --------------------2
K d + k 2 2
K s + k 1
----------------- 4 --------------- M
M
2 = -------------------------------------------------------------2
2
16M = ( K d + k 2 ) 4 ( K s + k 1 )
2
4 ( K s + k 1 ) = ( K d + 4M K d ) 16M
2
k 1 = 4M 4M K s
K = 4M 2 4M K 4M K
s
d
N
K s = 10 ---m
Ns
K d = 1 -----m
K = 4M 2 4M K 4M K
s
d
K = 4 ( 1 ) 2 4 ( 1 ) 10 4 ( 1 ) ( 1 )
K = 10 3
The desired response time for the system will be a time response of approximately
0.2sec.
x
1
G ( s ) = --- = ------------------------------------2
F
Ms + K d s + K s
0
A cl = A BK = K
---------s
M
1
0
1
0
=
Kd
Ks k1 Kd k2
1- k 1 k 2
-------------------------------- ---------------------M
M
M
M
e ( s ) = sI Acl =
0
1
s 0
Ks k1 Kd k2
0 s
-------------------- ---------------------M
M
s
1
K s + k 1 Kd + k2
----------------- ------------------ + s
M
M
K d + k 2 K s k 1
2
- + -------------------- = ( s + 0.2 ) 2
= s + s ---------------- M M
Ks k 1
------------------- = ( 0.2 ) 2
M
Kd + k2
------------------ = 2 ( 0.2 )
M
k 1 = 10.04
k 2 = 0.6
0
1
= K K + 0.2
---------s --------d
M M
0
1
= K K
---------s --------d M M
1 0
0
1 0 Ks
--------M
0
+ 2 ( 0.2 ) K
---------s
M
1
K d
--------M
0
1
1
2
K d + 0.2 I
--------M
Kd
--------M
Ks
--------- + 0.04
M
Kd
--------- + 0.4
M
2
Kd
------2
Ks
--------- + 0.04
M
Kd
--------- + 0.4
M
2
Kd
------2
0.4K d
K s K d 0.4K s K s
------------ ------------- --------- +
-------------- + 0.04
2
M
M M
M
M
1 1
0
1
0
0.4
0.04
0
+
+
0.4K
0.4K
2
K s Kd K s Kd
-----------------s ----------------d0 0.04
M
M
------------ --------- + ------
2
M M2
M
Ks
--------M
K s K d 0.4K s K s
0.4K d
------------ ------------- --------- +
-------------- + 0.04
2
M
M M
M
M
1 0
Kd
--------- + 0.4
M
2
Kd
------2
K s
0.4K d
--------- +
-------------- + 0.04
M M
M
0.6
9.36
1 0 0
1 1 1
0
1
1 0
1 1
0
1
Ac =
0
0 1
0.6
10 3
1 0
1
---- 10 1
9.36
M
Ac =
0 1 0 0 0.6 0 = 0.6 1
10 1
10 3
9.36 0
9.36 4
Bc = L =
0.6
9.36
C c = K = 10 3
The compensator transfer function can then be written,
1
G c ( s ) = C c ( sI A c ) B c
= 10 3 s 0 0.6 1
0 s
9.36 4
0.6 =
s 0.6 1
10 3
9.36
9.36 s 4
0.6
9.36
s 4 9.36
s 0.6 1
0.6
0.6
10 3
10 3
9.36 s 4 9.36
1 s 0.6 9.36
= ------------------------------------------------------------------------------ = --------------------------------------------------------------------------( s 0.6 ) ( s 4 ) ( 1 ) ( 9.36 )
s 0.6 1
9.36 s 4
s4 1
0.6
0.6s 6.96
10 3
10 3
9.36 s 0.6 9.36
9.36s
- = ------------------------------------------------------= -----------------------------------------------------------------------2
2
1s 4.6s 6.96
1s 4.6s 6.96
22.08s + 69.6
= -----------------------------------2
s 4.6s 6.96
#define
PERIOD
0.001
// period of the interrupt
interrupt_loop(){ // this routine is called at a regular interval PERIOD
}
double analog_in(){
}
d-
---x
dt
x
1
--s
d---x
dt
1
--s
BK
+
A BK LC
A
LC
21.6 LQR
- Linear Quadratic Regulator (LQR) allows the design of a regulating system that
tends to zero, but also considers the control effort.
- The basic function for evaluating the controller is defined below. The integral is
evaluated to reduce the value. Ideally the state value will go to zero (it is a regulator), and
the plant input will be minimized. It is assumed that D=0. Here a new state z is defined.
It should reflect the system state values that should be regulated to zero.
J LQR =
[ (y(t) )
y ( t ) + r ( u ( t ) ) ] dt
0
T
( y ( t ) ) y ( t ) = state cost
2
( u ( t ) ) = control cost
r = control penalty
z = Cz x
J LQR =
[ (x(t) )
C z C z x ( t ) + r ( u ( t ) ) ] dt
- the design process involves adjusting the r value to obtain the desired result.
-
x = Ax + Bu
y = Cx
The equivalent transfer function can be written, where a(s) is the characteristic eqn.
b(s)
b(s)
1
C ( sI A ) B = ---------- = ----------------a(s)
sI A
This is followed by.......
UNKNOWN STUFF
The desired charactersitic equation
n can be used to find the controller gain values.
( s pi )
sI A + BK LQR =
i=1
1
( s ) = a ( s )a ( s ) + r b ( s )b ( s )
The poles are then found using Ackermanns method.
( u ( t ) ) = control cost
r = control penalty
J LQR =
[ (x(t) )
C Cx ( t ) + r ( u ( t ) ) ] dt
C Cx ( t ) + r ( u ( t ) ) ] dt
J LQR =
[ (x(t) )
0
Z = N+P
where,
P = poles of G(s)Gc(s) in the right hand plane
Z = closed loop poles in the right hand plane
N = number of clockwise loops in the Nyquist Diagram about -1
Z = 0, N = P
For stability
Note: Larger loops indicate a more stable system. Tight loops are very sensitive to
parameter variations.
21.8.1 Stability
The stability function is defined using the Nyquist plot.
The boundedness of the system can be tested by ensuring that the sensitivity always
remains less than some given value gamma for all frequencies. Note: this requires
that D must be smaller than gamma.
S ( j ) <
1
C ( jI A ) B + D <
The Hamiltonian matrix is defined below. It does not have purely complex Eigenvalues.
2
A + B( I D D) D C
H =
1 T
B( I D D) B
1 T
C ( I + D ( I D D ) D )C A C D ( I D D ) B
Or, if D=0,
T
H =
BB
---------2
C C A
This matrix can then be tested, if there are no purely complex eigenvalues, the system is
bounded.
T
I H =
0
0
BB --------2
C C A
BB A --------2
BB T C T C
2
T
= ( 1 ) + ( A + A ) + --------------------
2
etc....
C C +A
XBB X
T
T
- = 0
A X + XA + C C + ----------------2
Parameter
Estimation
Regulator
Design
Regulator
Plant
Reference
Model
Adaptation
Algorithm
Controller
Plant
Reference
Model
Adaptation
Algorithm
+
Controller
Plant
+
Parallel MRAC
Reference
Model
Adaptation
Algorithm
Adjustable
Plant
Series MRAC
Adaptation
Algorithm
Reference
Model
Adjustable
Plant
The reference model is developed from the fundamentals of the system. It should be
controllable and stable.
d---X = Am Xm + Bm r
dt m
Parameters can be adpated using,
e = X m Xp
d---X = A p X p + Bp r
dt p
Where,
A p, B p = Adjustable parameter matrices
A m, B m = Reference model matrices
d---X = Ap Xp + Bp ( r + ua )
dt p
d---e = Am Xm + Bm r Ap Xp Bp ( r + ua ) = Am ( e + Xp ) + Bm r Ap Xp Bp ( r + ua )
dt
d---e = e ( Am ) + Xp ( A m A p ) + r ( B m B p ) + u a ( B p )
dt
d---e = e ( Am ) + f
dt
f = Xp ( Am Ap ) + r ( Bm Bp ) + ua ( Bp )
V = e Qe
d
dd T
T
T T
T
T
---V = ----- e Qe + e Q ----- e = ( e ( A m ) + f )Qe + e Q ( e ( A m ) + f )
dt
dt
dt
dT T
T
---V = e ( A m Q + QA m )e + 2e Qf
dt
dT
T
---V = e We + 2e Qf
dt
W = ( A m Q + QAm )
d---x = Ax + Bu + B w w
dt
y = Cx + v
where,
B w = process noise weighting matrix
w = process noise matrix
v = sensor noise
N(s)
1
G yw ( s ) = C ( sI A ) B w = ----------D( s)
R
D ( s )D ( s ) ------w N ( s )N ( s ) = 0
Rv
where,
21.11 SUMMARY
Topics:
Objectives:
22.1 INTRODUCTION
A simple example
d---= AX out + BX in
X
dt out
d---Y
= CX out + DXin
dt out
T
T
xi xi 2
x i 2x i 1 + x i 2
----------------------------------------- = cx i 1 + d -------------------- + fF
2
T
T
xi xi 2
x i 2x i 1 + x i 2
a ( 0 ) + b ( 0 ) + c ( x i 1 ) + d --------------------- + e ( 0 ) + f ( F ) = ---------------------------------------2
T
T
xi 3 xi 5
x i 3 2x i 4 + x i 5
a ( 0 ) + b ( 0 ) + c ( x i 4 ) + d --------------------------- + e ( 0 ) + f ( F ) = ----------------------------------------------2
T
T
x i 6 2x i 7 + x i 8
xi 6 xi 7
a ( 0 ) + b ( 0 ) + c ( x i 7 ) + d --------------------------- + e ( 0 ) + f ( F ) = ----------------------------------------------2
T
T
xi 7
xi xi 2
--------------------T
xi 3 xi 5
-------------------------T
xi 6 xi 8
-------------------------T
xi 1
xi 4
xi 7
xi 1
xi 4
xi xi 2
--------------------0
F 0
T
xi 3 xi 5
--------------------------0
F 0
T
a
xi 6 xi 8
b
--------------------------0
F 0
T
c =
x i 2x i 1 + x i 2
xi xi 2
d
---------------------------------------2
--------------------- 0 F
T
e
T
x i 3 2x i 4 + x i 5
f
xi 3 xi 5
-------------------------------------------------------------------------- 0 F
2
T
T
xi 6 xi 7
x i 6 2x i 7 + x i 8
--------------------------- 0 F
-----------------------------------------------T
2
T
Some sample data points can be calculated for F=1 and M=1,
1 2
1F 2
x = --- at = --- ----- t
T = 1
F = 1
2
2M
x(0) = 0
xi 8 = 0
x ( 1 ) = 0.5
x i 7 = 0.5
x(2) = 2
xi 6 = 2
x ( 3 ) = 4.5
x i 5 = 4.5
x(4) = 8
xi 4 = 8
x ( 5 ) = 12.5
x i 3 = 12.5
x ( 6 ) = 18
x i 2 = 18
x ( 7 ) = 24.5
x i 1 = 24.5
x ( 8 ) = 32
x i = 32
M = 1
The values can be substituted into the matrix, and the matrix solved
32 18
24.5 -----------------1
12.5 4.5
8 -----------------------1
2---------- 00.5
1
0
32 18
24.5 ------------------ 0
1
12.5 4.5
8 ------------------------ 0
1
2 0.5
0.5 ---------------- 0
1
24.5 14 0 0 1
a
8 8 0 0 1
b
c = 0.5 2 0 0 1
0 0 24.5 14 0
d
0 0 8 8 0
e
0 0 0.5 1.5 0
f
0
0
0
1
1
1
32
18----------------0
1
12.5
4.5----------------------0 a
1
2---------- 0b
0
1
c =
32
2 ( 24.5 ) + 18-----------------------------------------d
2
1
1
e
12.5
2 ( 8 ) + 4.5---------------------------------------1 f
2
1
2 2 ( 0.5 ) + 0
1
---------------------------------2
1
14
8
2
1
1
1
a
0
b
1
c = 0
d
0
e
0
f
1
Substituting these values back into the state equation results in the following
expression, which matches the exact state equation.
x
d- x i
---= a b i + e F
dt v
c d vi
f
i
x
= 10 i + 0 0 F
0 0 vi
0 0 1
0
xi
state equation
output equation
xi 1
xi 4
a
b
x
c = i7
d
0
e
f
0
0
xi xi 2
--------------------T
xi 3 xi 5
--------------------------T
xi 6 xi 8
--------------------------T
0
0
0
F 0
F 0
F 0
xi xi 2
x i 1 --------------------- 0 F
T
xi 3 xi 5
x i 4 --------------------------- 0 F
T
xi 6 xi 7
x i 7 --------------------------- 0 F
T
xi xi 2
--------------------T
xi 3 xi 5
--------------------------T
xi 6 xi 8
--------------------------T
x i 2x i 1 + x i 2
---------------------------------------2
T
x i 3 2x i 4 + x i 5
-----------------------------------------------2
T
x i 6 2x i 7 + x i 8
-----------------------------------------------2
T
This can then be used as a model to calculate a transfer function as shown below
using the previous example.
A = a b
c d
B = e
f
x
= 10 i + 0 0 F
C = 1 0
D = 0 0
0
0
0
0
1
v
0
0 0
0 0
i
these can be used to calculate a SISO system transfer function, and its inverse.
xi
G = C ( sI A ) B + D = 1 0 s 1 0 a b
0 0 01
c d
e + 0 0
f
0 0
G = 1 0 s a b
0 0 c s d
sd b
c sa
e = 1 0 -------------------------------------------- e
f
(
s
a
)
(
s
d
)
+
bc
f
0 0
sd b
0 0
( s d )e + bf
e = --------------------------------------------------------------------------------------G =
( s a ) ( s d ) + bc f
( s a ) ( s d ) + bc
(-------------------------------------------s a ) ( s d ) + bc1
G =
( s d )e + bf
+
+
xd
controller
1
plant
xa
The identified coefficients for the controller are substituted into the inverse transfer func.
0
a
b
1
c = 0
0
d
0
e
1
f
( s a ) ( s d ) + bc
(s 0)(s 0) + 1(0 )
2
= --------------------------------------------- = -------------------------------------------------- = s
( s d )e + bf
( s 0 )0 + 1 ( 1 )
The identified coefficients for the controller are substituted into the inverse transfer func.
d- 2
F
F
---x = ----- = -- dt
M
1
1
--x- = ---2
F
s
A state diagram for the system may then be written using a proportional controller,
s
+
+
xd
P
1
1--2
s
xa
--1P
xd
1--2
s
1
2
s--P
+
+
xd
-
xa
P--2
s
1
s--P
+
+
P
-------------2
s +P
xd
xd
P
-------------2
s +P
s---+1
P
xa
xa
xd
+ P -------------P
s-------------
P 2
s +P
xa
xd
xa
Note: This suggests that the result is an ideal controller, however in practice there are
errors, such as sensor noise, imperfect actuators, and changing system conditions
such as mass. This also assumes that the force actuator can generate any requested
force.
xa
22.2 SUMMARY
electromagnetics - 23.1
Topics:
Objectives:
23.1 INTRODUCTION
Magnetic fields and forces are extremely useful. The fields can allow energy storage, or transmit forces.
23.2.1 Induction
Magnetic fields pass through space.
electromagnetics - 23.2
I =
Hdl
where,
I = current flowing along a line (A)
H = magnetic field intensity (A/m)
l = A perpendicular path around the current flow (m)
For example, at a fixed radius (r) around a wire,
2
I =
Hr d
I = 2Hr
I
H = --------2r
Flux density can be calculated for low H values. As the value climbs the relationship becomes non-linear.
B = H = r 0 H
where,
B = Flux density (Wb/m*2 or T)
= permeability of material
7 Henry
0 = permeability of free space = 4 10 ---------------m
r = relative permeability
Permeability,
electromagnetics - 23.3
B
= ---H
7 H
0 = 410 ---m
= r 0
where,
A
H = Magnetic field intensity ----
m
0 = permeability of free space
r = relative permeability of a material
= permeability of a material
Permeability is approximately linear for smaller electric fields, but with larger
magnetic fields the materials saturate and the value of B reaches a maximum value.
B
1.5
linear
saturation
4000
Figure 23.1
electromagnetics - 23.4
1.5
silicon sheet steel
cast steel
1.0
B (T)
0.5
cast iron
0
0
500
H (At/m)
Figure 23.2
1000
electromagnetics - 23.5
I
B = --------2r
where,
Wb
B = Flux density -------2- or Tesla
m
I = Current in the conductor (A)
r = radial distance from the conductor
B dA
where,
= Flux density ( Wb )
A = Cross section area perpendicular to flux
When a material is used out of the saturation region the permeabilities may be
written as reluctances,
L
R = ------A
where,
R = reluctance of a magnetic path
L = length of a magnetic path
A = cross section area of a magnetic path
electromagnetics - 23.6
Ni
F
= ------------- = --R
l -
----- A
Example,
lc
R c = ----------c Ac
I
lg
lc
Faradays law,
lg
R g = -----------g Ag
Ni
= -----------------Rc + R g
electromagnetics - 23.7
d
e = N -----
dt
For a coil
where,
e = the potential voltage across the coil
N = the number of turns in the coil
Field energy,
B
B
W = ------ V = ------ Al
2
2
B
F = ------ A
2
The basic property of induction is that it will (in the presence of a magnetic field)
convert a changing current flowing in a conductor to a force or convert a force to a current
flow from a change in the current or the path.
electromagnetics - 23.8
F = ( I B )L
F = ( L B )I
I, L
F
where,
L = conductor length
F = force (N)
The FBD/schematic equivalent is,
F
M
Figure 23.3
electromagnetics - 23.9
e m = ( v B )L
where,
e m = electromotive force (V)
= magnetic flux (Wb - webers)
v = velocity of conductor
v
e
Figure 23.4
+
-
Hysteresis
electromagnetics - 23.10
5
4
a
axis of
rotation
Figure 23.5
electromagnetics - 23.11
For wire 3,
r cos ( t )
P 3 = r sin ( t )
b b
--- to --2 2
V3 =
r sin ( t )
r cos ( t )
0
0
B = B
0
de m3 = ( V B )dL
e m3 =
b
--2
b--2
e m3 =
b
--2
b--2
r sin ( t )
0
r cos ( t ) B dr
0
0
0
dr
0
Br sin ( t )
e m3
r
= B ---- sin ( t )
2
b
--2
b
--2
b
b 2
= B sin ( t ) --- --- = 0
2
2
Figure 23.6
electromagnetics - 23.12
0
B = B
0
de m2 = ( V B )dL
a
e m2 =
b
--- sin ( t )
2
0
b-- B dl
cos ( t )
2
0
0
0
0
e m2 =
b
B --- cos ( t )
2
b
dl = aB --- cos ( t )
2
e m4 = e m2
Figure 23.7
As can be seen in the previous equation, as the loop is rotated a voltage will be
generated (a generator), or a given voltage will cause the loop to rotate (motor).
In this arrangement we have to change the polarity on the coil every 180 deg of
rotation. If we didnt do this the loop the torque on the loop would reverse for half the
motion. The result would be that the motor would swing back and forth, but not rotate
fully. To make the torque push consistently in the same direction we need to reverse the
electromagnetics - 23.13
applied voltage for half the cycle. The device that does this is called a commutator. It is
basically a split ring with brushes.
e m = aBb cos ( t )
Real motors also have more than a single winding (loop of wire). To add this into
the equation we only need to multiply by the number of loops in the winding.
e m = NaBb cos ( t )
As with most devices the motor is coupled. This means that one change, say in
torque/force will change the velocity and hence the voltage. But a change in voltage will
also change the current in the windings, and hence the force, etc.
Consider a motor that is braked with a constant friction load of Tf.
F w = ( I B )L = IBa
b
T w = 2 F --- = Fb = IBab
= T w T f = J
IBab T f = J
Figure 23.8
We still need to relate the voltage and current on the motor. The equivalent circuit
for a motor shows the related components.
electromagnetics - 23.14
IA
RA
LA
+
em
eA
+
-
where,
I A, e A = voltage and current applied to the armature (motor supply)
R A, L A = equivalent resistance and inductance of windings
d
= e A I A R A L A ----- I A e m = 0
dt
Figure 23.9
Practice problem,
electromagnetics - 23.15
Ks
M
Kd
R
x
I
electromagnetics - 23.16
23.4 SUMMARY
fluids - 24.1
Topics:
Objectives:
24.1 SUMMARY
Fluids are a popular method for transmitting power (hydraulics). Basically, by
applying a pressure at one point, we can induce flow through a pipe/value/orifice.
w = flow rate
p = pressure
fluids - 24.2
24.2.1 Resistance
If fluid flows freely, we say it is without resistance. In reality, every fluid flow
experiences some resistance. Even a simple pipe has resistance. Of similar interest is the
resistance of a value.
w = K p
where,
w = flow rate through the pipe
p = pressure difference across valve/pipe
K = a constant specific to the pipe/valve/orifice
NOTE: In this case the relationship between pressure drop and flow are non-linear.
We have two choices if we want to analyze this system.
1. We can do a non-linear analysis (e.g. integration)
2. We can approximate the equation with a linear equation. This is only good for
operation near the chosen valve position. As the flow rate changes significantly
the accuracy of the equation will decrease.
w
p R ( w w )
( p )
R --------------w
w
w
R = 2 -----2K
w
p
( p )
Figure 24.1
Resistance may also result from valves. Valves usually restrict flow by reducing
an area for fluid to flow through.
A simple form of valve is a sliding plunger. The valve below is called a two way
fluids - 24.3
Fluid in
to cylinder
Fluid out
Figure 24.2
Four way valve allow fluid force to be applied in both directions of a cylinder
motion.
fluids - 24.4
to
reservoir
fluid in
fluid in
to
reservoir
no motion
advance
Figure 24.3
retract
24.2.2 Capacitance
Fluids are often stored in reservoirs or tanks. In a tank we have little pressure
near the top, but at the bottom the mass of the fluid above creates a hydrostatic pressure.
Other factors also affect the pressure, such as the shape of the tank, or whether or not the
top of the tank is open.
To calculate the pressure we need to integrate over the height of the fluid.
fluids - 24.5
A
Fp
h
F p + Fg
Fg
Fp + Fg
F p + Fg
Fg
P = -----A
F g = V = gAh
gAh
P = -------------- = gh
A
Figure 24.4
Consider a tank as a capacitor. As fluid is added the height of the fluid rises, and
the hydrostatic pressure increases. Hence we can pump fluid into a tank to store energy,
and letting fluid out recovers the energy. A very common application of this principle is a
municipal water tower. Water is pumped into these tanks. As consumers draw water
through the system these tanks provide pressure to the system. When designing these
tanks we should be careful to keep the cross section constant (e.g. a cylinder). If the cross
section varies then the fluid pressure will not drop at a linear rate and you wont be able to
use linear analysis techniques (eg., Laplace).
The mathematical equations for a constant cross section tank are,
A
C = -----g
d
w = C ----- p
dt
fluids - 24.6
Fluid In
Normally closed
inlet valve
Fluid Out
piston
Normally
Closed Valve
Figure 24.5
A geared hydraulic pump is pictured below. Other types use vanes and pistons.
fluids - 24.7
Figure 24.6
Vane based pumps can be used to create fluid flow. As the pump rotates the vanes
move to keep a good seal with the outer pump wall. The displacement on the advance and
return sides are unequal (aided by the sliding vanes). The relative displacement across the
pump determines the fluid flow.
Fluid in
Fluid out
fluids - 24.8
Figure 24.7
As with the resistance of valves, these are not linear devices. It is essential that
we linearize the devices. To do this we look at the pressure flow curves. (Note: most
motors and engines have this problem)
( p )
K = --------------w
1
w ---- ( p p )
K
( p )
w
w
Figure 24.8
pump
Figure 24.9
valve
tank
fluids - 24.9
fluids - 24.10
24.4 SUMMARY
Topics:
Objectives:
25.1 INTRODUCTION
Energy can be stored and transferred in materials in a number of forms. Thermal
energy (heat) is stored and transmitted through all forms of matter.
25.2.1 Resistance
A universal property of energy is that it constantly strives for equilibrium. This
means that a concentration of energy will tend to dissipate. When material separates
regions of different temperatures it will conduct heat energy at a rate proportional to the
difference. Materials have a measurable conductivity or resistance (note: this is different
than electrical conductivity and resistance.)
1
q = --- ( j i )
R
dq
------ = --1- d
------
dl
R dl
where,
q = heat flow rate from j to i (J/s or watts)
R = thermal resistance
= temperature
d
R = ------A
d = length of thermal conductor
A = cross section area of thermal conductor
= thermal conductivity (W/mK)
Figure 25.1
Thermal resistance
When dealing with thermal resistance there are many parallels to electrical resistance. The flow of heat (current) is proportional to the thermal difference (voltage). If we
have thermal systems in parallel or series they add as normal resistors do.
R1
R2
R1
R2
11
1
----= ------ + -----RT
R1 R2
Figure 25.2
RT = R1 + R2
25.2.2 Capacitance
Heat energy is absorbed at different rates in different materials - this rate is
referred to as thermal capacitance. And as long as the material has no major changes in
structure (i.e., gas-solid or phase transition) this number is relatively constant.
1
= ---- ( q in q out )
C
d
1- dq
------ = -------dt
C dt
where,
C = thermal capacitance
C = M
where,
M = mass of thermal body
= specific heat of material in mass
Figure 25.3
Thermal capacitance
One consideration when dealing with heating capacitance is that the heat will not
instantly disperse throughout the mass. When we want to increase the rate of heat absorption we can use a mixer (with a gas or fluid). A mixer is shown in the figure below. This
mixer is just a rotating propeller that will cause the liquid to circulate.
Figure 25.4
25.2.3 Sources
If we plan to design a thermal system we will need some sort of heat source. One
popular heating source is an electrical heating element.
heating coils are normally made out of some high resistance metal/alloy such as
nichrome (nickel and chrome) or tungsten.
If we run a current through the wire the resistance will generate heat. As these
metals get hotter the resistance rises, hence the temperature is self regulating. If we control
the voltage and current we can control the amount of heat delivered by the coil.
resistance
+
V
temp
P = IV
where,
P = power generated as heat
I = current into coil
V = voltage across coil
Figure 25.5
Consider an insulated sealed chamber with a resistive barrier between sides, and
a heating element in one side.
qi
1, C 1
2, C 2
Given
q 0 = heat flow into the left chamber from the heating element
1, C 1 = initial temperature and heat capacitance of left side
2, C 2 = initial temperature and heat capacitance of right side
Next,
q 1 = heat flow through the center barrier
1
q 1 = --- ( 1 2 )
R
1
d--- 1 = ------ ( q0 q 1 )
C1
dt
1
d--- 2 = ------ ( q 1 )
C2
dt
1
1
s 1 = ------ q 0 --- ( 1 2 )
C1
R
1
s2 = ---------- ( 1 2 )
RC 2
Rq o sRC 1 1 = 1 2
1 = 2 + sRC 2 2
1 ( 1 + sRC 1 ) = 2 + Rq o
2 + Rq o
1 = --------------------1 + sRC 1
Figure 25.6
2 + Rq o
1 = ---------------------- = 2 + sRC 2 2
1 + sRC 1
2 + Rq o
------------------------------ = 1 + sRC 1
2 + sRC 2 2
Rq o
------- 2
1
---------------------- + ---------------------- = 1 + sRC 1
1 + sRC 2 1 + sRC 2
Rq o
2 2
1 + --------- = 1 + s ( RC 1 + RC 2 ) + s R C 1 C 2
2
Rq o
--------- = s ( RC 1 + RC 2 ) + s 2 R 2 C 1 C 2
2
q
----o- = s ( C1 + C 2 ) + s 2 RC 1 C 2
2
qo
--------------- RC 1 C 2
A
B
2 = -------------------------------------- = --- + ---------------------------+
C
s
C
C
1
2
1 + C2
- + s2
s -----------------s + ----------------- RC 1 C 2
RC 1 C 2
qo
qo
------------------------------ RC C
RC 1 C 2
qo
1 2
qo
qo
--------------- ---------------
RC 1 C 2
RC 1 C 2
C 1 + C 2
qo
----------------------------------+
s
s
RC 1 C 2 RC C
RC 1 C 2
1 2
qo
qo
----------------------------------C1 + C2
C1 + C2
2 = ------------------- + ---------------------------s
C 1 + C2
s + -----------------RC 1 C 2
Figure 25.7
qo
1
1
1 1
1
L ( 2 ) = ------------------- L --- L ---------------------------s
C1 + C2
C 1 + C2
s + ------------------RC 1 C 2
C1 + C 2
------------------ t
qo
RC 1 C 2
2 ( t ) = ------------------- 1 e
C1 + C 2
Figure 25.8
25.4 SUMMARY
26. OPTIMIZATION
Topics:
Objectives:
26.1 INTRODUCTION
For simple examples use derivatives and find the maxima/minima
Consider the example below to find the
L =
$
C fence = 20 ---- ( 2w + 2d )
m
where,
C fence = cost to construct the fence ($)
w = width of the pasture (m)
C land
where,
C land = cost for the land
$ -
R = 0.05 ----------wd
2
m yr
wd < 300m
$ -
2
R = 0.01 ----------( wd 300m ) + 60$
2
m yr
wd 300m
where,
R = revenue generated by pasture land
V = R C fence C land
where,
V = total value
Figure 26.1
Figure 26.2
w 1600m
d 1600m
Figure 26.3
Figure 26.4
4
4
w
d
C penaly = ---------------- + ----------------
1600m
1600m
Figure 26.5
Figure 26.6
Search path
Objective Space
Constraints
Figure 26.7
Local searches
Global Search space. In this case the system becomes stuck in a local mimina.
Local Optima
Figure 26.8
Global Optima
Global searches
Global Search space. In this case the system searches all maxim.
Local Optima
Figure 26.9
Global Optima
A Global Search
26.4.3 Simplex
26.5 SUMMARY
Topics:
Objectives:
27.1 INTRODUCTION
Ks
na
Ks
x
ni
Ks
nc
Fi
Ks
nd
Equations can be written to relate the position of node i to the surrounding nodes
Ks ( x i x a )
F i = ( F i x, F i y )
Ks ( x c xi )
= Fix Ks ( xi xa ) + K s ( xc x i ) = 0
x i ( 2K s ) + x a ( K s ) + x c ( K s ) = F i x
+
Ks ( yi yd )
Fx
Fx
= F iy K s ( y i y d ) + K s ( y b y i ) = 0
y i ( 2K s ) + y b ( K s ) + y d ( K s ) = F i y
xi
yi
xa
ya
Fix
Fiy
2K s 0 K s 0 0 0 K s 0 0 0
xb
0 2K s 0 0 0 K s 0 0 0 K s y b
xc
yc
xd
yd
nb
Ks
Fa
Fb
Ks
Ks
y
x
Ks
nd
nc
Fd
Fc
Equations can be written to relate the position of node i to the surrounding nodes
and the applied force. The x and y values are deflections from the unloaded state. The
Ks value is based on the material stiffness and the geometry of the elements.
Ks ( xb xa )
( F a x, F a y )
Ks ( ya y d )
Fx
= F ax + K s ( x b x a ) = 0
F ax = x a ( K s ) + x b ( K s )
+
Fx
= F ay K s ( y a y d ) = 0
F ay = y a ( K s ) + y d ( K s )
Ks ( xb xa )
( Fb x, F by )
Ks ( y b yc )
Fx
= Fbx Ks ( xb xa ) = 0
F bx = x a ( K s ) + x b ( K s )
+
Fx
= Fby Ks ( yb yc ) = 0
F b y = y b ( K s ) + y c ( K s )
Ks ( yb yc )
Ks ( xc xd )
Fx
= F cx Ks ( x c x d ) = 0
F cx = x c ( K s ) + x d ( K s )
( F cx, F cy )
Fx
= F cy + Ks ( y b y c ) = 0
F cy = y b ( K s ) + y c ( K s )
Ks ( ya yd )
Ks ( x c xd )
Fx
= Fdx + Ks ( xc x d ) = 0
F d x = x c ( K s ) + x d ( K s )
( F d x, F d y )
Fx
= F dy + K s ( y a y d ) = 0
F dy = y a ( K s ) + y d ( K s )
Fax
Ks
0 Ks 0
Fay
Ks
0 K s ya
Ks 0
Ks
xb
yb
0 Ks 0
xc
Ks
yc
0 Ks 0
Ks
xd
Ks
yd
Fbx
Fby
Ks
0 Ks 0
Fcx
Ks
Fcy
0 Ks 0
Fdx
Fdy
0 Ks 0
xa
n1
n3
Ks
Ks
#1
Ks
n4
Ks
n7
#3
Ks
Ks
#2
Ks
Ks
n5
#4
Ks
n8
Ks
Ks
n6
Ks
n9
For element #1
Local stiffness matrix
F 1x
Ks
0 K s 0
F 1y
Ks
0 Ks y1
Ks 0
Ks
x2
y2
0 Ks 0
x4
Ks
y4
0 Ks 0
Ks
x5
Ks
y5
F 2x
F 2y
Ks
0 K s 0
F 4x
Ks
F 4y
0 Ks 0
F 5x
F 5y
0 Ks 0
x1
F 1x
F 1y
F 2x
F 2y
F 3x
Ks
0 Ks 0 0 0 0
Ks
0 0 0 0
0 Ks 0 0 0 0 0 0
K s 0
Ks
0 0 0 0
0 0 0 0 0 0 0
0 0 0 0 0 0 0
Ks 0 0 0 K s 0
0
0
0
0
0
0
0
0
0
0 0 0 0
0 0 0 0
0 0 0 Ks
0 Ks 0 0 0
F 5y
F 6x
F 6y
F 3y
F 4x
F 4y
F 5x
F 7x
F 7y
F 8x
F 8y
F 9x
F 9y
0 0 0 0 0 0 0
y1
x2
y2
x3
y3
0 0 0 0 0 0 0 0 0
x
0 0 0 0 0 0 0 0 0 4
0 K s 0 0 0 0 0 0 0 y4
K
0 0 0 0 0 0 0 0 x5
s
0 0 0 Ks 0
Ks
0 Ks 0
0 0 0 0
0 0 0 0 0 0 0 y5
K 0 0 0 0 0 0 x6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
y6
x7
y7
x8
y8
x9
y9
For element #2
local stiffness matrix
F2x
Ks
0 Ks 0
F2y
Ks
0 K s y2
Ks 0
Ks
x3
y3
0 Ks 0
x5
Ks
y5
0 K s 0
Ks
x6
Ks
y6
F3x
F3y
Ks
0 Ks 0
F5x
Ks
F5y
0 Ks 0
F6x
F6y
0 Ks 0
x2
F 1x
F 1y
Ks
0 K s 0
F 2x
Ks
K s 0 2K s 0 K s 0
F 2y
0 Ks
0 0 0 0 0 0 y1
0 K s 0 0 0 0 0 0 x2
0 0 0 0 0 0 0 0 y2
0 0 0 0 0 0 0 0 x3
F 3x
F 3y
0 K s 0
Ks 0
F 4x
0 Ks 0
F 4y
0 Ks
0 Ks 0
0 Ks 0
Ks
F 5y
0 K s 0 2K s 0
F 6x
0 Ks 0
F 6y
F 7x
0 0 0 0 0 0 y5
0 2K s 0 0 0 0 0 0 x 6
0 0 Ks 0 0 0 0 0 y6
0 0 0 K 0 0 0 0 x7
F 7y
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
F 5x
F 8x
F 8y
F 9x
F 9y
0 2K s 0
0 0 0 0 0 0 y3
0 0 0 0 0 0 x4
0 0 0 0 0 0 y4
0 0 0 0 0 0 x5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
y7
x8
y8
x9
y9
For element #3
F4x
Ks
0 Ks 0
F4y
Ks
0 K s y4
Ks 0
Ks
x5
y5
0 Ks 0
x7
Ks
y7
0 Ks 0
Ks
x8
Ks
y8
F5x
F5y
Ks
0 Ks 0
F7x
Ks
F7y
0 Ks 0
F8x
F8y
0 Ks 0
x4
For element #4
F5x
Ks
0 Ks 0
F5y
Ks
0 K s y5
Ks 0
Ks
x6
y6
0 Ks 0
x8
Ks
y8
0 K s 0
Ks
x9
Ks
y9
F6x
F6y
Ks
0 Ks 0
F8x
Ks
F8y
0 Ks 0
F9x
F9y
0 K s 0
x5
If the input forces are known, then the resulting displacements of the nodes can
be calculated by inverting the matrix. Consider the matrix for a single node.
Fax
Ks
0 Ks 0
ya
Ks
0 Ks
Fay
Ks 0
Ks
Fbx
Fby
0 Ks 0
F cx
Ks
F cy
0 Ks 0
Ks
Fdx
Ks
Fdy
xb
yb
Ks
0 Ks 0
xc
Ks
yc
0 Ks 0
xd
yd
0 K s 0
xa
If we assume that node c is fixed in the x and y directions, the matrix can
reflect this by setting the appropriate matrix rows to zero.
xa
Ks
0 Ks 0 0
ya
Ks
xb
Ks 0
0 0 K s
F ax
F ay
Ks 0 0
F bx
F by
yb
0 Ks 0
xd
0 Ks 0
F dx
yd
0 K s 0
0 0 Ks
F dy
The displacements can then be found by selecting values for the coefficents and
solving the matrix. We can select a value of 1000 for the stiffness, and a force of 10 will be
10
0
= 10
0
0
10
10
3
10
xd
yd
10
0 10
10
0 10
0
3
0
10
10
0
0
0
0
0
This method generates very large matrices, easily into the millions.
To reduce the matrix size techniques such as symmetry are commonly used.
Strains can be found by calculating the relative displacements of neighboring
points. These can then be used to calculate the stresses.
More complex elements are commonly used depending upon the stress conditions, part geometry and other factors.
There are a variety of finite element element methods and applications
Computation Fluid Dynamics
Nonlinear deformation
27.4 SUMMARY
27.8 BIBLIOGRAPHY
How, J, "16.31 Feedback Control Course Notes", MIT Opencourseware website.
Topics:
Fuzzy logic theory; sets, rules and solving
Objectives:
To understand fuzzy logic control.
Be able to implement a fuzzy logic controller.
28.1 INTRODUCTION
Fuzzy logic is well suited to implementing control rules that can only be expressed
verbally, or systems that cannot be modelled with linear differential equations. Rules and
membership sets are used to make a decision. A simple verbal rule set is shown in Figure
28.1. These rules concern how fast to fill a bucket, based upon how full it is.
Figure 28.1
The outstanding question is "What does it mean when the bucket is empty, half
full, or full?" And, what is meant by filling the bucket slowly or quickly. We can define
sets that indicate when something is true (1), false (0), or a bit of both (0-1), as shown in
Figure 28.2. Consider the bucket is full set. When the height is 0, the set membership is 0,
so nobody would think the bucket is full. As the height increases more people think the
bucket is full until they all think it is full. There is no definite line stating that the bucket is
full. The other bucket states have similar functions. Notice that the angle function relates
the valve angle to the fill rate. The sets are shifted to the right. In reality this would probably mean that the valve would have to be turned a large angle before flow begins, but after
that it increases quickly.
1
bucket is full
0
1
stop filling
0
1
fill slowly
0
1
bucket is empty
0
height
Figure 28.2
fill quickly
angle
Fuzzy Sets
Now, if we are given a height we can examine the rules, and find output values, as
shown in Figure 28.3. This begins be comparing the bucket height to find the membership
for bucket is full at 0.75, bucket is half full at 1.0 and bucket is empty at 0. Rule 3 is
ignored because the membership was 0. The result for rule 1 is 0.75, so the 0.75 membership value is found on the stop filling and a value of a1 is found for the valve angle. For
rule 2 the result was 1.0, so the fill slowly set is examined to find a value. In this case there
is a range where fill slowly is 1.0, so the center point is chosen to get angle a2. These two
results can then be combined with a weighted average to get
0.75 ( a1 ) + 1.0 ( a2 )
angle = ---------------------------------------------- .
0.75 + 1.0
1
stop filling
0
angle
a1
1
bucket is half full
fill slowly
0
height
a2
angle
1
bucket is empty
0
height
Figure 28.3
fill quickly
angle
vdesired
verror
+
Fuzzy
Vmotor Motor
Imotor
Logic
Power
Controller
Amplifier
vactual
Servo
Motor
verror
1
LN
0
-100 -50 0 50 100
rps
rps
0
-6
rps
0 3
rps/s
-6
-3
0 3
rps/s
0
-6
-3
0 3
rps/s
0
-3
0 3
Figure 28.4
rps
12 18 24
12 18 24
12 18 24
1
rps/s
-6
1
rps
0
1
LP
-3
1
-100 -50 0 50 100
SP
1
-100 -50 0 50 100
ZE
Vmotor
SN
dtverror
12 18 24
12 18 24
0
-6
-3
0 3
rps/s
Consider the case where verror = 30 rps and d/dt verror = 1 rps/s. Rule 1to 6 are calculated in Figure 28.5.
0
-100 -50 0 50 100
rps
0
-6
-3
0 3
rps/s
12 18 24
ANY VALUE
(so ignore)
30rps
17V
(could also
have chosen
some value
above 17V)
1
rps
0
-100-50 0 50 100
rps/s
0
-6 -3 0 3 6
rps/s 0
-6 -3 0 3 6
1rps/s
V
0
14V
1rps/s
30rps
12 18 24
0
-100 -50 0 50 100
30rps
rps
1
rps/s
0
-6
-3
0 3
1rps/s
0
0
12 18 24
0
-100 -50 0 50 100
rps
-6
-3
0 3
1rps/s
30rps
rps/s
12 18 24
0
-100 -50 0 50 100
rps
0
-6
-3
0 3
rps/s
0
0
12 18 24
1rps
30rps
0
-100 -50 0 50 100
rps
30rps
-6
-3
0 3
rps/s
0
0
12 18 24
ANY VALUE
Rule Calculation
The results from the individual rules can be combined using the calculation in Figure 28.6. In this case only two of the rules matched, so only two terms are used, to give a
( Vmotor ) ( membershipi )
i
V motor =
i--------------------------------------------------------------------=1
n
( membershipi )
i=1
28.3 REFERENCES
Li, Y.F., and Lau, C.C., Application of Fuzzy Control for Servo Systems, IEEE International
Conference on Robotics and Automation, Philadelphia, 1988, pp. 1511-1519.
28.4 SUMMARY
Fuzzy rules can be developed verbally to describe a controller.
Fuzzy sets can be developed statistically or by opinion.
Solving fuzzy logic involves finding fuzzy set values and then calculating a value
for each rule. These values for each rule are combined with a weighted average.
verror
1
SN
rps
0
-6
-3
0 3
rps/s
0
-100 -50 0 50 100
rps
SP
Vmotor
1
-100 -50 0 50 100
ZE
dtverror
0
-100 -50 0 50 100
12 18 24
12 18 24
0
-6
-3
0 3
rps/s
1
rps
1
rps/s
0
-6
-3
0 3
0
0
12 18 24
4. Develop a set of fuzzy control rules adjusting the water temperature in a sink.
5. Develop a fuzzy logic control algorithm and implement it in structured text. The fuzzy rule set
below is to be used to control the speed of a motor. When the error (difference between desired
and actual speeds) is large the system will respond faster. When the difference is smaller the
response will be smaller. Calculate the outputs for the system given errors of 5, 20 and 40.
100%
Big Error
error
0%
10
30
100%
Small Error
error
0%
10
30
20
Big Output
error
50
Small Output
5
error
50
if (big error) then (big output)
if (small error) then (small output)
Topics:
Basic laws of motion
Gravity, inertia, springs, dampers, cables and pulleys, drag, friction, FBDs
System analysis techniques
Design case
Objectives:
To be able to develop differential equations that describe translating systems.
I. Introduction
Over the past decade, the artificial intelligence community has undergone a resurgence of interest in the
research and development of artificial neural networks. An artificial neural network is an attempt to simulate the manner in which the brain interprets information as determined by the current knowledge of biology, physiology, and psychology [14]. Artificial neural networks behave in much the same manner as biological neural networks, giving many
of the same benefits. Artificial neural nets are fault tolerant, exhibit the ability to learn and adapt to new situations,
and have the ability to generalize based on a limited set of data.
neural nets to process information simultaneously, as opposed to the serial nature of traditional digital computers. This
parallel nature, inherent in neural networks, achieves the increase in speed by distributing the calculations among many
neurons. The network structure provides a level of fault tolerance, which allows the network to withstand component
failures without having the entire network fail.
This paper focuses on a popular feedforward model of neural networks. In this model a set of inputs are
applied to the network, and multiplied by a set of connection weights. All of the weighted inputs to the neuron are
then summed and an activation function is applied to the summed value. This activation level becomes the neurons
output and can be either an input for other neurons, or an output for the network. Learning in this network is done by
adjusting the connection weights based upon training vectors (input and corresponding desired output). When a training vector is presented to a neural net, the connection weights are adjusted to minimize the difference between the desired and actual output. After a network is trained with a set of training vectors, the network should produce a good
output match for the inputs.
Artificial neural networks are mainly used in two areas --- pattern recognition, and pattern matching.
Pattern recognition is performed by classifying an unknown pattern through comparisons with previously learned pat-
translation - 29.2
terns. This ability is termed associative recall. An example of this is the use of neural networks to recognize handwritten digits [4]. In pattern recognition, when a particular pattern is noisy or distorted, the network can generalize
and choose the closest match [3][4][10][13]. Pattern matching uses continuous input patterns to evoke continuous
output patterns (i.e. one-to-one mappings). An example is the use of a neural network as a basic controller for a plant.
The controller would accept the plant conditions as the inputs, and a set of control outputs would drive the current manufacturing process. This approach is discussed in more detail in [11]. This paper uses the pattern matching strategy
for the inverse kinematics of a typical three link manipulator.
This paper focuses on the kinematic control of a three link manipulator working in three dimensional
space, within a quarter of the robot workspace. When manipulators interact with the environment, the position of the
end effector must be described. Human users are more suited to working in cartesian coordinates than the joint coordinates of robots; hence, there is a need for conversion between the two coordinate systems. The conversion from
world coordinates to robot joint angles is called inverse kinematics. In the past, the problem of solving the inverse
kinematics has been done through the application of various methods. These methods take one of two basic forms (i)
Closed form equations or, (ii) Iterative. A closed form solution is found by decoupling the joints responsible for position and those responsible for orientation. These solution types are more desirable since explicit equations are available. Iterative techniques, such as, Newton-Raphson, Newton-Gauss and Linearization are available when no closed
form solutions can be obtained. Many of these solutions require that the kinematics be performed off-line because of
the computational demands of the algorithms. A brief survey of these various algorithms is given in [2] and [15]. Both
of the forms are generally rigid and do not account for uncontrollable variables such as manufacture tolerances, calibration error, and wear. A neural network approach could adapt to changes in the robot due to wear from extended
use over time. Such algebraic and algorithmic procedures cannot incorporate these situations.
Previous research in the application of neural networks to kinematic control was performed for a to a twolink planar manipulator (working in a plane), and a five degree of freedom robot working in three dimensional space.
The manipulators were controlled with a high accuracy; however, the training region for the two link robot was a small
square in the centre of the workspace, and the five link robot was limited to a small wedge volume within the robot
workspace [5][6][7]. This paper explains research extended to investigate the performance throughout the entire work
space of the robot. Investigations of a more complete representation of inverse kinematics throughout the entire workspace is provided.
The paper begins with a general introduction to the artificial neuron, a brief description of feedforward
neural networks, and a description of the backpropagation learning algorithm. The inverse kinematics problem is described, and the application of a two-layer neural network is discussed. The use of compensation networks is explored
and the results of the experimental work are presented. The implications of the neural network approach to the kinematic control problem are presented at the conclusion of this paper.
translation - 29.3
used in this paper, the connection weights between neurons are adjusted. The neuron receives inputs opi from neuron
ui while the network is exposed to input pattern p. Each input is multiplied by a connection weight wij, where wij is
the connection between neurons ui and uj. The connection weights correspond to the strength of the influence of each
of the preceding neurons. After the inputs have been multiplied by the connection weights for input pattern p, their
values are summed, netpj. Included in the summation is a bias value j to offset the basic level of the input to the
activation function, f(netpj), which gives the output opj . Figure 1 shows the structure of the basic neuron.
Neuron uj
wij
opi
activation function
where, wij
opi
j
opj
netpj
opj
opj = f(netpj)
netpj
netpj = wijopi + j
synaptic weight, or connection weight between neuron ui and neuron uj
output of neuron ui (input of neuron uj)
bias value of neuron uj
output of neuron uj (function of the summation f(netpj))
Figure 1 Basic structure of an artificial neuron.
In order to establish a bias value j, the bias term can appear as an input from a separate neuron with a
fixed value (a value of +1 is common). Each neuron requiring a bias value will be connected to the same bias neuron.
The bias values are then self-adjusted as the other neurons learn, without the need for extra considerations.
In calculating the output of the neuron, the activation function may be in the form of a threshold function,
in which the output of the neuron is +1 if a threshold level is reached and 0 otherwise. Squashing functions limit the
linear output between a maximum and minimum value. These linear functions, however, do not take advantage of
multi-layer networks [14]. Hyperbolic tangents and the sigmoidal functions are similar to real neural responses; however, the hyperbolic tangent is unbounded and hard to implement in hardware. In this paper, the Sigmoidal function
is used because of its ability to produce continuous non-linear functions, which can be implemented in hardware in
future research areas. Figure 2 shows some commonly used activation functions.
translation - 29.4
linear
sigmoidal logistic
hyperbolic tangent
f(netpj) =
Squashing Function
1
- net
1-e
linear threshold
-1
pj
A single neuron can only simulate 14 of the 16 basic boolean logic functions. It cannot emulate the X-OR,
or the X-NOR gates [9]. These limitations require that more than a single neuron be used, and thus, the architecture
becomes an important consideration. In Feedforward systems, networks propagate the inputs forward through the neural net. Each neuron has inputs that only come from neurons in preceding layers. A one-layer neural net consists of a
layer of input neurons and a layer of output neurons. A multi-layer network consists of an input layer of neurons, one
or more hidden layers, and an output layer.
The circular nodes in figure 3 represent basic neurons, as described in the previous section. The input neurons are shown as squares because they only act as terminal points (i.e. opi = input). The input layer does not process
information; thus, it is not considered to be a part of the structure and is numbered layer 0. A simple one-layer network
can effectively map many sets of inputs to outputs. In pattern recognition problems, this depends upon the linear separability of the problem domain. In pattern matching problems, this depends upon continuity, and topography of the
function. If such a set of connection weights cannot be found using one-layered networks, then multi-layered networks
must be considered.
translation - 29.5
input
layer
output
layer
input
layer
Figure 3
hidden
layer
output
layer
Learning methods may be either supervised, or unsupervised. Supervised learning is required for pattern
matching, and backpropagation is the most popular of the supervised learning techniques. In the backpropagation
Training Paradigm, the connection weights are adjusted to reduce the output error. In the initial state, the network has
a random set of connection weights. When a system starts with all connection weights equal, the network begins at a
sort of local optimum, and will not converge to the global solution. In order for the network to learn, a set of inputs
are presented to the system and a set of outputs are calculated. A difference between the actual outputs and desired
outputs is calculated and the connection weights are modified to reduce this difference [14].
After inputs have been applied and the network output solution has been calculated, the estimated error
contribution, , by each neuron must be calculated. The calculations begin at the output layer of the network. The
delta value for any output neuron is computed as,
pj = (tpj - opj) f (netpj)
(1)
where, tpj is the desired output of output neuron uj, and opj is the actual output, and f (netpj) is the first derivative of
the activation function with respect to the total input (netpj) for the given input pattern p evaluated at neuron uj.
For the output layer, the change in connection weights can easily be calculated since the of the output
neurons is easily determined. With the introduction of hidden layers, the desired outputs of these hidden neurons becomes more difficult to estimate. In order to estimate the delta of a hidden neuron, the error signal from the output
translation - 29.6
layer must be propagated backwards to all preceding layers. The delta, pj , for the hidden neurons is calculated as,
pj = f (netpj)
pkwjk
(2)
k
where, f (netpj) is the first derivative of the activation function with respect to netpj at hidden neuron uj, pk is the
delta value for the subsequent neuron uk, and wjk is the connection weight for the link between hidden neuron uj and
subsequent neuron uk. This process of calculating the s is performed for all layers until the input layer is reached
After all the error terms have been calculated for all the neurons, the weights may be adjusted. The estimated weight change pwij(n+1) is calculated for each input connection to neuron uj from neuron ui by,
pwij(n+1) = pj opi + wij(n)
new delta
(3)
momentum
term
where is the learning rate, is the smoothing term, pj is the delta of neuron uj, opi is the output of preceding neuron
ui, and wij(n) is the weight change in the previous training interval. The estimated errors are used to estimate the
weight changes with the least mean squares method.
The backpropagation method is essentially a gradient descent approach which minimizes the errors. In
order to adjust the connection weights, the gradient descent approach requires that steps be taken in weight space. If
the steps are too large, then the weights will overshoot their optimum values. If the steps are too small, it may take
longer to converge and may become caught in local minima. The step size is then a function of the learning rate. The
learning rate, which varies between 0 and 1, should be large enough to come up with a solution in a reasonable amount
of time without having the solution oscillate within weight space. By introducing the momentum term, which varies
between 0 and 1, the learning rate can be increased while preventing oscillation. The momentum term is introduced
by adding to the weight change some percentage of the last weight change. Common values for range from 0.1 to
0.9, and common values of range from 0.2 to 0.8
The training set is comprised of points which have inputs and desired outputs, which may or may not be
in a random order. In this paper, the training points were chosen at random from an ordered table. This allows a weight
correction after each point is trained. This method helps avoid local optimum points in weight space, by creating a
random walk type of weight optimization, and gives a much faster global convergence. This process is then repeatedly
performed for all training examples until a satisfactory output error is achieved. For an in-depth derivation of the back
propagation method, the reader is suggested to consult Rumelhart and McClelland [14].
In order for a robotic manipulator to perform tasks within space, a user must specify a location in three
dimensional space. The robotic controller must then determine the correct joint coordinates to locate the manipulator.
This area of robotics, Inverse Kinematics, has been well researched and many good solutions exist. In all cases the
* note: The subscript notation (i, j, k) means that neuron ui precedes neuron uj, which precedes uk, etc.
translation - 29.7
resultant solutions are highly specific to a particular robot configuration, with exact dimensions. Such explicit solutions do not tolerate changes over time. These changes may be caused by poor tolerances in manufacture, wear over
extended periods of operation, damage to the robot, and poor calibration. Computations are often complex, can be
quite slow, and require a computer capable of performing complex mathematical functions.
An example of a robot which may be used with an artificial neural network control system is the three link
manipulator as shown in figure 4. The forward and inverse kinematic equations, used by the authors in testing, are
given in the equations (4) and (5).
Z0
1
a2
variable
1
2
3
1
2
3
90o
0o
0o
0
a2
a3
0
0
0
a3
Z1
3
Z2
Denavit-Hartenburg
Transformations
T3 = A1A2A3
c1c23
s1c23
s23
-c1s23
-s1c23
c23
s1
-c1
0
c1(a3c23 + a2c2)
s1(a3c23 + a2c2)
a3s23 + a2s2
translation - 29.8
px = c1(a3c23 + a2c2)
py = s1(a3c23 + a2c2)
pz = a3s23 + a2s2
py
1 = tan-1
px
2 = tan-1
pz2
pz - a2s2
3 = tan-1
- a 2 c2
tan-1
pz
s1 = sin 1
where
c1 = cos 1
c12 = cos (1 + 2)
= +
px2
and
s12 = sin (1 + 2)
+ py 2
It should be noted that the inverse kinematics is not determinate. The sources of indeterminacy for this
manipulator occur:
for Elbow up and Elbow Down Configurations, and
when the end of the arm is directly over the origin, 1 is undefined.
Other examples of indeterminacy occur in PUMA style robots where left and right arm configurations would give indeterminacy. These sources of indeterminacy may all be overcome by constraining the inverse kinematics to a particular case.
Neural Network solutions have the benefit of having faster processing times since information is processed
in parallel. The solutions may be adaptive and still be implemented in hardware by using specialized electronics. Neural systems can generalize to approximate solutions from small training sets. Neural systems are fault tolerant and
robust. The network will not fail if a few neurons are damaged, and the solutions may still retain accuracy. When
implementing a neural network approach, complex computers are not essential and robot controllers need not be spe-
translation - 29.9
then the network will be very accurate when dealing with points near the cluster. In this case, the robot should be familiar with points throughout the entire workspace; thus, points should be evenly distributed. The order in which
points are presented to the network also affect the speed and quality of convergence. If the points are not presented in
a random order, each training update will tend to train the network for the current section of space which the points are
from.
The neural network architecture chosen for this problem is a simple network with one hidden layer. The
network has three input neurons for the desired position, and three output neurons for the estimated joint angles. A
bias neuron (with a value of +1) is attached to all of the neurons in the hidden layer, and to all of the output neurons.
The number of neurons in the hidden layer varied from 10, 20, and 40. This network architecture is pictured in figure 5.
input
layer
hidden
layer
output
layer
+1
bias
neuron
Figure 5
29.1 SUMMARY
translation - 29.10
29.5 REFERENCES
translation - 29.11
translation - 29.12
analog IO - 30.1
Objectives:
30.1 INTRODUCTION
Elements of embedded systems (with program examples of each)
NOTE: emphasize a top-down program structure with subroutines for each
one.
logical IO - digital inputs and outputs
analog outputs - immediate
analog inputs - delayed in (show use of nops and loop to wait), use potentiometer
timers including PWM to control transisor/h-bridge for motor control,
show sound generation also.
counters including encoder decoding and tachometer decoding
serial IO
other peripherals such as displays, sounds, etc.
Control system fundamentals
a simple PID feedback loop using nops and time calcs to ensure time
a multiple step process that waits for an input and does a task with the PID
loop. Show an executive subroutine and calls to
dealing with events
event types; asynchronous, delayed
polling
interrupts
Concurrent processes
single thread vs concurrent processing
how to implement a single control thread
how to create multiple processes
real-time
the need for multiple processes
non-time critical
time critical; regular updates and minimum time between runs
analog IO - 30.2
priority levels
hard vs. soft
Structured Design
General systems design topics
show the general structure of a program with an executive routine
that calls task routines.
show a program that mixes an asynchronous GUI mixed with a
realtime routine
Modal System design
show the use of global mode bits to track the mode of the system
show how the mode bits changes the flow of execution in the executive routine.
Flowcharts
map flowchart structures to program structure
using a register value to track the location in the flowchart
State Diagrams
show the use of state diagrams to model a process and then how the
program is written for it.
Failure Analysis
show basic probability theory
show parallel vs serial failures
show failure estimation using theory for single and chained failure
modes
show the methods for categorizing failures as a hazard, danger, etc.
Communication
Receiving and sending strings
String Handling
Parsing strings
Composing strings
Command and response structures
Error Checking
Non byte oriented data
Networked structured, ie destination address header
A full feedback control system
Other topics
Keyboard multiplexing
output refreshing, LEDs
analog IO - 30.3
30.3 SUMMARY
31. WRITING
Reports are the most common form of document written by engineering. Report
writing is an art that we often overlook, but in many cases can make a dramatic impact on
how our work is perceived. Your reports are most likely to find their way to a superiors
desk than you are to meet the individual.
get beyond the first paragraph, but they have already lost the interest of the
reader.
Transitions are not that important - Students are often coached to create clean transitions between sentences and paragraphs. As a result they often add unnecessary sentences and words. Words that are often warning signs are also and then.
You dont need to keep the good stuff to the end - Many authors try to write the
report so that there is a climax. It can be very frustrating for a reader who
reads 90% of a report before they encounter some discussion of the results.
(note: a report is not like a mystery novel.)
Saying it more than once is acceptable- Over 3/4 of students feel that it is unacceptable to state facts more than once. In truth you want to state fact as many
times as necessary to make a technical point. In the case of very important
details, this will be in the abstract, introduction discussion and conclusion.
A better procedure for writing engineering reports is detailed below. The method
leaves writing to one of the last stages, but it becomes much easier when done this way.
1. Plan and do the work as normal. Regardless of what the report requires this will
often include drawings, sketches, graphs/charts of collected data, pictures, etc.
2. Do the analysis (preferably on computer) of the data and results. These should
be organized into a logical sequence.
3. Review the results to ensure they make sense and follow a logical flow.If necessary add figures to help clarify.
4. Review the materials to verify that they make sense without the text.
5. Put in point form notes to lay out the document.
6. Write the text for the report.
7. Verify that the report conforms to guidelines.
8. Proofread.
should also contain the analysis and conclusions. The completeness of detail allows you
(and others) to review these and verify the correctness of what has been done. These have
been historically used for hundreds of years and are accepted as a form of scientific and
legal evidence. It is completely unacceptable to make incorrect entries or leave out important steps or data.
Standard Format:
1. Title, Author, Date - these make it clear what the labs contain, who did the work,
and when it was done. If the report is done for a course it should also include
the course name/number, the lab instructor and the lab section.
2. Purpose - a brief one line statement that allows a quick overview of what the
experiment is about. This is best written in the form of a scientific goal using
the scientific methods.
3. Theory - a review of applicable theory and calculations necessary. Any design
work is done at this stage
4. Equipment - a list of the required equipment will help anybody trying to replicate the procedure. Specific identifying numbers should be listed when possible. If there are problems in the data, or an instrument is found to be out of
calibration, we can track the problems to specific sets of data and equipment.
5. Procedure - these are sequential operations that describe what was done during
the experiment. The level of detail should be enough that somebody else could
replicate the procedure. We want to use this as a scientific protocol.
6. Results (Note: sometimes procedure and results are mixed) - the results are
recorded in tables, graphs, etc. as appropriate. It will also be very helpful to note
other events that occur (e.g. power loss, high humidity, etc.)
7. Discussion - At this stage the results are reviewed for trends and other observations. At this point we want to consider the scientific method.
8. Conclusions - To conclude we will summarize the significant results, and make
general statements either upholding or rejecting our purpose.
Style: These are meant to be written AS the work is done. As a result the work
should be past tense
Laboratory reports should have one or more hypotheses that are to be tested. If
testing designs these are the specifications. Examples might be,
- what is the thermal capacity of a material?
- what is the bandwidth of an amplifier?
- will the counter increment/decrement between 0 to 9?
NOTE: These reports are much easier to write if you prepare all of the calculations, graphs, etc. before you start to write. If you sit down and decide to do things as you
write it will take twice as long and get you half the marks...... believe me, I have written
many in the past and I mark them now.
1
q = --- ( coffee air )
R
where,
q = heat flow rate from coffee to air (J/s)
R = thermal resistance between air and coffee
= temperatures in the coffee and air
We can also consider that coffee has a certain thermal capacity for the heat energy.
As the amount of energy rises, there will be a corresponding temperature increase. This is
known as the thermal capacitance, and this value is unique for every material. The basic
relationships are given below. I will assume that the energy flow rate into the coffee is
negligible.
1
1
coffee = ---------------- ( q in qout ) = ---------------- q
C coffee
C coffee
where,
d coffee
1 ----------------- = --------------q
dt
C coffee
d coffee
1
----------------- = ---------------q
dt
C coffee
d coffee
1
1
-----------------= --------------------------------- --- ( coffee air )
dt
M coffee coffee R
d coffee
1
1
------------------ + ------------------------------------- coffee = ------------------------------------- air
dt
M coffee coffee R
M coffee coffee R
This differential equation can then be solved to find the temperature as a function
of time.
Guess
= A + Be
BCe
Ct
Ct
d--- = BCe
dt
Ct
1
1
Ct
+ ------------------------------------- ( A + Be ) = ------------------------------------- air
M coffee coffee R
M coffee coffee R
B
A
1
Ct
e BC + ------------------------------------- + ------------------------------------- + ------------------------------------- air = 0
M coffee coffee R
M coffee coffee R
M coffee coffee R
B
BC + ------------------------------------- = 0
M coffee coffee R
1
C = ------------------------------------M coffee coffee R
1
A
------------------------------------- + ------------------------------------- air = 0
M coffee coffee R M coffee coffee R
A = air
C(0)
= air + B
B = 0 air
t
---------------------------------M coffee coffee R
The time constant of this problem can be taken from the differential equation
above.
= M coffee coffee R
Equipment:
1 ceramic coffee cup (14 oz.)
2 oz. ground coffee
1 coffee maker - Proctor Silex Model 1234A
1 thermocouple (gvsu #632357)
1 temperature meter (gvsu #234364)
1 thermometer
2 quarts of tap water
1 standard #2 coffee filter
1 clock with second hand
1 small scale (gvsu# 63424)
Procedure and Results:
1. The coffee pot was filled with water and this was put into the coffee maker. The
coffee filter and grounds were put into the machine, and the machine was turned on. After
five minutes approximately the coffee was done, and the pot was full.
2. The mass of the empty coffee cup was measured on the scale and found to be
214g.
3. The air temperature in the room was measured with the thermometer and found
to be 24C. The temperature of the coffee in the pot was measured using the thermocouple
and temperature meter and found to be 70C.
4. Coffee was poured into the cup and, after allowing 1 minute for the temperature
to equalize, the temperature was measured again. The temperature was 65C. Readings of
the coffee temperature were taken every 10 minutes for the next 60 minutes. These values
were recorded in Table 1 below. During this period the cup was left on a table top and
allowed to cool in the ambient air temperature. During this period the mass of the full coffee cup was measured and found to be 478g.
temperature (deg C)
65
53
43
35
30
28
26
Discussion:
The difference between the temperature of the coffee in the pot and in the cup was
5C. This indicates that some of the heat energy in the coffee was lost to heating the cup.
This change is significant, but I will assume that the heating of the cup was complete
within the first minute, and this will have no effect on the data collected afterwards.
The readings for temperature over time are graphed in Figure 1 below. These show
the first-order response as expected, and from these we can graphically estimate the time
constant at approximately 32 minutes.
temp
(deg C)
60
40
24
t (min)
20
0
20
32min
40
60
We can compare the theoretical and experimental models by using plotting both on
the same graph. The graph clearly shows that there is good agreement between the two
curves, except for the point at 30 minutes, where there is a difference of 3.5 degrees C.
temp
(deg C)
experimental data
mathematical model
60
max. difference
of 3.5 deg. C
40
t (min)
20
0
20
40
60
3.5
error = ------------------ 100 = 8.5%
65 24
Finally, the results can be used to calculate a thermal resistance. If we know the
mass of the coffee and assume that the coffee has the same specific heat as water, and have
the time constant, the thermal resistance is found to be 1731sC/J.
= M coffee coffee R
sC
Conclusion:
In general the models agreed well, except for a single data point. This error was
relatively small, only being 8.5% of the entire data range. This error was most likely
caused by a single measurement error. The error value is greater than the theoretical value,
which suggests that the temperature might have been read at a "hot spot". In the procedure
the temperature measuring location was not fixed, which probably resulted in a variation
in measurement location.
31.6 RESEARCH
Purpose: After looking at a technical field we use these reports to condense the
important details and differences. After reading a research report another reader should be
able to discuss advanced topics in general terms.
Strategy A:
1. Clearly define the objectives for the report
2. Outline what you know on a word processor in point form and find the holes
3. Do research to find the missing information
4. Incorporate the new and old information (still in point form)
5. Rearrange the points into a logical structure
6. Convert point form into full text
7. Proof read and edit
Purpose: These reports allow the developer or team to document all of the design
decisions made during the course of the project. This report should also mention avenues
not taken. Quite often the projects that we start will be handed off to others after a period
of time. In many cases they will not have the opportunity to talk to us, or we may not have
the time. These reports serve as a well known, central document that gathers all relevant
information.
Strategy A:
1. Define the goals for the project clearly in point form
2. Examine available options and also add these in point form
3. Start to examine engineering aspects of the options
4. Make engineering decisions, and add point form to the document
5. As work continues on the project add notes and figures
6. When the project is complete, convert the point form to full text.
7. Proof read and edit
31.9.1 Executive
Purpose: These reports condense long topics into a very brief document, typically
less than one page in length. Basically these save a manager from having to
read a complete report to find the details that interest him/her.
31.9.2 Consulting
Purpose: These reports are typically commissioned by an independent third party
to review a difficult problem. The consultant will review the details of the problem, do tests as required, and summarize the results. The report typically ends
with conclusions, suggestions or recommendations.
31.9.3 Memo(randum)
Purpose: An internal business letter designed to convey a business policy or infor-
mation.
- begin with Date, From, To, Subject:, cc:
31.9.4 Interim
Purpose: This report is normally a formal report to track the progress of a project.
When a project is initially planned, it will be given a timeline to follow. The
interim report will indicate progress relative to the initial timeline, as well as
major achievements and problems.
31.9.5 Poster
Purpose: A presentation format that conveys the key concept visually so that a
spectator would have a concept of the project in a glance, and be able to review
the key concepts in under one minute.
- use a minmimum of 16pt, or higher font.
- abundant usage of figures
- put on foan core backing board, or worst, a purchased cardboard backing board.
- The format uses bullets (still full sentences) to abreviate the presentation.
- the poster should be self-explanatory
- be neat
- use color in a tasteful way.
Typical sections include;
purpose/motivation - a brief summary of the purpose of the work. A summary of the major work detials and outcomes.
approach - the work that was done
conclusions - the results of the work that was assigned
acknowledgements - indicate others who played a significant role in the
results.
31.9.7 Oral
Purpose: Work may be presented to an audience in a formal presentation.
- This can be a very effective method of comminucation when done well, but it can
also be very cruel because an audience may be forced to sit through an uncomfortable experience.
- use a minmimum of 16pt, or higher font.
- abundant usage of figures
- The format uses bullets (still full sentences) to abreviate the presentation.
- the presentation should be self-explanatory
- use color, visuals and sounds in a tasteful way.
- present concepts, by introducing, presenting then reviewing.
31.9.8 Patent
Purpose: A document to describe a design for legal purposes.
- a patent documents that an individual/company owns the idea.
- it gives them the right to sue those who use the idea and infringe
- the patent is submitted to the patent office and reviewed. If the reviewers are satisfied, then the patent is granted, and a patent number is assigned.
- a patent only gives the holder the right to sue.
- Normally a patent lists the inventor (normally an engineer), but is assigned to the
employer. If the inventor is an employee, they assign the rights to the employer
in exchange for $1.
- Patents are publically available after they are granted to encourage innovation.
- Patents have a life of 20 years from the date of application.
- A patent is only available for ideas that can be patented.
- the patent is structured into the follwing sections;
Abstract - a brief description of the patent
References - A list of other patents, or other publications that are related.
These are expected as most patents are built upon previous concepts.
Claims - An itemized list of features that make the design unique from previous ideas. Normally the are at least a dozen claims. These are the items
that are the legal basis for the patent.
Description - A techibcally detailed description of the patentable
work.Tihis normally includes sections such as Summary of the Invention, Description of the Drawings, Detailed Description of the Invention.
It is common for the description to include drawings and equations.
- figures
- tables
- equations
When these elements are included, there MUST be a mention of them in the written text.
These days it is common to cut and paste figures in software. Make sure
- the resolution is appropriate
- the colors print properly in the final form or print well as black and white
- the smallest features are visible
- scanned drawings are clean and cropped to size
- scanned photographs are clear and cropped to size
- digital photographs should be properly lit, and cropped to size
- screen captures are clipped to include only relevant data
31.11.1 Figures
Figures include drawings, schematics, graphs, charts, etc.
They should be labelled underneath sequentially and given a brief title to distinguish it from other graphs. For example Figure 1 - Voltage and currents for 50 ohm resistor
In the body of the report the reference may be shortened to Fig. 1
The figures do not need to immediately follow the reference, but they should be
kept in sequence. We will often move figures to make the type setting work out better.
If drawing graphs by computer,
- if fitting a line/curve to the points indicate the method used (e.g. linear regression)
- try not to use more than 5 curves on the same graph
- use legends that can be seen in black and white
- clearly label units and scales
- label axes with descriptive term. For example Hardness (RHC) instead of
RHC
31.11.2 Graphs
Units on axes
Should have points drawn and connected with straight (or no) lines if experimental.
Smooth lines are drawn for functions or fitted curves. If a curve has been fitted
the fitting method should be described.
if using graphing software dont put a title on the graph.
- titles should clearly, and disctinct, indicate why the content of the figure is signif-
icant.
31.11.3 Tables
Tables are often treated as figures.
They allow dense information presentation, typically numerical in nature.
Number
Color
Shape
Material
car
red
rectangular
die cast
truck
blue
long
polyprop.
motorcycle
green
small
wood
31.11.4 Equations
When presenting equations, use a good equation editor, and watch to make sure
subscripts, etc are visible.
Number equations that are referred to in the text.
Box in equations of great significance.
left justify equations
Fx
Fy
= T 1 sin 60 + F R sin R = 0
= T 1 T 1 cos ( 60 ) + F R cos R = 0
T 1 sin 60
T 1 + T 1 cos 60
F R = ---------------------- = -----------------------------------sin R
cos R
(1)
(2)
sub (1) into (2)
sin R
sin 60
--------------------------- = --------------- = tan R
1 + cos 60
cos R
0.866
tan R = ---------------1 + 0.5
R = 30
98 sin 60 = F R sin 30
F R = 170N
31.11.7 References
References help provide direction to the sources of information when the information may be questioned, or the reader may want to get additional detail.
Reference formats vary between publication sources. But, the best rule is be consistent.
One popular method for references is to number them. The numbers are used in
the body of the paper (eg, [14]), and the references are listed numerically at the end.
Another method is to list the author name and year (eg, [Yackish, 1997]) and then
list the references at the end of the report.
Footnotes are not commonly used in engineering works.
References on the internet should include a working web address, a description of
the page (normally the page title), a date viewed and the page author if available.
31.11.8 Acknowledgments
When others have contributed to the work but are not listed as authors we may
31.11.9 Abstracts
An abstract is a brief summary of the results of the project. It us used by potential
readers to determine if they are interested in reading a report. It should repeat a number of
key details. An abstract should not omit key findings, as if it were an advertisement for a
movie.
31.11.10 Appendices
Reports often include appendices to reduce bulk from the body. When putting
material in an appendix it should stand alone and be referred to in the body of the report. It
is customary to provide a summary of the results in the appendix, unless it is tutorial in
nature. Examples of common appendices are given below.
sample calculations - these are redundant numerical calculations, or a prolonged
derivation of equations. The body of the report has a summary of key assumptions,sample calculations and results. The calculations are often provided so
that the reader may verify the calculations.
long tables of data - tables of numerical data are often put in appendices. Typically
a sample of the table is included in the body for discussion purposes. These are
often provided for the reader who wants to use the data beyond the uses in the
report.
program listings - long listings of computer programs are often put in appendices.
They are referenced in the body on the report near the algorithm/calculation/
method they implement. These listings are provided for readers who want to
use the program.
multiple data graphs - multiple sets of data graphs are often put in appendices and
summarized in a report body. The graphs are often provided so that the reader
may use the graphs for verification or further analysis.
reviews of basic theory - these are often referenced in the body of the report for
readers who may not have seen a topic previously. These are uncommon in student reports.
addition.
- holes that form patterns must be dimensioned relative to each other and
relative to a major feature
- smaller dimensions should be closer to the part
- chained dimensions must be aligned
- hole sizes and dimensions should be on the profile view
- arcs/circles more than 180 degrees are sized by diameter, otherwise radius
is used.
- redundant dimensions should be eliminated
Tolerancing
- Tolerances must be reasonable for manufacturing
- Tolerances must ensure proper assembly and operation at maximum/minimum material conditions
- mating parts should not have identical dimensions, they should be free
running or press fits.
- Smaller tolerances should be used for mating parts
- a general part tolerance should be defined for the part, and smaller tolerances indicated for critical dimensions to reduce clutter
31.11.14 Discussions
- discussions are expected to state all of the details. Sometines these might be obvious when the data is observed, but they must be stated.
31.11.15 Conclusions
- This is the section to restate the key numbers that support the purpose of the lab.
- clear, concise recommendations are needed.
- You must not make statements such as this lab was a wonderful learning experience or all students should enjoy this lab.
31.11.16 Recomendations
- This is the section to restate the key numbers that support the purpose of the lab.
- clear, concise recommendations are needed.
-
31.11.17 Appendices
- Some material is too bulky for a report body, these are normally moved to an
appendix.
- when material is placed in an appendix, it must be summarized in the body of the
report.
- The report must briefly summarize (ussually a figure or equation or more) and
then refer to the appendix.
- it is expected that there will be some duplication between the appendix and the
summary in the appendix.
31.11.18 Units
- Put a apce between numbers an units
- verify that units match the numerical results]
- radians are one of the units that may not observe norma; conventions.
developed for efficiency and clarity. Examples include DMM, HTTP, kitted,
parted, etc. All acronyms should be defined at their first use.
Colloquialisms - Avoid informal language in technical reports. Use of informal
language such as show me the beef will look unprofessional, confuse some
readers and make material easily dated.
Bad
calculated
found,
measured
read
Bad
chose
optimized
parallax error
human error
Bad
it became obvious that
came in at
representing
Bad
implementation of
important
precise
exact
perfect
noted to be
invoved
allowed for it
was found to be
was looked thorough ---> reviewed
along with - with
also
then
decided on - selected
found - measured, calculated
A = _____ / 10
Communication:
clear and concise
good use of figures, graphs and tables
use of english
professional style
passive voice/past tense
suitable use of formatting
B = _____ / 10
Overall:
A * B = ______ / 100
Team:
no
Follows defined formats and requirements
Comments:
31.16 PATENTS
yes
A method of swing on a swing is disclosed, in which a user positioned on a standard swing suspended by two chains from a substantially horizontal tree branch induces side to side motion by
pulling alternately on one chain and then the other.
Inventors: Olson; Steven (337 Otis Ave., St. Paul, MN 55104)
Appl. No.: 715198
Filed: November 17, 2000
Current U.S. Class: 472/118
Intern'l Class: A63G 009/00
Field of Search: 472/118,119,120,121,122,123,125
References Cited [Referenced By]
U.S. Patent Documents
242601Jun., 1881Clement472/118.
5413298May., 1995Perreault248/228.
Primary Examiner: Nguyen; Kien T.
Attorney, Agent or Firm: Olson; Peter Lowell
Claims
I claim:
1. A method of swinging on a swing, the method comprising the steps of:
a) suspending a seat for supporting a user between only two chains that are hung from a tree
branch;
b) positioning a user on the seat so that the user is facing a direction perpendicular to the tree
branch;
c) having the user pull alternately on one chain to induce movement of the user and the swing
toward one side, and then on the other chain to induce movement of the user and the swing toward
the other side; and
d) repeating step c) to create side-to-side swinging motion, relative to the user, that is parallel to
the tree branch.
2. The method of claim 1, wherein the method is practiced independently by the user to create the
side-to-side motion from an initial dead stop.
3. The method of claim 1, wherein the method further comprises the step of:
e) inducing a component of forward and back motion into the swinging motion, resulting in a
swinging path that is generally shaped as an oval.
4. The method of claim 3, wherein the magnitude of the component of forward and back motion is
less than the component of side-to-side motion.
Description
TECHNICAL FIELD
The present invention relates to a method of swinging on a swing.
BACKGROUND OF THE INVENTION
A few basic types of swings have been around for generations. Perhaps the most common is one
that includes a seat suspended between two ropes or chains that are hung from a tree branch or
other substantially horizontal support. These swings are often found in side-by-side sets of two or
three or more on, for example, a school playground.
Young children often need help to climb onto a swing, and may need a push (sometimes even an
"underdog" push) to begin swinging. Others may be able to begin the swinging movement on
their own by pushing with their feet against the ground, and once moving may coordinate the
motion of their legs and body in what may be called "pumping" to sustain the movement of the
swing. When swinging in this manner, the user travels along a path as generally shown in the
cross-section of FIG. 1. Another method of swinging on a swing involves twisting the seat around
repeatedly so that the chains or ropes are wound in a double helix. When allowed to unwind, the
swing spins quickly, which can be entertaining for the user.
These methods of swinging on a swing, although of considerable interest to some people, can lose
their appeal with age and experience. A new method of swinging on a swing would therefore represent an advance of great significance and value.
SUMMARY OF THE INVENTION
In accordance with one embodiment of the present invention, a method is provided for swinging
on a swing. The swing comprises a seat for supporting a user that is suspended between two
chains that are hung from a substantially horizontal tree branch. The method comprises the steps
of: a) positioning a user on the seat; and b) having the user pull alternately on one chain to induce
movement of the user and the swing toward one side, and then on the other chain to induce movement of the user and the swing toward the other side, to create side-to-side motion. In another
embodiment of the invention, the swinging method may be practiced independently by the user to
create the side-to-side motion from an initial dead stop. These and other features of the invention
are described in greater detail below.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic top view of the swinging path of a swing used in accordance with conventional swinging methods.
FIG. 2 is a front view of a swinging path of a swing used in accordance with one embodiment of
the swinging method of the present invention.
FIG. 3 is a schematic top view of a swinging path of a swing used in accordance with a second
embodiment of the swinging method of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
The present inventor has created, through experimentation on a standard swing, a new and
improved method of swinging. The swing is of the type described above, in which a seat is suspended between two chains that are hung from a substantially horizontal tree branch. As is apparent to those of ordinary skill in the area of swinging, the chains could be replaced with ropes,
cables, or the like, or the tree branch could be replaced with another substantially horizontal support such as a metal bar or pole.
The standard swing should be a single swing that is suspended sufficiently far away from obstructions to make the practice of the inventive swinging method completely safe. That is, the swing
should be suspended a sufficient distance away from the trunk of the tree from which it suspended, and from any other swing, building, support, overhead wire, or other obstruction or threat
to safety that may be present.
The standard method of swinging on a swing is defined by oscillatory motion of the swing and the
user along an axis that is substantially perpendicular to the axis of the tree branch from which the
swing is suspended. This "forward and back" movement has been known for generations, and is
illustrated in FIG. 1. In contrast to the conventional method of swinging, the present inventor has
discovered that much greater satisfaction can be obtained by alternately pulling on one chain to
move the swing and the user toward that side, and then pulling on the other chain to move the
swing and the user toward that side. This side-to-side oscillatory motion of the swing and the user
is thus along an axis that is substantially parallel to the axis of the tree branch from which the
swing is suspended, and is illustrated in FIG. 2. This side to side swinging method has the added
benefit that it can be continued for long periods of time simply by alternately pulling on one chain
and then the other. The importance of sufficient clearance between the swing and any obstructions
or threats to the user's safety is apparent.
The present inventor has discovered certain other improvements in the art of swinging on a swing,
either or both of which can be used in conjunction with the swinging method described immediately above. The first is that the inventive swinging method can be initiated from a dead stop without pushing, and without the user having to contact the ground. That is, the user can climb onto
the swing, and begin from an initial dead stop to pull first on one chain, and then on the other
chain, alternately until the user and the swing have begun to swing side-to-side in accordance with
the inventive swinging method described herein. This enables even young users to swing independently and joyously, which is of great benefit to all.
Another improvement on the swinging method described above is the induction into the side-toside swinging movement of a component of forward-and-back motion. That is, by skillful manipulation of the body, the present inventor has found it possible to add a relatively minor component
of forward-and-back motion to the side-to-side swinging motion, resulting in a swinging path that
is generally shaped like an oval, as is shown in FIG. 3. It is preferred that the magnitude of the forward-and back motion (shown in FIG. 3 as being along the Y axis) be less than the magnitude of
the side-to side motion (shown in FIG. 3 as being along the X axis), so that the latter predominates. In this manner, the motion can be more easily continued simply by alternately pulling on
one chain and then the other in the manner described.
Lastly, it should be noted that because pulling alternately on one chain and then the other resembles in some measure the movements one would use to swing from vines in a dense jungle forest,
the swinging method of the present invention may be referred to by the present inventor and his
sister as "Tarzan" swinging. The user may even choose to produce a Tarzan-type yell while swinging in the manner described, which more accurately replicates swinging on vines in a dense jungle
forest. Actual jungle forestry is not required.
Licenses are available from the inventor upon request.
32. PROJECTS
32.1
The project is intended to emphasize proper project management and team skills to
produce a complex engineering system from concept to completion.
32.2 OVERVIEW
Cranes are used to move large loads in industrial environments. To do this they
normally lift a load and then move it to a new position. At the end of the motion
the load is put down. If the load is swaying, the operator may have to pause
before putting down the load. This delay is undesirable because it increases the
total cycle time for the crane.
This design project will endeavor to build a proof-of-concept prototype for an antisway system for a gantry crane as shown in Figure 32.1. The crane will ride on
a wooden beam constructed with 2x4 lumber (actually smaller). An electric
motor will drive the cart left or right. The mass will hang below the cart.
motor driven cart
pivot point
1Kg mass
Figure 32.1
The equipment
The system will use active compensation to move the cart between positions, and
to move the cart to compensate for swaying of the suspended mass. The overall
system performance will be evaluated on how quickly the crane can move the
load, and then eliminate sway. Additional factors used in assessment of performance will be the mass, cost and overall design quality.
width of the 2x4. None of the cart, except for the arms, should be less than 2cm
below the bottom of the 2x4.
The objectives for the design are itemized below. The following equation will be
used to assess the overall score for the design.
- The equipment cost should be minimized. The goal of $150 or less has
been specified as very important.
- Mass should be minimized, the target value is 0.2Kg.
- A clearly justified and supported design process. This will be judged
before the competition.
- The settling time, ts, to less than a +/- 0.5 inch margin around the commanded position. This time will be determined using the best ts/d for two
arbitrary target motions.
- The test will start with the mass stationary at one position. The gantry system must move the mass to another position once started. The total
movement range will be 2 to 20 inches in 2 inch increments.
C
------t s 2 -------B
T
200
0.2
-score =
( 4 ) ( 10 ) ( 10 ) ( 2 )
d
where,
t s = the time to settle (s)
32.3 MANAGEMENT
The management details for the project are outlined below. These may evolve as
the semester progresses.
32.3.2 Teams
The teams are composed of up to 4 students from EGR 345 and one student from
EGR 101. The EGR 345 teams are grouped using the skills survey attached in
the appendices.
Teams are expected to divided tasks for members to work in parallel. It is also
expected that team members will review the work of others to ensure accuracy
and completeness. This is particularly true of calculations, materials lists drawings and budgets.
All team members are expected to work in a professional manner. The general
rules of conduct in a team are,
- treat others as you want to be treated
- communicate expectations and problems clearly
- be polite and accommodating
- when problems arise, help to solve them, even if they are not your fault.
Dont lay the blame for problems on others.
Personal conflicts must be resolved by team members in a professional manner.
The performance of the team will be assessed using peer evaluations on a regular basis and may impact individual grades. In the case of non-participation the
penalty may be up to 100% of the project grade.
In the event that team members cannot resolve differences with a team member,
the team may fire the team member by a vote (it must be unanimous, except
for the member in question). In this case the fired student is responsible for
finding another team that will accept (or hire) them. In the event they cannot
find another team to join, they will be expected to perform all of the work themself. The firing mechanism is intended to deal with individuals who harm the
progress of the team, not for non-participants.
32.4 DELIVERABLES
The following items are to be submitted at various stages of the project.
tion.
- a table of contents
- three view drawings of each significant part
- block diagram of the control system
- block diagrams showing the system architecture
- circuit schematic
- if a motion profile will be used, it should be documented
- an assembly drawing of the mechanism, including a BOM
- a budget listing each of the parts that must be purchased/acquired. Catalog pages
and quotes can be used to validate the budget. In the final report, copies of
receipts, or catalog pages will be required.
- a weight inventory, itemized by each part of the design
- additional calculations for mechanical design issues, such as stress that may
result in failure. Normally these result in a factor of safety.
- the equations of motion for the system
- a Scilab program that verifies the operation of the system using the equations of
motion.
- a C program that implements the controller as designed.
32.5.2 Drawings
All drawings will observe the standards used in EGR 101 (see the EGR 101 or 345
course pages). This includes dimensions and tolerances that can be produced
using the available equipment and materials. Please note that sketch means
that it is done by hand, approximately, while drawing means it is done formally
in a CAD package. Normally you should create solid models, and then generate
multiview drawings. Note: All drawings must have a title block.
Shaded views have very little value and should be avoided, wireframe drawings
are much more useful. If there is a definite need to include a shaded drawing,
change the background to change the quantity of toner used.
32.5.4 Calculations
Calculations are required to justify the design work. These should follow the conventions used in EGR 345. When computer programs are written, they should
be commented and included.
32.6 APPENDICES
68HC11
Digital in
motor position
Analog in
mass angle
Figure 32.2
H-bridge V s Motor
PWM
cart
hanging mass
mass
1--D
quadrature
input
encoder
V
motor
potentiometer
System Architecture
A block diagram for the system is shown in Figure 32.3. A motion generator is
used to generate a smooth path for the mass. There are two feedback control loops. The
first loop adjusts the motor for general positioning of the system. The inner loop adjusts
the output to reduce sway of the cart. The two control signals are added to produce a command signal the drives the load to the final position and then reduces vibrations.
Cp
generate
setpoint
table
update
setpoints
Cd
where,
C p = commanded position
C d = desired position setpoint
encoder
K enc
Cp
Cd
C pe
C pc
Cc
Vs
K pp
PWM
motor
C Lc
K sp
where,
CL
A/D
VL
Figure 32.3
potentiometer
K pot L V zero
cart and
mass
x
L
M p l L
Fp
x
l
2
L Mp l
Mc x
Fw
Mc
Mp
Fp
L
where,
M p g sin L
x cos L M p
M p g cos L
x sin L M p
M c = mass of cart
M p = mass of payload
F p = force in suspension arm
F w = force from wheels
L = angle of payload from vertical
l = length of suspension arm
r w = radius of cart wheel
F = Mp g sin L Mp lL x cos L Mp = 0
l L = g sin L x cos L
(1)
F
=
F
M
g
cos
p
p
L
L M p l + x sin L M p = 0
F p = M p g cos L + L M p l x sin L M p
Fx
= M c x + F p sin L + F w = 0
M c x + F p sin L + F w = 0
2
Figure 32.4
(2)
2
Fw rw
K
K
F w
K2
K
x = w rw
given
2
K
J K -
Fw = V s --------- x ----2- x -------
r w R
r w
r 2w R
(1) --->
2
2
K
J K -
M c x + ( M p g cos L + L M p l x sin L M p ) sin L + V s --------- x ----2- x ------- = 0
r w R
r 2w R
r w
2
J
2
K
2
K -
x M c + ----2- + ( sin L ) M p + cos L sin L M p g + L ( M p l sin L ) + V s --------- x ------- = 0
r w R
r 2w R
rw
2
2
2
2
M c r w + J + ( sin L ) M p r w
2
K - -------K -
-------------------------------------------------------------
-------x
x 2
= ( cos L sin L )M p g + L ( M p l sin L ) + V s
2
r w R
r w R
rw
2
M p gr w
M p l sin L r w
- + L -------------------------------------------------------------- +
x = ( cos L sin L ) ------------------------------------------------------------- M c r 2w + J + ( sin L ) 2 M p r 2w
M c r 2w + J + ( sin L )2 M p r 2w
2
Kr w
K
- + x -----------------------------------------------------------------------
V s ---------------------------------------------------------------------- R ( M c r 2w + J + ( sin L ) 2 M p r 2w )
R ( M c r 2w + J + ( sin L ) 2 M p r 2w )
Figure 32.5
State Equations:
x = v
2
2
M p gr w
L = L
2
M p l sin L r w
2
-
L ------------------------------------------------------------- M c r 2w + J + ( sin L ) 2 M p r 2w
Kr w
-
+ V s ---------------------------------------------------------------------- R ( Mc r 2w + J + ( sin L ) 2 M p r 2w )
v cos L
-----gL =
sin L ----------------l
l
Figure 32.6
A Scilab program that simulates the given system is given in Figure 32.7.
// contest.sce
// System component values
l = 0.4; // 40cm
Mp = 1.0; // 1kg
Mc = 0.2; // 200g
g = 9.81; // good old gravity
rw = 0.03; // 6cm dia tires
K = 0.5; // motor speed constant
R = 7; // motor resistance
J = 0.005; // rotor inertia
// System state
x0 = 0;
// initial conditions for position
v0 = 0;
theta0 = 0.0;// the initial position for the load
omega0 = 0.0;
X=[x0, v0, theta0, omega0];
// The controller definition
PI = 3.14159;
ppr = 16;// encoder pulses per revolution;
Kpot = 1.72;// the angle voltage ratio
Vzero = 2.5;// the voltage when the pendulum is vertical
Vadmax = 5;// the A/D voltage range
Vadmin = 0;
Cadmax = 255;// the A/D converter output limits
Cadmin = 0;
tolerance = 0.5;// the tolerance for the system to settle
Kpp = 20;// position feedback gain
Ksp = 2;// sway feedback gain
Vpwmmax = 12; // PWM output limitations in V
Cpwmmax = 255; // PWM input range
Cdeadpos = 100;// deadband limits
Cdeadneg = 95;
function foo=control(Cdesired)
Cp = ppr * X($, 1)/(2*PI*rw);
Cpe = Cdesired - Cp;
Cpc = Kpp * Cpe;
VL = Kpot * X($,3) + 2.5; // assume the zero angle is 2.5V
CL = ((VL - Vadmin) / (Vadmax - Vadmin)) * (Cadmax - Cadmin);
if CL > Cadmax then CL = Cadmax; end// check for voltages over limits
if CL < Cadmin then CL = Cadmin; end
CLc = Ksp * (CL - (Cadmax + Cadmin) / 2);
Cc = Cpc + CLc;
Cpwm = 0;
if Cc > 0.5 then// deadband compensation
Cpwm = Cdeadpos + (Cc/Cpwmmax)*(Cpwmmax - Cdeadpos);
end
if Cc <= -0.5 then
Cpwm = -Cdeadneg + (Cc/Cpwmmax)*(Cpwmmax - Cdeadneg);
end
foo = Vpwmmax * (Cpwm / Cpwmmax) ; // PWM output
if foo > Vpwmmax then foo = Vpwmmax; end // clip voltage if too large
if foo < -Vpwmmax then foo = -Vpwmmax; end
endfunction
Figure 32.7
// Integration Set the time length and step size for the integration
steps = 5000;// The number of steps to use
t_start = 0;// the start time - normally use zero
t_end = 10;// the end time
h = (t_end - t_start) / steps;// the step size
t = [t_start];// an array to store time values
for i=1:steps,
t = [t ; t($,:) + h];
X = [X ; X($,:) + h * derivative(X($,:), t($,:), h)];// first order
end
Figure 32.8
Figure 32.9
contract Dr. Farris and Dr. Jack will acts as arbiters. If this occurs, one or both of the parties
will be penalized. This may involve actions as severe as receiving a failing grade in the project.
Exhibits:
1. Schedule of Actions
Oct 15-24 - Cart designs are developed by 101 and 345 students
resulting in submission of the Formal Proposal
Oct 30, 2003 - Cart Build Approval submitted
Nov 11 - Initial build completed
Nov 12 - First test completed and First List of Deficiencies submitted
Nov 15 - Design Report Draft submitted for review
Nov 19 - Final test completed and Final List of Deficiencies submitted
Nov 25 - Competition
Dec 4 - Submit Final Report
proficient
The ability to build components with wood, plastic, metal or other materials.
Hands-on Electrical:
Basic wiring skills, soldering, etc.
Writing:
Layout and write complex documents
Teamwork Skills:
The ability to work with others in a team environment.
Leadership Skills:
The ability to act as a role model that teammates will follow.
Design Skills:
Work in unstructured/semistructured problem solving.
Personal/Technical Strengths:
Personal/Technical Weaknesses:
People you would like to work with:
People you would NOT like to work with:
Other Commitments (courses, work, etc. - give hours for each)
Other Items of Interest:
poor
Did the teammate return e-mails and other forms of communication promptly? Could the teammate
understand, explain and evaluate the technical aspects of the project in a clear concise manner?
Did your teammate come to meetings on time? Did the teammate participate in all aspects of the
project? How much did the teammates efforts contribute to the overall success of the project?
Meets deadlines:
Did you teammate complete individual tasks on time? Did the teammate keep the project progressing forward in a timely
manner with a consistent effort throughout the project or was the teammate only available when the team was in trouble?
Quality of work:
Was you r teammate willing to accept and carry out individual tasks on time? How well were these
individual tasks carried out? Did your teammate do his or her fair share of the work?
Overall:
Would you be happy working with the person again? Would you give this person a job reference?
YES / NO
Height of Payload:
= 40cm
Mass of Cart:
not including the harness
M =
Kg
C =
C =
[ 0, 1 ]
T =
[ 0, 1 ]
d =
ts =
Cost:
proof of costs
Build Quality:
assigned by judges. 0 is the best
Technical Analysis:
0 is the best
Test Distance:
the distance specified in the test
Settling Times:
The time required to move within +/- 0.5in. from the target.
Overall Score:
based upon other values
------t s 2 -------B
T
200
0.2
score = --- ( 4 ) ( 10 ) ( 10 ) ( 2 )
d
Comments:
SCORE =
Team Number:
average
excellent
poor
6 7
10
10
10
10
10
10
10
10
10
10
10
11
10
12
10
33.3.1 Graphs
- Label all axis, including units
- Provide a legend for surves with multiple traces
- Equations are plotted as smooth functions.
- Data is plotted as discrete points.
33.4 SCILAB
Basic ruules are,
- when solving problems using computational tools, such as C programs, or
numerical simulations, provide copies of the programs.
- when used to solve problems, include program listings
33.5 TERMINOLOGY
The terminology is significant,
graph: a plot,
numbers: a value
equation: a function, a relationship, an expression,
proof: a derivation
34.1 INTRODUCTION
This section has been greatly enhanced, and tailored to meet our engineering
requirements.
The section outlined here is not intended to teach the elements of mathematics,
but it is designed to be a quick reference guide to support the engineer required to use
techniques that may not have been used recently.
For those planning to write the first ABET Fundamentals of Engineering exam,
the following topics are commonly on the exam.
- quadratic equation
- straight line equations - slop and perpendicular
- conics, circles, ellipses, etc.
- matrices, determinants, adjoint, inverse, cofactors, multiplication
- limits, LHospitals rule, small angle approximation
- integration of areas
- complex numbers, polar form, conjugate, addition of polar forms
- maxima, minima and inflection points
- first-order differential equations - guessing and separation
- second-order differential equation - linear, homogeneous, non-homogeneous,
second-order
- triangles, sine, cosine, etc.
- integration - by parts and separation
- solving equations using inverse matrices, Cramers rule, substitution
- eigenvalues, eigenvectors
- dot and cross products, areas of parallelograms, angles and triple product
- divergence and curl - solenoidal and conservative fields
- centroids
- integration of volumes
name
lower case
upper case
alpha
beta
gamma
delta
epsilon
zeta
eta
theta
iota
kappa
lambda
mu
nu
xi
omicron
pi
rho
sigma
tau
upsilon
phi
chi
psi
omega
Figure 34.1
The constants listed are amount some of the main ones, other values can be
derived through calculation using modern calculators or computers. The values are typically given with more than 15 places of accuracy so that they can be used for double precision calculations.
1 n
e = 2.7182818 = lim 1 + --- = natural logarithm base
n
n
= 3.1415927 = pi
= 0.57721566 = Eulers constant
1radian = 57.29578
Figure 34.2
commutative
a+b = b+a
distributive
a ( b + c ) = ab + ac
associative
a ( bc ) = ( ab )c
Figure 34.3
a + (b + c) = (a + b) + c
34.1.2.1 - Factorial
A compact representation of a series of increasing multiples.
n! = 1 2 3 4 n
0! = 1
Figure 34.4
(x )( x ) = x
n+m
(x )
nm
---------= x
m
(x )
x = 1 , if x is not 0
x
1
= ----px
n
n m
(x )
Figure 34.5
= x
nm
1
--n
x =
x
( xy ) = ( x ) ( y )
Properties of exponents
m
---n
x-- = n------xy
n y
x = 10
x = n
The basic natural logarithm (e is a constant with a value found near the start of this section:
ln x = log e x = y
Figure 34.6
x = e
Definitions of logarithms
logn ( n ) = 1
logn ( 1 ) = 0
x
logn -- = log n ( x ) logn ( y )
y
y
log n ( x ) = ylogn ( x )
log m ( x )
log n ( x ) = ------------------logm ( n )
ln ( A ) = ln ( A ) + ( + 2k )j
Figure 34.7
Properties of logarithms
kI
( a + x ) = a + na
Figure 34.8
n1
n( n 1 ) n 2 2
n
x + -------------------- a
x ++x
2!
c)
A ( 5 + B ) C = 5A + B C
A + BA B
------------= ---- + ---C+D
C D
log ( ab ) = log ( a ) + log ( b )
d)
5(5 ) = 5
e)
3 log ( 4 ) = log ( 16 )
f)
( x + 6 ) ( x 6 ) = x + 36
log ( 5 )
10
= -----10
5
6
(-----------------x + 1 ) - = x 2 + 2x + 1
a)
b)
g)
h)
(x + 1)
x ( x + 2 ) 3x
2
(-------------------------------------x + 3 ) ( x + 1 )x -
h)
(x + 1) x
3
log ( x )
64
-----16
15
3------ + ----21 28
(x y )
2
4x 8y
i)
5---
4
-------5
j)
(y + 4) (y 2)
k)
2 3 4
4
5--- 8---
3 9
l)
x y
x----------+ 1= 4
x+2
(ans.
+ BA
B
1 1
a) A
-----------= ------- + ------- = --- + --AB
AB AB
B A
AB b) -----------A + B cannot be simplified
4 5
x y )
c) (------------- 2 -
x
2 5 3
6 15
= (x y ) = x y
a)
(ans.
a)
5. Rearrange the following equation so that only y is on the left hand side.
y---------+ x= x+2
y+z
(ans.
y---------+ x= x+2
y+z
y + x = (x + 2)(y + z)
y + x = xy + xz + 2y + 2z
y xy 2y = xz + 2z x
y ( x 1 ) = xz + 2z x
xz + 2z x
y = -------------------------x1
t3 + 5
lim ----------------
t 0 5t 3 + 1
b)
t3 + 5
lim ----------------
t 5t 3 + 1
(ans.
a)
b)
3
t3 + 5
0 +5 ---------------------lim ---------------=
= 5
3
t 0 5t 3 + 1
5(0) + 1
3
3
t3 + 5
+
5
- = -------------- = 0.2
lim ---------------- = ----------------------3
3
t 5t 3 + 1
5() + 1
5()
34.2 FUNCTIONS
Binomial
P(m) =
t p q
n
t nt
q = 1p
q, p [ 0, 1 ]
tm
Poisson
P(m) =
tm
e
-----------t!
>0
Hypergeometric
r s
t n t
P ( m ) = ----------------------r + s
tm
n
Normal
1 x t2
P ( x ) = ---------- e dt
2
Figure 34.9
Distribution functions
ax + bx + c = 0 = a ( x r 1 ) ( x r 2 )
b b 4ac
r 1, r 2 = -------------------------------------2a
Cubic equations also appear on a regular basic, and as a result should also be considered.
x + ax + bx + c = 0 = ( x r 1 ) ( x r 2 ) ( x r 3 )
First, calculate,
2
3b a
Q = ----------------9
9ab 27c 2a
R = --------------------------------------54
S =
R+ Q +R
T =
R Q +R
S+T a j 3
r 2 = ------------ --- + --------- ( S T )
2
3
2
S+T a j 3
r 3 = ------------ --- --------- ( S T )
2
3
2
x + ax + bx + cx + d = 0 = ( x r 1 ) ( x r 2 ) ( x r 3 ) ( x r 4 )
First, solve the equation below to get a real root (call it y),
3
y by + ( ac 4d )y + ( 4bd c a d ) = 0
Next, find the roots of the 2 equations below,
a + a 2 4b + 4y
y + y2 4d
r 1, r 2 = z + ------------------------------------------- z + ------------------------------ = 0
2
2
2
y y2 4d
2 a a 4b + 4y
r 3, r 4 = z + ------------------------------------------- z + ------------------------------ = 0
2
2
is the
order of the
numerator >=
denominator?
yes
no
Find roots of the denominator
and break the equation into
partial fraction form with
unknown values
OR
use limits technique.
If there are higher order
roots (repeated terms)
then derivatives will be
required to find solutions
Done
A B
1 C
x ( s ) = -------------------= ----2 + --- + ----------2
s s+1
s
s (s + 1)
1 -
C = lim ( s + 1 ) -------------------2
s 1
s (s + 1)
1 -
2
A = lim s -------------------2
s0
s (s + 1)
= 1
1
= lim ----------- = 1
s0 s + 1
d 2
1 -
B = lim ----- s -------------------2
s 0 ds
s (s + 1)
d
1
= lim ----- -----------
s 0 ds s + 1
= lim [ ( s + 1 ) ] = 1
s0
5s + 3s + 8s + 6x ( s ) = ------------------------------------------2
s +4
This cannot be solved using partial fractions because the numerator is 3rd order
and the denominator is only 2nd order. Therefore long division can be used to
reduce the order of the equation.
5s + 3
2
s +4
5s + 3s + 8s + 6
3
5s + 20s
2
3s 12s + 6
2
3s + 12
12s 6
This can now be used to write a new function that has a reduced portion that can be
solved with partial fractions.
12s 6x ( s ) = 5s + 3 + --------------------2
s +4
solve
--------------------12s 6A
B
= ------------- + ------------2
s + 2j s 2j
s +4
Figure 34.15 Solving partial fractions when the numerator order is greater than the
denominator
When the order of the denominator terms is greater than 1 it requires an expanded
partial fraction form, as shown below.
5
F ( s ) = -----------------------32
s (s + 1)
5 A B
C
D
E
----------------------= ----2 + --- + ------------------3- + ------------------2- + ---------------3
2
s (s + 1)
(s + 1)
s
(s + 1)
s (s + 1)
5
A B
C
D
E
------------------------ = ---+ --- + ------------------3- + ------------------2- + ---------------3
2
2
s
(
s
+
1)
s
(s + 1)
(s + 1)
s (s + 1)
3
A ( s + 1 ) + Bs ( s + 1 ) + Cs + Ds ( s + 1 ) + Es ( s + 1 ) = ----------------------------------------------------------------------------------------------------------------------------------------3
2
s (s + 1)
4
s ( B + E ) + s ( A + 3B + D + 2E ) + s ( 3A + 3B + C + D + E ) + s ( 3A + B ) + ( A )= -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------3
2
s (s + 1)
0
1
3
3
1
1
3
3
1
0
0
0
1
0
0
0
1
1
0
0
1
2
1
0
0
A
0
B
0
C = 0
D
0
E
5
A
0 1 0
B
1 3 0
C = 3 3 1
D
3 1 0
E
1 0 0
0
1
1
0
0
1
2
1
0
0
0
5
0
15
0 = 5
0
10
5
15
5
5 15
5
10
15
------------------------ = ---- + --------- + ------------------- + ------------------- + ---------------3
2
3
2 (s + 1)
2
s
s
(s + 1)
(s + 1)
s (s + 1)
xi
The notation
is equivalent to x a + x a + 1 + x a + 2 + + x b assuming a
i=a
and b are integers and b a . The index variable i is a placeholder whose name does not
matter.
Operations on summations:
b
xi
i=a
xi
i=b
xi
i=a
b
= xi
i=a
b
xi + yj
i=a
b
j=a
c
xi +
i=a
( xi + yi )
i=a
c
xi =
i = b+1
x
y =
i j
i = a j = c
b
xi
i=a
b
xi yj
i=a j=c
1
= --- N ( N + 1 )
2
i=1
1 N
----------------, 1
for both real and complex .
= 1
i=0
N, = 1
N1
1
= ------------, < 1 for both real and complex . For 1 , the summation
1
i=0
x + 2x + 1 = ( x + 1 ) ( x + 1 )
a)
x 2x + 1
b)
x 1
c)
x +1
2
2
d)
x + x + 10
e)
f)
2x + 8x = 8
(ans.
2x + 8x = 8
2
x + 4x + 4 = 0
2
(x + 2) = 0
x = 2, 2
3. Reduce the following expression to partial fraction form.
x + 4---------------2
x + 8x
(ans.
x+4
A
BAx
+ 8A + Bxx + 4------------------------------------ + -----------------------------------------=
=
=
2
2
x(x + 8)
x x+8
x + 8x
x + 8x
( 1 )x + 4 = ( A + B )x + 8A
8A = 4
A = 0.5
1 = A+B
B = 0.5
x+4
0.5 ----------------- = 0.5
------- + ----------2
x
x
+8
x + 8x
34.3.1 Trigonometry
The basic trigonometry functions are,
y
1
sin = -- = ----------r
csc
x
1
cos = -- = ----------r
sec
r
y
y
1
sin
tan = -- = ----------- = -----------x
cot
cos
Pythagorean Formula:
2
r = x +y
Sine - sin
1
-270
-180
-90
0
-1
90
180
270
360
450
Cosine - cos
1
-270
-180
-90
90
180
270
360
450
-1
Tangent - tan
1
-270
-180
-90
0
-1
90
180
270
360
450
Cosecant - csc
-270
-180
-90
0
-1
90
180
270
360
450
Secant - sec
1
-270
-180
-90
90
180
270
360
450
-1
Cotangent - cot
-270
-180
-90
0
-1
90
180
270
360
450
NOTE: Keep in mind when finding these trig values, that any value that does not
lie in the right hand quadrants of cartesian space, may need additions of 90 or 180.
Cosine Law:
2
c = a + b 2ab cos c
c
Sine Law:
b
c
a ------------= -------------- = -------------sin A
sin B
sin C
C
a
Now a group of trigonometric relationships will be given. These are often best
used when attempting to manipulate equations.
sin ( ) = sin
cos ( ) = cos
tan ( ) = tan
OR
sin sin
cos ( 1 2 ) = cos 1 cos 2 +
1
2
OR
tan 1 tan 2
tan ( 1 2 ) = -----------------------------------1
+ tan 1 tan 2
cot 1 cot 2
+1
cot ( 1 2 ) = ------------------------------------tan 2 tan 1
sin
1 cos
tan --- = --------------------- = --------------------2
1 + cos
sin
2
( cos ) + ( sin ) = 1
(deg)
-90
-60
-45
-30
0
30
45
60
90
sin
cos
tan
-1.0
-0.866
-0.707
-0.5
0
0.5
0.707
0.866
1.0
0.0
0.5
0.707
0.866
1
0.866
0.707
0.5
0.0
-infinity
-1
0
1
infinity
e +e
cos = ---------------------2
e e
sin = --------------------2j
e e
sinh ( x ) = ------------------ = hyperbolic sine of x
2
x
e +e
cosh ( x ) = ------------------ = hyperbolic cosine of x
2
x
sinh ( x )
e e
- = hyperbolic tangent of x
tanh ( x ) = ------------------- = ----------------x
x
cosh ( x )
e +e
1
2 = hyperbolic cosecant of x
csch ( x ) = ------------------ = ----------------x
x
sinh ( x )
e e
1
2
- = hyperbolic secant of x
sech ( x ) = ------------------- = ----------------x
x
cosh ( x )
e +e
x
cosh ( x )
e +e = hyperbolic cotangent of x
coth ( x ) = ------------------- = ----------------x
x
sinh ( x )
e e
sinh ( x ) = sinh ( x )
cosh ( x ) = cosh ( x )
tanh ( x ) = tanh ( x )
csch ( x ) = csch ( x )
sech ( x ) = sech ( x )
coth ( x ) = coth ( x )
sin ( jx ) = j sinh ( x )
j sin ( x ) = sinh ( jx )
cos ( jx ) = cosh ( x )
cos ( x ) = cosh ( jx )
tan ( jx ) = j tanh ( x )
j tan ( x ) = tanh ( jx )
1. Find all of the missing side lengths and corner angles on the two triangles below.
3
10
(ans.
( s + 3j ) ( s 3j ) ( s + 2j ) =
( s + 3j ) ( s 3j ) ( s + 2j ) = ( s 9j ) ( s + 4js + 4j )
4
2 2
2 2
s + 4js + 4j s 9j s 9j 4js 9j 4j
4
s + ( 4j )s + ( 5 )s + ( 36j )s + ( 36 )
3. Solve the following partial fraction
1 Note: there was a typo here, so --------------an acceptable answer is also. x + 0.5
4
-------------------------=
2
x + 3x + 2
(ans.
A
B
Ax + 2A + Bx + B( 2A + B ) + ( A + B )x
4
-------------------------= ------------ + ------------ = -----------------------------------------= ------------------------------------------------2
2
2
x
+
1
x
+
2
x + 3x + 2
x + 3x + 2
x + 3x + 2
A+B = 0
A = B
2A + B = 4 = 2B + B = B
B = 4
A = 4
4 - ----------4 ----------+
x+1 x+2
34.3.3 Geometry
A set of the basic 2D and 3D geometric primitives are given, and the notation
used is described below,
A = contained area
P = perimeter distance
V = contained volume
S = surface area
x, y, z = centre of mass
x, y, z = centroid
I x, I y, I z = moment of inertia of area (or second moment of inertia)
AREA PROPERTIES:
Ix =
Iy =
xy dA
I xy =
JO =
dA =
(x
x dA
A
- = centroid location along the x-axis
x = ----------- dA
A
y dA
A
- = centroid location along the y-axis
y = ----------- dA
A
Rectangle/Square:
A = ab
P = 2a + 2b
a
x
Centroid:
Moment of Inertia
(about centroid axes):
3
b
x = --2
ba
I x = -------12
3
a
y = --2
b
Moment of Inertia
(about origin axes):
3
ba
I x = -------3
3
b a
I y = -------12
b a
I y = -------3
I xy = 0
b a
I xy = ----------4
2 2
Triangle:
y
bh
A = -----2
P =
h
x
b
Moment of Inertia
(about centroid axes):
Centroid:
a+b
x = -----------3
h
y = --3
Moment of Inertia
(about origin axes):
bh
I x = -------36
bh
I x = -------12
bh 2
2
I y = ------ ( a + b ab )
36
bh 2
2
I y = ------ ( a + b ab )
12
bh
I xy = -------- ( 2a b )
72
Circle:
bh
I xy = -------- ( 2a b )
24
y
A = r
P = 2r
x
Centroid:
Moment of Inertia
Moment of Inertia
(about centroid axes): (about origin axes):
4
x = r
r
I x = -------4
y = r
r
I y = -------4
I xy = 0
Ix =
Iy =
I xy =
Mr
J z = ---------2
Half Circle:
r
A = -------2
P = r + 2r
r
x
Centroid:
x = r
4r
y = -----3
Moment of Inertia
(about origin axes):
Moment of Inertia
(about centroid axes):
8 4
I x = --- ------ r
8 9
r
I x = -------8
r
I y = -------8
r
I y = -------8
I xy = 0
I xy = 0
Quarter Circle:
r
A = -------4
r
P = ----- + 2r
2
r
x
Centroid:
Moment of Inertia
(about centroid axes):
4r
x = -----3
I x = 0.05488r
4r
y = -----3
I y = 0.05488r
r
I y = -------16
I xy = 0.01647 r
Moment of Inertia
(about origin axes):
4
r
------Ix =
16
I xy
r
= ---8
Circular Arc:
y
2
r
A = -------2
P = r + 2r
r
x
Centroid:
2r sin --2
x = ----------------3
y = 0
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
4
Ix =
r
I x = ---- ( sin )
8
Iy =
r
I y = ---- ( + sin )
8
I xy =
I xy = 0
Ellipse:
r1
A = r 1 r 2
P = 4r 1
--2
r2
2
r1 + r 2
2
1 -------------------- ( sin ) d
a
r 1 + r2
P 2 --------------2
Centroid:
Moment of Inertia
Moment of Inertia
(about centroid axes): (about origin axes):
3
x = r2
y = r1
r 1 r 2
I x = ------------4
3
r 1 r 2
I y = ------------4
I xy =
Ix =
Iy =
I xy =
Half Ellipse:
r 1 r 2
A = -----------2
P = 2r 1
--2
2
r1
r1
2
r2
+
2
1 -------------------- ( sin ) d + 2r 2
a
r2
r 1 + r2
P --------------- + 2r 2
2
Centroid:
x = r2
4r 1
y = -------3
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
3
I x = 0.05488r 2 r 1
3
I y = 0.05488r 2 r 1
2 2
I xy = 0.01647r 1 r 2
r 2 r 1
I x = -----------16
3
r 2 r 1
-----------Iy =
16
2 2
r1 r2
I xy = --------8
Quarter Ellipse:
r 1 r 2
A = ------------4
P =
r1
--2
r1
r1 + r 2
2
1 -------------------- ( sin ) d + 2r 2
a
r2
r 1 + r2
P --- --------------- + 2r 2
2
2
Centroid:
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
4r 2
x = -------3
Ix =
I x = r 2 r 1
4r 1
y = -------3
Iy =
I y = r 2 r 1
I xy =
r2 r1
= --------8
2 2
I xy
Parabola:
2
A = --- ab
3
a
4a + b + 16a
b
b + 16a
P = --------------------------- + ------ ln ----------------------------------------
2
b
8a
x
b
Centroid:
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
b
x = --2
Ix =
Ix =
2a
y = -----5
Iy =
Iy =
I xy =
I xy =
Half Parabola:
ab
A = -----3
a
4a + b2 + 16a 2
b + 16a
b
P = --------------------------- + --------- ln ----------------------------------------
4
16a
b
x
b
Centroid:
3b
x = -----8
2a
y = -----5
Moment of Inertia
Moment of Inertia
(about centroid axes): (about origin axes):
3
3
8ba
2ba
I x = -----------I x = -----------175
7
3
19b a
I y = --------------480
2 2
b a
I xy = ----------60
2b a
I y = -----------15
2 2
b a
I xy = ----------6
A general class of geometries are conics. This for is shown below, and can be
used to represent many of the simple shapes represented by a polynomial.
A = B = C = 0
B = 0, A = C
2
B AC < 0
2
B AC = 0
2
B AC > 0
straight line
circle
ellipse
parabola
hyperbola
VOLUME PROPERTIES:
Ix =
rx
Iy =
ry
Iz =
rz
x dV
V
- = centroid location along the x-axis
x = ----------- dV
V
y dV
V
- = centroid location along the y-axis
y = ----------- dV
V
z dV
V
- = centroid location along the z-axis
z = ----------- dV
V
Parallelepiped (box):
V = abc
S = 2 ( ab + ac + bc )
b
x
a
Moment of Inertia
(about centroid axes):
Centroid:
a
x = --2
M(a + b )
I x = --------------------------12
b
y = --2
M(a + c )
I y = --------------------------12
c
z = --2
M(b + a )
I z = --------------------------12
Moment of Inertia
(about origin axes):
Ix =
Jx =
Iy =
Jy =
Iz =
Jz =
Sphere:
4 3
V = --- r
3
S = 4r
Centroid:
x
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
x = r
2Mr
I x = ------------5
y = r
2Mr
I y = ------------5
z = r
2Mr
I z = ------------5
Ix =
2Mr
J x = ------------5
Iy =
2Mr
J y = ------------5
Iz =
2Mr
J z = ------------5
Hemisphere:
2 3
V = --- r
3
z
r
S =
Centroid:
x = r
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
83
2
I x = --------- Mr
320
Ix =
3r
y = ----8
z = r
2Mr
I y = ------------5
Iy =
83
2
I z = --------- Mr
320
Iz =
Cap of a Sphere:
y
h
1 2
V = --- h ( 3r h )
3
z
r
S = 2rh
Centroid:
x = r
y =
z = r
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
Ix =
Ix =
Iy =
Iy =
Iz =
Iz =
Cylinder:
V = hr
S = 2rh + 2r
Centroid:
Moment of Inertia
(about centroid axis):
x = r
h
r
I x = M ------ + ----
12 4
h
y = --2
Mr
I y = ---------2
z = r
r
h
I z = M ------ + ----
12 4
Moment of Inertia
(about origin axis):
2
h
r
I x = M ----- + ----
3 4
M ( 3r + h )
J x = -----------------------------12
Iy =
Mr
J y = ---------2
M ( 3r + h )
J z = -----------------------------12
r
h
I z = M ----- + ----
3 4
Cone:
1 2
V = --- r h
3
2
S = r r + h
Centroid:
Moment of Inertia
(about centroid axes):
3
Moment of Inertia
(about origin axes):
x = r
3h
3r
I x = M -------- + --------
80
20
h
y = --4
3Mr
I y = ------------10
z = r
3h
3r
I z = M -------- + --------
80
20
Ix =
Iy =
3
Iz =
Tetrahedron:
1
V = --- Ah
3
h
A
Centroid:
Moment of Inertia
(about origin axes):
Moment of Inertia
(about centroid axes):
x =
Ix =
Ix =
h
y = --4
Iy =
Iy =
Iz =
Iz =
z =
Torus:
y
r2
1 2
2
V = --- ( r 1 + r 2 ) ( r 2 r 1 )
4
r1
z
2
S = ( r2 r1 )
Centroid:
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
x = r2
Ix =
Ix =
r 2 r 1
y = ------------- 2
Iy =
Iy =
Iz =
Iz =
z = r2
Ellipsoid:
y
r2
4
V = --- r 1 r 2 r 3
3
r3
r1
z
S =
Centroid:
x = r1
y = r2
z = r3
Moment of Inertia
(about origin axes):
Moment of Inertia
(about centroid axes):
Ix =
Ix =
Iy =
Iy =
Iz =
Iz =
Paraboloid:
1 2
V = --- r h
2
h
z
r
S =
Centroid:
x = r
y =
z = r
Moment of Inertia
(about centroid axes):
Moment of Inertia
(about origin axes):
Ix =
Ix =
Iy =
Iy =
Iz =
Iz =
y = mx + b
1
m perpendicular = ---m
y2 y1
m = ---------------x2 x1
--x- + --y- = 1
a b
If we assume a line is between two points in space, and that at one end we have a
local reference frame, there are some basic relationships that can be derived.
( x 2, y 2, z 2 )
( x 1, y 1, z 1 )
d =
( x2 x1 ) + ( y2 y1 ) + ( z2 z1 )
( x 0, y 0, z 0 )
z
The direction cosines of the angles are,
x2 x1
= acos ----------------
d
2
y2 y1
= acos ----------------
d
2
z2 z1
= acos ---------------
d
Explicit
Parametric t=[0,1]
y
b
P 1 = ( x 1, y 1, z 1 )
where the coefficients defined by the intercepts are,
1
A = --a
1
B = --b
1
C = --c
N
P 2 = ( x 2, y 2, z 2 )
D = 1
P 3 = ( x 3, y 3, z 3 )
z
The determinant can also be used,
x x1 y y1 z z1
det x x 2 y y 2 z z 2 = 0
x x3 y y3 z z3
det
y2 y1 z2 z1
y3 y1 z3 z1
( x x 1 ) + det
+ det
z2 z1 x2 x1
z3 z1 x3 x1
x2 x1 y2 y1
x3 x1 y3 y1
( y y1 )
( z z1 ) = 0
2. What is the equation of a line that passes through the points below?
a) (0, 0) to (2, 2)
b) (1, 0) to (0, 1)
c) (3, 4) to (2. 9)
3. Find a line that is perpendicular to the line through the points (2, 1) and (1, 2). The perpendicular line passes through (3, 5).
4. Manipulate the following equations to solve for x.
a) x 2 + 3x = 2
b) sin x = cos x
(ans.
a) x 2 + 3x = 2
2
x + 3x + 2 = 0
2
3 3 4( 1)( 2 )
3 98
31
x = ------------------------------------------------ = ------------------------------ = ---------------- = 1, 2
2(1 )
2
2
b) sin x = cos x
sin x---------= 1
cos x
tan x = 1
x = atan 1
x = , 135, 45, 225,
5. Simplify the following expressions.
a)
(ans.
a)
( cos 2 )
sin 2 ---------------------- + sin 2
sin 2
( cos 2 )
2
2
sin 2 ---------------------- + sin 2 = ( cos 2 ) + ( sin 2 ) = 1
sin 2
6. A line that passes through the point (1, 2) and has a slope of 2. Find the equation for the line,
and for a line perpendicular to it.
(ans.
given,
y = mx + b
m = 2
x = 1
y = 2
substituting
2 = 2(1) + b
y = 2x
b = 0
perpendicular
1
m perp = ---- = 0.5
m
y = 0.5 x
j = 1
Complex Numbers:
a + bj
where,
N* = a bj
Basic Properties:
( a + bj ) + ( c + dj ) = ( a + c ) + ( b + d )j
( a + bj ) ( c + dj ) = ( a c ) + ( b d )j
( a + bj ) ( c + dj ) = ( ac bd ) + ( ad + bc )j
Na + bj
N N*
a + bj c dj
ac + bd bc ad
---- + ------------------ j
= -------------- = ----- ------- = -------------- ------------- = ----------------2
2
2
2
M
c + dj
M N*
c + dj c dj
c +d
c +d
We can also show complex numbers graphically. These representations lead to
alternative representations. If it in not obvious above, please consider the notation above
uses a cartesian notation, but a polar notation can also be very useful when doing large calculations.
CARTESIAN FORM
Ij
imaginary (j)
N = R + Ij
R
real
A =
R +I
I
= atan ---
R
POLAR FORM
R = A cos
I = A sin
Ij
imaginary (j)
N = A
R
real
A = amplitude
= phase angle
We can also do calculations using polar notation (this is well suited to multiplication and division, whereas cartesian notation is easier for addition and subtraction),
A = A ( cos + j sin ) = Ae
( A 1 1 ) ( A 2 2 ) = ( A 1 A 2 ) ( 1 + 2 )
( A 1 1 )
A
--------------------- = -----1- ( 1 2 )
A 2
( A 2 2 )
n
( A ) = ( A ) ( n )
(DeMoivres theorem)
Note that DeMoivres theorem can be used to find exponents (including roots) of
complex numbers
Eulers formula: e
= cos + j sin
e +e
cos = ---------------------2
j
e e
sin = --------------------2j
z
( x, y, z ) ( r, , z )
z
z
y
x
x
x +y
x = r cos
r =
y = r sin
y
= atan --
x
z
( x, y, z ) ( r, , )
r
z
y
x
x
r =
x = r sin cos
y = r sin sin
z
= acos -
r
b)
c)
16 -------------------2
( 4j + 4 )
3j + 5
--------------------2
( 4j + 3 )
( 3 + 5j )4j
y
= atan --
x
z = r cos
a)
x +y +z
where,
j =
(ans. a)
16
16
16--------------------- = ------------------------------------ = ------= 0.5j
2
16 + 32j + 16
32j
( 4j + 4 )
3j + 5
3j + 5 3j + 5 7 24j
35 141j + 72
37 141j
b) -------------------- = --------------------------------= --------------------- --------------------- = --------------------------------------- = ----------------------2
16
+
24j
+
9
7
+
24j
24j
49
+
576
625
( 4j + 3 )
c) ( 3 + 5j )4j = 12j + 20j 2 = 12j 20
y = (x + 2)
34.5.1 Vectors
Vectors are often drawn with arrows, as shown below,
head
terminus
A vector is said to have magnitude (length or
strength) and direction.
origin
tail
j
x
k
becomes
y
A = ( 2, 3, 4 )
A + B = ( 2 + 7, 3 + 8, 4 + 9 )
B = ( 7, 8, 9 )
A B = ( 2 7, 3 8, 4 9 )
B
A+B
Parallelogram Law
A
F2 = 5i + 3j
y
F1 F2
cos = ----------------F1 F2
F1 = 2i + 4j
( 2 ) ( 5 ) + ( 4 )( 3)
= acos -------------------------------------------
2 2 + 4 2 5 2 + 3 2
22
= acos ----------------------- = 32.5
( 4.47 ) ( 6 )
We can use a dot product to project one vector onto another vector.
z
We want to find the component of
force F1 that projects onto the
vector V. To do this we first convert V to a unit vector, if we do
not, the component we find will
be multiplied by the magnitude
of V.
F 1 = ( 3i + 4j + 5k )N
V = 1j + 1k
y
F1
1j + 1k
V
= ------ = --------------------- = 0.707j + 0.707k
V
2
2
1 +1
F 1V = V F 1 = ( 0.707j + 0.707k ) ( 3i + 4j + 5k )N
F 1V = ( 0 ) ( 3 ) + ( 0.707 ) ( 4 ) + ( 0.707 ) ( 5 ) = 6N
F 1V
We can consider the basic properties of the dot product and units vectors.
Unit vectors are useful when breaking up vector magnitudes and direction. As an example consider the vector, and the displaced x-y axes shown below as x-y.
F = 10N
y
y
45
60
x
We could write out 5 vectors here, relative to the x-y axis,
x axis = 2i
y axis = 3j
x axis = 1i + 1j
y axis = 1i + 1j
F = 10N 60 = ( 10 cos 60 )i + ( 10 sin 60 )j
None of these vectors has a magnitude of 1, and hence they are not unit vectors. But, if
we find the equivalent vectors with a magnitude of one we can simplify many tasks.
In particular if we want to find the x and y components of F relative to the x-y axis we
can use the dot product.
x = 1i + 0j (unit vector for the x-axis)
F x = x F = ( 1i + 0j ) [ ( 10 cos 60 )i + ( 10 sin 60 )j ]
= ( 1 ) ( 10 cos 60 ) + ( 0 ) ( 10 sin 60 ) = 10N cos 60
This result is obvious, but consider the other obvious case where we want to project a
vector onto itself,
= 10 ( ( cos 60 ) + ( sin 60 ) ) = 10
Correct
Now consider the case where we find the component of F in the x direction. Again,
this can be done using the dot product to project F onto a unit vector.
u x' = cos 45i + sin 45j
F x' = F x' = ( ( 10 cos 60 )i + ( 10 sin 60 )j ) ( ( cos 45 )i + ( sin 45 )j )
= ( 10 cos 60 ) ( cos 45 ) + ( 10 sin 60 ) ( sin 45 )
= 10 ( cos 60 cos 45 + sin 60 sin 45 ) = 10 ( cos ( 60 45 ) )
Here we see a few cases where the dot product has been applied to find the vector projected onto a unit vector. Now finally consider the more general case,
V2
V1
2
V 2V 1
1
x
First, by inspection, we can see that the component of V2 (projected) in the direction
of V1 will be,
V 2V 1 = V 2 cos ( 2 1 )
Next, we can manipulate this expression into the dot product form,
= V 2 ( cos 1 cos 2 + sin 1 sin 2 )
= V 2 [ ( cos 1 i + sin 1 j ) ( cos 2 i + sin 2 j ) ]
V1 V2
V 1 V2
V1 V2
= V 2 --------- --------- = V 2 ------------------ = ------------------ = V2 V1
V1 V2
V1 V2
V1
Or more generally,
V1 V 2
V2 V 1 = V 2 cos ( 2 1 ) = V 2 -----------------V1 V2
V1 V2
V 2 cos ( 2 1 ) = V 2 -----------------V1 V2
V1 V2
cos ( 2 1 ) = -----------------V 1 V2
*Note that the dot product also works in 3D, and similar proofs are used.
F = ( 6.43i + 7.66j + 0k )N
d = ( 2i + 0j + 0k )m
i
M = dF =
2m
0m 0m
6.43N 7.66N 0N
NOTE: note that the cross product here is for the right hand
rule coordinates. If the left
handed coordinate system is
used F and d should be
reversed.
The cross (or vector) product of two vectors will yield a new vector perpendicular
to both vectors, with a magnitude that is a product of the two magnitudes.
V1 V2
V1
V2
V1 V2 = ( x 1 i + y1 j + z1 k ) ( x2 i + y 2 j + z 2 k )
i j k
V1 V2 =
x1 y1 z1
x2 y2 z2
V 1 V 2 = ( y 1 z 2 z 1 y 2 )i + ( z 1 x 2 x 1 z 2 )j + ( x 1 y 2 y 1 x 2 )k
We can also find a unit vector normal n to the vectors V1 and V2 using
a cross product, divided by the magnitude.
V1 V2
n = --------------------V1 V2
When using a left/right handed coordinate system,
The positive orientation of angles and moments about an axis can be determined
by pointing the thumb of the right hand along the axis of rotation. The fingers
curl in the positive direction.
y
+
z
+
y
+
x
The cross product is distributive, but not associative. This allows us to collect terms
in a cross product operation, but we cannot change the order of the cross product.
r 1 F + r 2 F = ( r 1 + r2 ) F
DISTRIBUTIVE
rFFr
but
r F = ( F r )
NOT ASSOCIATIVE
When we want to do a cross product, followed by a dot product (called the mixed triple
product), we can do both steps in one operation by finding the determinant of the following. An example of a problem that would use this shortcut is when a moment is found
about one point on a pipe, and then the moment component twisting the pipe is found
using the dot product.
ux uy uz
( d F) u =
dx dy dz
F x F y Fz
34.5.5 Matrices
Matrices allow simple equations that drive a large number of repetitive calculations - as a result they are found in many computer applications.
A matrix has the form seen below,
n columns
a 11 a21 a n1
a 12 a22 a n2
a 1m a 2m a nm
m rows
Matrix operations are available for many of the basic algebraic expressions,
examples are given below. There are also many restrictions - many of these are indicated.
A = 2
3 4 5
B = 6 7 8
9 10 11
12 13 14
C = 15 16 17
18 19 20
Addition/Subtraction
21
D = 22
23
E = 24 25 26
3 + 12 4 + 13 5 + 14
B + C = 6 + 15 7 + 16 8 + 17
9 + 18 10 + 19 11 + 20
B + D = not valid
32 42 52
BA = 62 72 82
9 2 10 2 11 2
B D = not valid
3 12 4 13 5 14
B + C = 6 15 7 16 8 17
9 18 10 19 11 20
Multiplication/Division
3--2
B
--- = 6--A
2
9--2
BD =
BC =
( 3 21 + 4 22 + 5 23 )
( 6 21 + 7 22 + 8 23 )
( 9 21 + 10 22 + 11 23 )
4--2
7--2
10
-----2
5--2
8--2
11
-----2
D E = 21 24 + 22 25 + 23 26
( 3 12 + 4 15 + 5 18 ) ( 3 13 + 4 16 + 5 19 ) ( 3 14 + 4 17 + 5 20 )
( 6 12 + 7 15 + 8 18 ) ( 6 13 + 7 16 + 8 19 ) ( 6 14 + 7 17 + 8 20 )
( 9 12 + 10 15 + 11 18 ) ( 9 13 + 10 16 + 11 19 ) ( 9 14 + 10 17 + 11 20 )
B- --B D
--, -, ---- , etc = not allowed (see inverse)
C D B
Note: To multiply matrices, the first matrix must have the same number
of columns as the second matrix has rows.
Determinant
B = 3
7 8 4 6 8 +5 6 7
10 11
9 11
9 10
7 8
10 11
= ( 7 11 ) ( 8 10 ) = 3
6 8
9 11
= ( 6 11 ) ( 8 9 ) = 6
6 7
9 10
= ( 6 10 ) ( 7 9 ) = 3
= 3 ( 3 ) 4 ( 6 ) + 5 ( 3 ) = 0
Transpose
3 6 9
T
B = 4 7 10
5 8 11
D = 21 22 23
24
T
E = 25
26
Adjoint
7 8
10 11
B = 4 5
10 11
4 5
7 8
6 8
10 11
3 5
9 11
6 7
9 10
3 4
9 10
3 5
6 8
3 4
6 7
Inverse
x
D = B y
z
B D = B
x
B y
z
x
x
1
B D = I y = y
z
z
B
= --------B
Identity Matrix
This is a square matrix that is the matrix equivalent to 1.
BI = IB = B
DI = ID = D
B
B = I
1 0 0
1 0,
,
1
0 1 0 , etc=I
0 1
0 0 1
The eigenvalue of a matrix is found using,
A I = 0
Given,
2x+4y+3z = 5
9x+6y+8z = 7
11 x + 13 y + 10 z = 12
Write down the matrix, then rearrange, and solve.
2 4 3 x
5
=
9 6 8 y
7
11 13 10 z
12
x
2 4 3
y = 9 6 8
z
11 13 10
5
7 =
12
Given,
2x+4y+3z = 5
9x+6y+8z = 7
11 x + 13 y + 10 z = 12
Write down the coefficient and parameter matrices,
A =
2 4 3
9 6 8
11 13 10
B =
5
7
12
Calculate the determinant for A (this will be reused), and calculate the determinants
for matrices below. Note: when trying to find the first parameter x we replace the
first column in A with B.
5 4 3
7 6 8
12 13 10
x = ------------------------------ =
A
2 5 3
9 7 8
11 12 10
y = ------------------------------ =
A
2 4 5
9 6 7
11 13 12
z = ------------------------------ =
A
Multiply
ANS.
1 2 3 10
4 5 6 11 =
7 8 9 12
Determinant
12 3
42 6
78 9
1 2 3 10
68
=
4 5 6 11
167
7 8 9 12
266
1 2 3
4 2 6
7 8 9
Inverse
12 3
42 6
78 9
1 2 3
4 2 6
7 8 9
= 36
6
B = 2
1
Cross Product
AB =
Dot Product
AB =
ANS.
A B = ( 4, 17, 10 )
A B = 13
ANS.
x= 0.273
y= -0.072
z= -0.627
5. Solve the following set of equations using a) Cramers rule and b) an inverse matrix.
2x + 3y = 4
5x + 1y = 0
(ans.
a)
2 3 x = 4
5 1 y
0
b)
4 3
0 1
4
x = ---------------- = -------- 13
2 3
5 1
2 4
5 0
20
20
y = ---------------- = --------- = ----- 13
13
2 3
5 1
x = 2 3
y
5 1
1 3
1 ( 5 )
4
-----
5
2
(
3
)
2
1- 4
4 = -------------------------------- 4 = ------------------- 4 = ----= 13
13
13
0
0
20
20
0
2 3
-----13
5 1
A B C X
L
+
D E F Y
M
GH I Z
N
(ans.
A B C X
L
D E F Y + M
GH I Z
N
AX + BY + CZ
L
DX + EY + FZ + M
GX + YH + IZ
N
AX + BY + CZ + L
= DX + EY + FZ + M
GX + YH + IZ + N
= AX + BY + CZ + L DX + EY + FZ + M GX + YH + IZ + N
7. Perform the matrix calculations given below.
a)
A B C X
D E F Y =
GH I Z
b)
A B C
D E F X Y Z =
GH I
8. Find the dot product, and the cross product, of the vectors A and B below.
x
A = y
z
(ans.
p
B = q
r
A B = xp + yq + zr
AB =
i j k
x y z
p q r
yr zq
= i ( yr zq ) + j ( zp xr ) + k ( xq yp ) = zp xr
xq yp
a
b
c
b)
d e f
g h k
mnp
a b
c d
c)
a b
c d
(ans.
a)
b)
a
b
c
d e f
d e f
=
g h k
a b c g h k = ( ad + bg + cm ) ( ae + bh + cn ) ( af + bk + cp )
mnp
mnp
a b
c d
= ad bc
adj
c)
a b
c d
a b
d c
d - ----------------c ----------------c d
b a
ad
bc
ad
bc
= ------------------- = ------------------- =
ad bc
b - ----------------a a b
----------------ad bc ad bc
c d
(ans.
12 3 x
5
14 8 y = 0
42 1 z
1
5 2 3
0 4 8
1 2 1
5 ( 4 16 ) + 2 ( 8 0 ) + 3 ( 0 4 )
60 + 16 12
56
x = ------------------ = ---------------------------------------------------------------------------------- = ---------------------------------- = --------- = 7
1 ( 4 16 ) + 2 ( 32 1 ) + 3 ( 2 16 )
12 + 62 42
8
1 2 3
1 4 8
4 2 1
A B C X
D E F Y =
GH I Z
b)
A B C
D E F X Y Z =
GH I
c)
d)
e)
det A B =
C D
A B
C D
A B
C D
=
1
12. Solve the following set of equations with the specified methods.
3x + 5y = 7
4x 6y = 2
a) Inverse matrix
b) Cramers rule
c) Gauss-Jordan row reduction
d) Substitution
34.6 CALCULUS
NOTE: Calculus is very useful when looking at real systems. Many students are
turned off by the topic because they "dont get it". But, the secret to calculus is to remember that there is no single "truth" - it is more a loose collection of tricks and techniques.
Each one has to be learned separately, and when needed you must remember it, or know
where to look.
Both u, v and w are functions of x, but this is not shown for brevity.
Also note that C is used as a constant, and all angles are in radians.
d----(C) = 0
dx
dd
----( Cu ) = ( C ) ------ ( u )
dx
dx
dd
d
----( u + v + ) = ------ ( u ) + ------ ( v ) +
dx
dx
dx
d- n
n1 d
----( u ) = ( nu
) ------ ( u )
dx
dx
dd
d
----( uv ) = ( u ) ------ ( v ) + ( v ) ------ ( u )
dx
dx
dx
d- u---
v d
u- ----d -------- ------ ( u ) ---=
2 dx
2 dx ( v )
dx v
v
v
dd
d
d
----( uvw ) = ( uv ) ------ ( w ) + ( uw ) ------ ( v ) + ( vw ) ------ ( u )
dx
dx
dx
dx
dd
d
----( y ) = ------ ( y ) ------ ( u ) = chain rule
dx
du dx
d1
----( u ) = -------------dx
d----( x)
du
d----(y)
d
du
------ ( y ) = -------------dx
d----(x)
du
dd
----( sin u ) = ( cos u ) ------ ( u )
dx
dx
dd
----( cos u ) = ( sin u ) ------ ( u )
dx
dx
d1 2d
----( tan u ) = ----------- ------ ( u )
dx
cos u dx
d- u
u d
----( e ) = ( e ) ------ ( u )
dx
dx
d1
----( ln x ) = --dx
x
d2 d
----( cot u ) = ( csc u ) ------ ( u )
dx
dx
d
d----( sec u ) = ( tan u sec u ) ------ ( u )
dx
dx
dd
----( csc u ) = ( csc u cot u ) ------ ( u )
dx
dx
dd
----( sinh u ) = ( cosh u ) ------ ( u )
dx
dx
dd
----( cosh u ) = ( sinh u ) ------ ( u )
dx
dx
d2 d
----( tanh u ) = ( sech u ) ------ ( u )
dx
dx
d-
d- 2
--- ---f(x)
f( x)
dt
dt
f
(
x
)
- =
lim ---------- = lim --------------------- = lim ----------------------x a g(x)
x a d
x a d 2
---- g ( x )
g
(
x
)
dt
---dt
Some techniques used for finding derivatives are,
Leibnitzs Rule, (notice the form is similar to the binomial equation) can be used for
finding the derivatives of multiplied functions.
d- 0 ----d- n
d- 1 ----d- n 1
d- n
---- ---- n ----(
uv
)
=
(
u
)
(
v
)
+
(
u
)
(v)
dx
dx
dx
dx
1 dx
d n2
d 0
d- 2 ----n d n
n ----( u ) -
( v ) + + ------ ( u ) ------ ( v )
n dx
2 dx
dx
dx
34.6.1.2 - Integration
Some basic properties of integrals include,
C dx
= ax + C
Cf ( x ) dx
= C f ( x ) dx
( u + v + w + ) dx
u dv
u dx + v dx + w dx +
= uv v du = integration by parts
1
u = Cx
= ---- f ( u ) du
C
dF(u)
( x ) du = ----------- du
F ( f ( x ) ) dx = F ( u ) ----du
f' ( x )
f ( Cx ) dx
n+1
x n
+C
x dx = ----------n+1
x
x
a+C
a dx = ------ln a
--x- dx
e dx
x
= ln x + C
x
= e +C
u = f(x)
sin x dx
= cos x + C
cos x dx
= sin x + C
( sin x )
( cos x )
3x sin 2x sin 4x
dx = ------ + ------------- + ------------- + C
8
4
32
n+1
( sin x ) n
+C
cos x ( sin x ) dx = -----------------------n+1
( cos x )
sin x cos x + x
dx = ------------------------------- + C
2
sinh x dx
= cosh x + C
sin x cos x + x
dx = ------------------------------- + C
2
cosh x dx
= sinh x + C
tanh x dx
= ln ( cosh x ) + C
cos x ( ( sin x ) + 2 )3
+C
( sin x ) dx = ------------------------------------------3
2
x cos ( ax ) dx
2
2x cos ( ax -) ------------------a x 2
+
sin ( ax ) + C
x cos ( ax ) dx = ------------------------2
3
a
a
n+1
x
n
-+C
x dx = ----------n+1
( a + bx )
ln a + bx
dx = ----------------------- + C
b
2 1
1
a + 2 b
(
a
+
bx
) dx = ---------------------- ln --------------------------- + C, a > 0, b < 0
2 ( b )a
a x b
2
ln ( bx + a -)
+C
x ( a + bx ) dx = --------------------------2b
2 1
(a
x
a
x ab
2 1
( a + bx ) dx = --- ------------- atan ------------- + C
b b ab
a
2
1
a+x
2 1
2
2
x ) dx = ------ ln ------------ + C, a > x
2a
ax
( a + bx )
x(x
2 a + bx
dx = ---------------------- + C
b
1
--2
a ) dx =
CORRECT??
x a +C
2 1
1
(
a
+
bx
+
cx
) dx = ------ ln
2
b
a + bx + cx + x c + ---------- + C, c > 0
2 c
2 1
1
2cx b
(
a
+
bx
+
cx
) dx = ---------- asin ------------------------- + C, c < 0
2
c
b 4ac
1
--2
3
---
1--2
3---
22
( a + bx )
( a + bx ) dx = ----3b
22
( a + bx )
( a + bx ) dx = ----3b
x ( a + bx )
1
--2
3
--2
2 ( 2a 3bx ) ( a + bx ) dx = ---------------------------------------------------2
15b
1
---
2 2
1
ln x + ----2- + x
1
--a
2 2 2
1--x ( 1 + a x ) + --------------------------------------------2
2 2
a
( 1 + a x ) dx = ---------------------------------------------------------------------------------2
3--2 2
1
a ----2- + x
a
--------------------------(
+
x
)
d
x
x
1
a
=
3
1
--2 2 2
---
1
2 2
----3
1
1 ln x + 2 + x
------a
ax
18
2
2 2 2
2 2
2 2 2
---------------------------------------------------------(
1
+
a
x
)
d
x
=
+
x
x
(
1
+
a
x
)
2
2
3
4 a
8a
8a
1--2 2 2
1---
1--asin ( ax )
2 2 2
(
x
)
d
x
=
x
(
1
a
x ) + --------------------1
a
2
a
1--2 2 2
3---
1
--2 2
3
---
a 1
2 2
x ( 1 a x ) dx = --3- a----2- x
1
---
x- 2 2 2 1--x
2
2 2
2
-x
(
a
x
)
d
x
x
)
x
(
a
x
) + a asin ---
=
a
+
8
4
a
2
1
--2
1
---
1
1
2 2
( 1 + a x ) dx = --a- ln x + a----2- + x
2 2
(1 a
1
--2
1
1
x ) dx = --- asin ( ax ) = --- acos ( ax )
a
a
2 2
ax
ax
e +C
e dx = -----a
ax
e ( ax 1 ) + C
xe dx = -----2
a
ax
y = f ( x, y, z )
We can write this in differential form, but the right hand side must contain partial
derivatives. If we separate the operators from the function, we get a simpler form.
We can then look at them as the result of a dot product, and divide it into two vectors.
( d )y = f ( x, y, z ) dx + f ( x, y, z ) dy + f ( x, y, z ) dz
x
y
z
( d )y = dx + dy + dz f ( x, y, z )
x
y
z
i+ j+ k
x y z
F = F x i + F y j + Fz k
Gausss or Greens or divergence theorem is given below. Both sides give the
flux across a surface, or out of a volume. This is very useful for dealing with magnetic
fields.
( F ) dV
V
FdA
A
where,
V, A = a volume V enclosed by a surface area A
F = a field or vector value over a volume
Stokes theorem is given below. Both sides give the flux across a surface, or out
of a volume. This is very useful for dealing with magnetic fields.
( F ) dA
FdL
L
where,
A, L = A surface area A, with a bounding parimeter of length L
F = a field or vector value over a volume
e.g.,
d
d- 2
---x + ----- x = y
dt
dt
e.g.,
d-
d-
--- --- dt x + dt y = 2
d
d-
---x + ----- y = 2
dt
dt
e.g.,
higher order differential equations - have at least one derivative that is higher than
second-order.
partial differential equations - these equations have partial derivatives
Note: when solving these equations it is common to hit blocks. In these cases
backtrack and try another approach.
linearity of a differential equation is determined by looking at the dependant variables in the equation. The equation is linear if they appear with an exponent other than 1.
eg.
y'' + y' + 2 = 5x
2
( y'' ) + y' + 2 = 5x
3
linear
non-linear
y'' + ( y' ) + 2 = 5x
non-linear
non-linear
y' + 2xy 4x = 0
y' 2y = 0
homogeneous
34.6.3.1.1 - Guessing
In this technique we guess at a function that will satisfy the equation, and test it to
see if it works.
y' + y = 0
y = Ce
the guess
t
t
y' + y = Ce + Ce
= 0
The previous example showed a general solution (i.e., the value of C was not
found). We can also find a particular solution.
y = Ce
y = 5e
a general solution
a particular solution
e.g.,
dx
------ + y 2 + 2y + 3 = 0
dy
2
dx = ( y 2y 3 )dy
3
y
2
x = -------- y 3y + C
3
e.g.,
dx
------ + x = 0
dy
1
--- dx = dy
x
ln ( x ) = y
e.g.,
dy
------ = y-- 1
dx
x
y
u = -x
d- 2
d-
--- ---y
+
A
dt
dt y + By = 0
e.g.,
d- 2
d-
--- --- dt y + 6 dt y + 3y = 0
Guess,
y = e
Bt
d-
Bt
--- dt y = Be
2
d-
2 Bt
--- dt y = B e
substitute and solve for B,
2 Bt
Bt
B e + 6Be + 3e
Bt
= 0
B + 6B + 3 = 0
B = 3 + 2.449j, 3 2.449j
substitute and solve for B,
y = e
y = e
y = e
( 3 + 2.449j )t
3 t 2.449jt
3 t
Bt
+ C 2 te
Bt
d- 2
d-
--- ---y
+
A
dt
dt y + By = Cx
to solve these equations we need to find the homogeneous and particular solutions and then add the two solutions.
y = yh + yp
to find yh solve,
d- 2
d
--- --- dt y + A dt y + B = 0
to find yp guess at a value of y and then test for validity, A good table of guesses is,
Cx form
Guess
Ax + B
Ax
e
Cx + D
Ax
Ax
Ce
Cxe
B sin ( Ax )
or B cos ( Ax )
C sin ( Ax ) + D cos ( Ax )
or Cx sin ( Ax ) + xD cos ( Ax )
d- 2
d-
2x
--- ---y
+
y
6y
=
e
dt
dt
First solve for the homogeneous part,
d- 2
d-
--- --- dt y + dt y 6y = 0
try
y = e
Bx
d-
Bx
---= Be
y
dt
2
d-
2 Bx
---y
=
B
e
dt
2 Bx
B e
+ Be
Bx
6e
Bx
= 0
B +B6 = 0
B = 3, 2
yh = e
3x
+e
2x
Next, solve for the particular part. We will guess the function below.
y = Ce
2x
d-
2x
---y = 2C e
dt
2
d-
2x
--- dt y = 4Ce
4Ce
2x
+ 2C e
2x
6Ce
2x
= e
2x
4C 2C 6C = 1
C = 0.25
y p = 0.25e
2x
Finally,
y = e
3x
+e
2x
+ 0.25e
2x
(n 1)
n1
f
f'' ( a ) ( x a )
( a)(x a)
f ( x ) = f ( a ) + f' ( a ) ( x a ) + ------------------------------- + + -----------------------------------------------2!
( n 1 )!
(ans. d 3t
d-
d
2
2t
2
3
2t
----- ------------------ + e 2t = ---( 3t ( 2 + t ) ) + ----- ( e ) = 3 ( 2 + t ) 6t ( 2 + t ) + 2e
2
dt
dt
dt
(2 + t)
2. Solve the following differential equation, given the initial conditions at t=0s.
x'' + 4x' + 4x = 5t
x0 = 0
x' 0 = 0
(ans.
homogeneous solution:
guess:
x'' + 4x' + 4x = 0
xh = e
At
x h' = Ae
2 At
At
A e + 4Ae + 4e
xh = C1 e
2t
+ C 2 te
At
At
2 At
x h'' = A e
2
= 0 = A + 4A + 4 = ( A + 2 ) ( A + 2 )
2t
particular solution:
guess:
x p = At + B
x p' = A
x p'' = 0
( 0 ) + 4 ( A ) + 4 ( At + B ) = 5t
( 4A + 4B ) + ( 4A )t = ( 0 ) + ( 5 )t
5
A = --- = 1.25
4
x p = 1.25t 1.25
combine and solve for constants:
x ( t ) = xh + xp = C1 e
for x(0) = 0
2t
0 = C1 e
for (d/dt)x(0) = 0
4 ( 1.25 ) + 4B = 0
+ C 2 te
2 ( 0 )
2t
+ 1.25t 1.25
+ C 2 ( 0 )e
0 = ( 2 )C 1 e
2 ( 0 )
2 ( 0 )
+ ( 2 )C 2 ( 0 )e
2t
B = 1.25
2( 0)
+ C2 e
2 ( 0 )
+ 1.25
C 2 = 1.25
(ans.
d
d
d
a) ----- ( sin t + cos t ) = ----- ( sin t ) + ----- ( cos t ) = cos t sin t
dt
dt
dt
d
2
3
b) ----- ( ( t + 2 ) ) = 2 ( t + 2 )
dt
d
8t
8t
8t
c) ----- ( 5te ) = 5e + 40te
dt
d
5
d) ----- ( 5 ln t ) = --dt
t
4. Find the following integrals
6 t dt
7t
b) 14 e dt
c) sin ( 0.5t ) dt
2
a)
d)
--x- dx
(ans.
3
t
3
a) 6 t dt = 6 ---- = 2t + C
3
2
7t
e
7t
7t
b) 14 e dt = 14 ------ + C = 2e + C
7
c)
sin ( 0.5t ) dt
d)
--x- dx
cos ( 0.5t )
= -------------------------- + C = 2 cos ( 0.5t ) + C
0.5
= 5 ln ( x ) + C
1 -
d- --------------dx x + 1
b)
d- t
---( e sin ( 2t 4 ) )
dt
(ans.
a)
d- ----------1 -
1
----= ------------------2
dx x + 1
(x + 1)
b)
d
----- ( e t sin ( 2t 4 ) ) = e t sin ( 2t 4 ) + 2e t cos ( 2t 4 )
dt
b)
(ans.
2t
dt
( sin + cos 3 ) d
a)
b)
( sin + cos 3 ) d
2t
2t
dt = 0.5e + C
1
= cos + --- sin 3 + C
3
x( 0) = 1
x' ( 0 ) = 1
(ans.
x( 0) = 1
x'' + 5x' + 3x = 3
x' ( 0 ) = 1
Homogeneous:
2
A + 5A + 4 = 0
5 25 16
53
A = ------------------------------------ = ---------------- = 4, 1
2
2
xh = C1 e
4t
+ C2e
Particular:
xp = A
x'p = 0
x'' p = 0
0 + 5 ( 0 ) + 3A = 3
A = 1
xp = 1
Initial values:
x = x h + xp = C1 e
4t
+ C2 e + 1
1 = C1( 1 ) + C2( 1 ) + 1
x' = 4C 1 e
4t
C2e
1 = 4C 1 ( 1 ) C 2 ( 1 )
4 ( C2 ) + C2 = 1
1 4t 1 t
x = --- e + --- e + 1
3
3
C1 = C2
1
C 2 = --3
1
C 1 = --3
9. Set up an integral and solve it to find the volume inside the shape below. The shape is basically
a cone with the top cut off.
b
x
a
(ans.
V =
dV =
0 A dx
cb
r = b + ----------- x
a
2
cb
cb
cb 2
2
2
A = r = b + ----------- x = b + 2 ----------- x + ----------- x
a
a
a
a
V =
cb 2
2
c b-
x + ----------- x dx
0 b + 2 ---------a
a
cb 2 cb 3
V = b x + ----------- x + --- ----------- x
3 a
a
2
cb 2 cb 2 3
2
V = b a + ----------- a + --- ----------- a
a
3 a
2
2 2
V = b a + ( c b )a + --- ( c b ) a
3
10. Solve the first order non-homogeneous differential equation below. Assume the system starts
at rest.
2x' + 4x = 5 sin 4t
11. Solve the second order non-homogeneous differential equation below.
2x'' + 4x' + 2x = 5
where,
x( 0) = 2
x' ( 0 ) = 0
y(t)
yi + 1
yi
yi 1
ti 1
ti
ti + 1
ti
T
i + y i 1
y-------------------(
)
y
t
ti 1 i 2 - ( ti ti 1 ) = --2- ( y i + yi 1 )
y i y i 1
y i + 1 y i
d---- = -------------------- = --1- ( y i y i 1 ) = --1- ( y i + 1 y i )
y ( t i ) ------------------- ti ti 1
t i + 1 ti
dt
T
T
1
--1- ( y i + 1 y i ) --- ( y i y i 1 )
2 yi + yi 1 + y i + 1
d-
T
T
---------------------------------------------------------------------------------------------------------------
y
(
t
)
=
i
2
dt
T
T
2
1 2 d 2
d
1 3 d 3
1 4 d 4
x ( t + h ) = x ( t ) + h ----- x ( t ) + ----- h ----- x ( t ) + ----- h ----- x ( t ) + ----- h ----- x ( t ) +
dt
2! dt
3! dt
4! dt
d-
---= 0
dt x 0
d-
2
3
---x = 1+x +t
dt
x0 = 0
t (s)
x(t)
d/dt x(t)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
h = 0.1
1
x ( t + h ) = x ( t ) + --- ( F 1 + 2F 2 + 2F 3 + F4 )
6
F 1 = hf ( t, x )
F1
h
F 2 = hf t + ---, x + ------
2
2
F2
h
F 3 = hf t + ---, x + ------
2
2
F 4 = hf ( t + h, x + F 3 )
where,
x = the state variables
f = the differential function
t = current point in time
h = the time step to the next integration point
f ( xi )
x i + 1 = x i ----------------------d ----f ( x )
dx i
The function f(x) is supplied by the user.
This method can become divergent if the function has an inflection point near the
root.
The technique is also sensitive to the initial guess.
This calculation should be repeated until the final solution is found.
F(s) =
0 f ( t )e
st
dt
where,
f ( t ) = the function in terms of time t
F ( s ) = the function in terms of the Laplace s
TIME DOMAIN
FREQUENCY DOMAIN
Kf ( t )
Kf ( s )
f1 ( t ) + f2 ( t ) f3 ( t ) +
df
(t)
----------dt
f1 ( s ) + f2 ( s ) f 3 ( s ) +
sf ( s ) f ( 0 )
d------------f ( t )2
dt
df ( 0 )
2
s f ( s ) sf ( 0 ) -----------------dt
f ( t )d------------n
dt
s f( s) s
0 f ( t ) dt
f-------( s -)
s
f ( t a )u ( t a ), a > 0
at
f(t)
f ( at ), a > 0
tf ( t )
n
t f(t)
f------( t )t
as
n1
f(0 ) s
n 2 df ( 0
f( s)
f(s a)
1 s
--- f --a a
-------------df ( s )ds
n
f ( s -)
( 1 ) -------------n
ds
nd
f ( 0 )-----------------)- d------------------n
dt
dt
f ( u ) du
s
TIME DOMAIN
A
--s
1
---2
s
1---------s+a
t
e
at
----------------2
2
s +
s
----------------2
2
s +
sin ( t )
cos ( t )
te
e
e
FREQUENCY DOMAIN
at
at
at
Ae
sin ( t )
cos ( t )
at
Ate
(s + a) +
s+a ------------------------------(s + a) +
A---------sa
at
2Ae
1
------------------2
(s + a)
-------------------------------
2t A e
A ----------------2
(s a)
cos ( t + )
cos ( t + )
complex conjugate
A
A
----------------------+ -------------------------------------s + + j
s + j
complex conjugate
A
A
------------------------------ + ------------------------------------2
2
( s + j )
( s + + j )
34.9 z-TRANSFORMS
For a discrete-time signal x [ n ] , the two-sided z-transform is defined by
X(z) =
n =
x [ n ]z
. In
n=0
z-Transform
X(z)
ROC
[n]
All z
u[n]
1 --------------1
1z
z >1
nu [ n ]
z
---------------------1 2
(1 z )
n u[n ]
a u[n ]
z >1
( 1 + z )z---------------------------1 3
(1 z )
z >1
1 -----------------1
1 az
z > a
na u [ n ]
( a )u [ n 1 ]
az
-------------------------1 2
( 1 az )
z > a
1 -----------------1
1 az
z < a
( na )u [ n 1 ]
az
-------------------------1 2
( 1 az )
cos ( 0 n ) u [ n ]
1 z cos 0
-----------------------------------------------1
2
1 2z cos 0 + z
sin ( 0 n ) u [ n ]
z sin 0
-----------------------------------------------1
2
1 2z cos 0 + z
z < a
z >1
z >1
a cos ( 0 n ) u [ n ]
1 az cos 0
--------------------------------------------------------1
2 2
1 2az cos 0 + a z
z > a
z-Transform
X(z)
a sin ( 0 n ) u [ n ]
az sin 0
--------------------------------------------------------1
2 2
1 2az cos 0 + a z
n! ---------------------u[n]
k! ( n k )!
z
----------------------------1 k + 1
(1 z )
ROC
z > a
z >1
The z-transform also has various properties that are useful. The table below lists
properties for the two-sided z-transform. The one-sided z-transform properties can be
derived from the ones below by considering the signal x [ n ]u [ n ] instead of simply x [ n ] .
Table 2: Two-sided z-Transform Properties
Property
Notation
Linearity
Time Domain
z-Domain
x[ n]
x1 [ n ]
X(z)
X1 ( z )
x2 [ n ]
X2 ( z )
x 1 [ n ] + x 2 [ n ]
X 1 ( z ) + X 2 ( z )
ROC
r2 < z < r 1
ROC 1
ROC 2
At least the intersection of ROC 1 and
ROC 2
Time Shifting
x[ n k]
z-Domain Scaling
a x[ n]
X(a z)
Time Reversal
x [ n ]
z-Domain
Differentiation
nx [ n ]
z X( z)
That of X ( z ) , except
z = 0 if k > 0 and
z = if k < 0
a r2 < z < a r1
X(z )
1
1
---< z < ---r2
r1
dX ( z )
z -------------dz
r2 < z < r 1
Time Domain
x 1 [ n ]*x 2 [ n ]
z-Domain
ROC
X 1 ( z )X 2 ( z )
Multiplication
x 1 [ n ]x 2 [ n ]
1z 1
------X 1 ( v )X 2 -- v dv
v
j2
x [ n ] causal
x [ 0 ] = lim X ( z )
At least
r 1l r 2l < z < r 1u r 2u
a0
f ( x ) = ----- +
2
nx
nx
a n cos --------- + b n sin ---------
L
L
n=1
1 L
nx
a n = --- f ( x ) cos --------- dx
L L
L
1 L
nx
b n = --- f ( x ) sin --------- dx
L L
L
34.12 REFERENCES/BIBLIOGRAPHY
Spiegel, M. R., Mathematical Handbook of Formulas and Tables, Schaums Outline Series,
McGraw-Hill Book Company, 1968.
C programming - 35.1
Machine Portable, which means that it requires only small changes to run on
other computers.
C programming - 35.2
35.2 BACKGROUND
The main program is treated like a function, and thus it has the name main().
C programming - 35.3
C programming - 35.4
Major Data Types for variables and functions are (for IBM PC):
int (2 byte integer),
short (1 byte integer),
long (4 byte integer),
char (1 byte integer),
float (4 byte IEEE floating point standard),
double (8 byte IEEE floating point standard).
int, short, long, char can be modified by the addition of unsigned, and register.
An unsigned integer will not use 1 bit for number sign. A register variable will use a data
register in the microprocessor, if possible, and it will speed things up (this is only available
for integers).
C programming - 35.5
Every variable has a scope. This determines which functions are able to use that
variable. If a variable is global, then it may be used by any function. These can be modified by the addition of static, extern and auto. If a variable is defined in a function, then it
will be local to that function, and is not used by any other function. If the variable needs to
be initialized every time the subroutine is called, this is an auto type. static variables can
be used for a variable that must keep the value it had the last time the function was called.
Using extern will allow the variable types from other parts of the program to be used in a
function.
C programming - 35.6
Some basic control flow statements are while(), do-while(), for(), switch(), and
if(). A couple of example programs are given below which demonstrate all the C flow
statements.
C programming - 35.7
C programming - 35.8
C programming - 35.9
#include <filename.h> will insert the file named filename.h into the program.
The *.h extension is used to indicate a header file which contains C code to define functions and constants. This almost always includes stdio.h. As we saw before, a function
must be defined (as with the add function). We did not define printf() before we used it,
this is normally done by using #include <stdio.h> at the top of your programs. stdio.h
contains a line which says int printf();. If we needed to use a math function like y =
sin(x) we would have to also use #include <math.h>, or else the compiler would not know
what type of value that sin() is supposed to return.
#define CONSTANT TEXT will do a direct replacement of CONSTANT in the program with TEXT, before compilation. #undef CONSTANT will undefine the CONSTANT.
#ifdef, #ifndef, #if, #else and #else can be used to conditionally include parts of a
program. This is use for including and eliminating debugging lines in a program.
#define, #include, #ifdef, #ifndef, #if, #else, /* and */ are all handled by the Preprocessor, before the compiler touches the program.
C programming - 35.10
Matrices are defined as shown in the example. In C there are no limits to the
matrix size, or dimensions. Arrays may be any data type. Strings are stored as arrays of
characters.
C programming - 35.11
Pointers are a very unique feature of C. First recall that each variable uses a real
location in memory. The computer remembers where the location of that variable is, this
memory of location is called a pointer. This pointer is always hidden from the programmer, and uses it only in the background. In C, the pointer to a variable may be used. We
may use some of the operations of C to get the variable that the pointer, points to. This
allows us to deal with variables in a very powerful way.
A C compiler has three basic components: Preprocessor, First and Second Pass
C programming - 35.12
The Preprocessor
#include files
(like stdio.h)
Library files
(*.lib)
Executable Code
(*.exe)
C programming - 35.13
Use indents, spaces and blank lines, to make the program look less cluttered, and
give it a block style.
Descriptive variable names, and defined constants make the purpose of the variable obvious.
All declarations for the program should be made at the top of the program listing.
C programming - 35.14
35.6.1 How?
A program should be broken into fundamental parts (using functions for each
part) and then assembled using functions. Each function consists of programs written
using the previous simpler functions.
C programming - 35.15
Body
Engine
Wheel
Never use gotos, they are a major source of logic errors. Functions are much easier to use, once written.
Try to isolate machine specific commands (like graphics) into a few functions.
35.6.2 Why?
A top-down design allows modules to be tested as they are completed. It is much
easier to find an error in a few lines of code, than in a complete program.
When programs are complete, errors tend to be associated with modules, and are
thus much easier to locate.
C programming - 35.16
Updates to programs are much easier, when we only need to change one function.
2. Define Problem - Write out the relevant theory. This description should include
variables, calculations and figures, which are necessary for a complete solution to the
problem. From this we make a list of required data (inputs) and necessary results (output).
3. Design User Interface - The layout of the screen(s) must be done on paper. The
method of data entry must also be considered. User options and help are also considered
here. (There are numerous factors to be considered at this stage, as outlined in the course
notes.)
4. Write Flow Program - This is the main code that decides when general operations occur. This is the most abstract part of the program, and is written calling dummy
program stubs.
C programming - 35.17
5. Expand Program - The dummy stubs are now individually written as functions. These functions will call another set of dummy program stubs. This continues
until all of the stubs are completed. After the completion of any new function, the program
is compiled, tested and debugged.
6. Testing and Debugging- The program operation is tested, and checked to make
sure that it meets the objectives. If any bugs are encountered, then the program is revised,
and then retested.
Golden Rule: If you are unsure how to proceed when writing a program, then work
out the problem on paper, before you commit yourself to your programmed solution.
C programming - 35.18
35.8.1 Objectives:
The program is expected to aid the design of beams by taking basic information
about beam geometry and material, and then providing immediate feedback. The beam
will be simply supported, and be under a single point load. The program should also provide a printed report on the beam.
The basic theory for beam design is available in any good mechanical design
textbook. In this example it will not be given.
The inputs were determined to be few in number: Beam Type, Beam Material,
Beam Thickness, Beam Width, Beam Height, Beam Length, Load Position, Load Force.
The possible outputs are Cross Section Area, Weight, Axial Stiffness, Bending
Stiffness, and Beam Deflection, a visual display of Beam Geometry, a display of Beam
Deflection.
C programming - 35.19
The small number of inputs and outputs could all be displayed, and updated, on a
single screen.
The left side of the screen was for inputs, the right side for outputs.
The screen is divided into regions for input(2), input display and prompts(1),
Beam Cross section(3), Numerical Results(4), and Beam Deflection(5).
35.8.3.2 - Input:
Current Inputs were indicated by placing a box around the item on the display(1).
The cursor keys could be used to cursor the input selector up or down.
C programming - 35.20
and <RETURN>. In the spirit of robustness it was decided to screen all other keys.
35.8.3.3 - Output:
Equations, calculations, material types, and other relevant information were
obtained from a text.
Proper textual descriptions were used to ensure clarity for the user.
For a printed report, screen information would be printed to a printer, with the
prompt area replaced with the date and time.
35.8.3.4 - Help:
A special set of help information was needed. It was decided to ensure that the
screen always displays all information necessary(2).
C programming - 35.21
35.8.3.6 - Miscellaneous:
The screen was expressed in normalized coordinates by most sub-routines.
C programming - 35.22
main()
/*
* EXECUTIVE CONTROL LEVEL
*
* This is the main terminal point between the
* various stages of setup, input, revision
* and termination.
*
* January 29th, 1989.
*/
{
static int error;
if((error = setup()) != ERROR) {
screen(NEW);
screen(UPDATE);
while((error = input()) != DONE) {
if(error == REVISED) {
screen(NEW);
screen(UPDATE);
}
}
error = NO_ERROR;
}
kill();
if(error == ERROR) {
printf(EGA Graphics Driver Not Installed);
}
C programming - 35.23
The routines were written in a top down fashion, in a time of about 30 hours.
These routines are listed below.
picture() - draws the beam cross section and deflection of beam. For the sake
of speed, this section will use low level commands.
input() - A function which controls the main input loop for numbers, con-
Condition and error flags were used to skip unnecessary operations, and thus
speed up response. A response of more than 0.5 seconds will result in loss of attention by
the user.
C programming - 35.24
Low Level
Specific Subroutines
High Level
Executive Subroutine
Top
Down
main()
screen()
box()
setup()
input()
printer()
kill()
picture()
enter()
draw_line()
text()
calculations()
printes()
C programming - 35.25
The testing and debugging was very fast, with only realignment of graphics being
required. This took a couple of hours.
35.8.7 Documentation
The theory for beam design was given for the reference of any program user, who
wanted to verify the theory, and possible use it.
A manual was given which described key layouts, screen layout, basic sequence
of operations, inputs and outputs.
A walk through manual was given. This allowed the user to follow an example
which displayed all aspects of the program.
C programming - 35.26
1. What are the basic components of a C compiler, and what do they do?
C programming - 35.27
2. You have been asked to design a CAD program which will choose a bolt and a
nut to hold two pieces of sheet metal together. Each piece of sheet metal will have a hole
drilled in it that is the size of the screw. You are required to consider that the two pieces are
experiencing a single force. State your assumptions about the problem, then describe how
you would produce this program with a Top Down design.
4. Describe some of the reasons for Using Top-Down Design, and how to do it.
units - 36.1
d y = 5ft
d x = 10m
Find the distance d,
dx + dy
d =
( 10m ) + ( 5ft )
d =
100m + 25ft
d =
2
2 0.3048m 2
100m + 25ft ---------------------
1ft
d =
m
100m + 25ft ( 0.092903 ) ------2
ft
multiply by 1
d =
100m + 25 ( 0.092903 )m
d =
102.32m = 10.12m
units - 36.2
units - 36.3
1 furlong = 660 ft
1 lightyear = 9.460528e15 m
1 parsec = 3.085678e16 m
Area
1 acre = 43,559.66 ft2
1 Hectare (ha) = 10,000 m2
1 Hectare (ha) = 10,000 m2
1 Hectare (ha) = 10,000 m2
1 Hectare (ha) = 10,000 m2
Velocity
1 mph = 0.8689762 knot
Angle
1 rev = 2PI radians = 360 degrees = 400 gradians
1 degree = 60 minutes
1 minute = 60 seconds
Volume
1 US gallon = 231 in3
1 CC = 1 cm3
1 IMP gallon = 277.274 in3
1 barrel = 31 IMP gal. = 31.5 US gal.
1 US gal. = 3.785 l = 4 quarts = 8 pints = 16 cups
1 liter (l) = 0.001 m3 = 2.1 pints (pt) = 1.06 quarts (qt) = 0.26 gallons (gal)
1 qt (quart) = 0.9464 l
1 cup (c) = 0.2365882 l = 8 USoz
1 US oz = 8 U.S. drams = 456.0129 drops = 480 US minim = 1.040842
IMP oz
= 2 tablespoons = 6 teaspoons
1 IMP gal. = 1.201 U.S. gal.
1 US pint = 16 US oz
1 IMP pint = 20 IMP oz
1tablespoon = 0.5 oz.
1 bushel = 32 quarts
1 peck = 8 quarts
Force/Mass
1 N (newton) = 1 kgm/s2 = 100,000 dyne
1 dyne = 2.248*10-6 lb. (pound)
1 kg = 9.81 N (on earth surface) = 2.2046 lb
1lb = 16 oz. (ounce) = 4.448N
units - 36.4
units - 36.5
1 ev = 1.60219*10-19 J
1 erg = 10-7 J
Temperature
F = [(C*9)/5]+32, C = Celsius (Centigrade), F = Fahrenheit
K = Kelvin
Rankine (R) = F - 459.666
0.252 calories = 1 BTU (British Thermal Unit)
-273.2 C = -459.7 F = 0 K = 0 R = absolute zero
0 C = 32 F = 273.3 K = 491.7 R = Water Freezes
100C = 212F = 373.3 K = 671.7 R = Water Boils (1 atm. pressure)
1 therm = 100,000 BTU
Mathematical
Time
1 Hz (hertz) = 1 s-1
1 year = 365 days = 52 weeks = 12 months
1 leap year = 366 days
1 day = 24 hours
1 fortnight = 14 days
1 hour = 60 min.
1 min = 60 seconds
1 millenium = 1000 years
1 century = 100 years
1 decade = 10 years
Physical Constants
R = 1.987 cal/mole K = ideal gas law constant
K = Boltzmanns constant = 1.3x10-16 erg/K = 1.3x10-23 J/K
h = Plancks constant = 6.62x10-27 erg-sec = 6.62x10-34 J.sec
Avagadros number = 6.02x1023 atoms/atomic weight
density of water = 1 g/cm3
electron charge = 1.60x10-19 coul.
electron rest mass = 9.11*10**-31 kg
proton rest mass = 1.67*10**-27 kg
units - 36.6
decimal
hexadecimal
binary
ASCII
decimal
hexadecimal
binary
ASCII
units - 36.7
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
10
11
12
13
14
15
16
17
18
19
1A
1B
1C
1D
1E
1F
00000000
00000001
00000010
00000011
00000100
00000101
00000110
00000111
00001000
00001001
00001010
00001011
00001100
00001101
00001110
00001111
00010000
00010001
00010010
00010011
00010100
00010101
00010110
00010111
00011000
00011001
00011010
00011011
00011100
00011101
00011110
00011111
NUL
SOH
STX
ETX
EOT
ENQ
ACK
BEL
BS
HT
LF
VT
FF
CR
S0
S1
DLE
DC1
DC2
DC3
DC4
NAK
SYN
ETB
CAN
EM
SUB
ESC
FS
GS
RS
US
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
20
21
22
23
24
25
26
27
28
29
2A
2B
2C
2D
2E
2F
30
31
32
33
34
35
36
37
38
39
3A
3B
3C
3D
3E
3F
00100000
00100001
00100010
00100011
00100100
00100101
00100110
00100111
00101000
00101001
00101010
00101011
00101100
00101101
00101110
00101111
00110000
00110001
00110010
00110011
00110100
00110101
00110110
00110111
00111000
00111001
00111010
00111011
00111100
00111101
00111110
00111111
space
!
#
$
%
&
(
)
*
+
,
.
/
0
1
2
3
4
5
6
7
8
9
:
;
<
=
>
?
decimal
hexadecimal
binary
ASCII
decimal
hexadecimal
binary
ASCII
units - 36.8
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
40
41
42
43
44
45
46
47
48
49
4A
4B
4C
4D
4E
4F
50
51
52
53
54
55
56
57
58
59
5A
5B
5C
5D
5E
5F
01000000
01000001
01000010
01000011
01000100
01000101
01000110
01000111
01001000
01001001
01001010
01001011
01001100
01001101
01001110
01001111
01010000
01010001
01010010
01010011
01010100
01010101
01010110
01010111
01011000
01011001
01011010
01011011
01011100
01011101
01011110
01011111
@
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
[
yen
]
^
_
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
60
61
62
63
64
65
66
67
68
69
6A
6B
6C
6D
6E
6F
70
71
72
73
74
75
76
77
78
79
7A
7B
7C
7D
7E
7F
01100000
01100001
01100010
01100011
01100100
01100101
01100110
01100111
01101000
01101001
01101010
01101011
01101100
01101101
01101110
01101111
01110000
01110001
01110010
01110011
01110100
01110101
01110110
01110111
01111000
01111001
01111010
01111011
01111100
01111101
01111110
01111111
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
s
t
u
v
w
x
y
z
{
|
}
r arr.
l arr.
36.5 G-CODES
Note that G and M codes not universal standards, and may vary between
machines. In any case of doubt, the manuals for the machine should be checked.
units - 36.9
A basic list of G operation codes is given below. These direct motion of the
tool.
G00 - Rapid move (not cutting)
G01 - Linear move
G02 - Clockwise circular motion
G03 - Counterclockwise circular motion
G04 - Dwell
G05 - Pause (for operator intervention)
G08 - Acceleration
G09 - Deceleration
G17 - x-y plane for circular interpolation
G18 - z-x plane for circular interpolation
G19 - y-z plane for circular interpolation
G20 - turning cycle or inch data specification
G21 - thread cutting cycle or metric data specification
G24 - face turning cycle
G25 - wait for input #1 to go low (Prolight Mill)
G26 - wait for input #1 to go high (Prolight Mill)
G28 - return to reference point
G29 - return from reference point
G31 - Stop on input (INROB1 is high) (Prolight Mill)
G33-35 - thread cutting functions (Emco Lathe)
G35 - wait for input #2 to go low (Prolight Mill)
G36 - wait for input #2 to go high (Prolight Mill)
G40 - cutter compensation cancel
G41 - cutter compensation to the left
G42 - cutter compensation to the right
G43 - tool length compensation, positive
G44 - tool length compensation, negative
G50 - Preset position
G70 - set inch based units or finishing cycle
G71 - set metric units or stock removal
G72 - indicate finishing cycle (EMCO Lathe)
G72 - 3D circular interpolation clockwise (Prolight Mill)
G73 - turning cycle contour (EMCO Lathe)
G73 - 3D circular interpolation counter clockwise (Prolight Mill)
G74 - facing cycle contour (Emco Lathe)
G74.1 - disable 360 deg arcs (Prolight Mill)
G75 - pattern repeating (Emco Lathe)
G75.1 - enable 360 degree arcs (Prolight Mill)
G76 - deep hole drilling, cut cycle in z-axis
G77 - cut-in cycle in x-axis
G78 - multiple threading cycle
G80 - fixed cycle cancel
units - 36.10
units - 36.11
materials - 37.1
Element
copper
iron
aluminum
tungsten
zinc
silicon
carbon
titanium
Atomic #
29
26
13
74
30
14
6
22
Atomic weight
63.57
55.85
26.97
183.85
65.37
28.09
12.01
47.9
density g/cm3
8.96
7.86
2.67
19.3
7.13
2.33
2.1
4.51
Valances Hw
1/2
2/3
3
6/8
2
4
2/4
3/4
F = 96,500 coulombs
26-30
0.33
35-500
100-550
1-45
2014-T6
2.8
73
36-41
0.33
410
480
13
6061-T6
2.8
70
26
0.33
270
310
17
7075-T6
2.7
72
26
0.33
480
550
11
1100-h14
2024-T6
5456-h116
thermal expansion
10-6/deg C
70-79
2.6-2.8
typical
poisson
(GPa)
(GPa)
aluminum
Density
(Mg/m3)
Type
Material
Table 1:
23
materials - 37.2
Brass
typical
thermal expansion
10-6/deg C
poisson
G
(GPa)
E
(GPa)
Density
(Mg/m3)
Type
Material
Table 1:
8.4-8.6
96-110
36-41
0.34
70-550
200-620
4-60
19.121.2
18-21
Yellow Brass
cold rolled
annealed
Red Brass
cold rolled
annealed
Bronze
typical
8.2-8.8
96-120
36-44
0.34
82-690
200-830
5-60
Cast Iron
typical
7.0-7.4
83-170
32-69
0.2-0.3
120
69-480
0-1
Concrete
typical
2.3
17-31
na
0.1-0.2
reinforced
2.4
17-31
na
0.1-0.2
lightweight
1.1-1.8
17-31
na
0.1-0.2
typical
8.9
110-120
40-47
0.330.36
2.4-2.8
48-83
19-35
0.170.27
Copper
55-760
3401400
9.9-12
10-70
7-14
230-830
4-50
16.617.6
30-1000
5-11
annealed
hard drawn
Glass
typical
Plate
70
Fibers
700020000
Magnesium
typ. alloys
1.761.83
41-45
15-17
0.35
80-280
140-340
2-20
26.128.8
Monel
67%Ni,30%
Cu
8.8
170
66
0.32
1701100
4501200
2-50
14
8.8
210
80
0.31
100-620
310-760
2-50
13
Nickel
materials - 37.3
thermal expansion
10-6/deg C
poisson
G
(GPa)
E
(GPa)
Density
(Mg/m3)
Type
Material
Table 1:
Plastics
Nylon
.88-1.1
2.1-3.4
na
0.4
40-80
20-100
70-140
Polyethylene
.96-1.4
0.7-1.4
na
0.4
7-28
15-300
140-290
Rock
Rubber
5-9
Granite
2.6-2.9
40-100
na
0.2-0.3
50280
Marble
2.6-2.9
40-100
na
0.2-0.3
50280
Quartz
2.6-2.9
40-100
na
0.2-0.3
50280
Limestone
2.0-2.9
20-70
na
0.2-0.3
20200
Sandstone
2.0-2.9
20-70
na
0.2-0.3
20200
typical
.96-1.3
0.00070.004
0.00020.001
0.450.50
190-210
75-80
0.270.30
Sand,soil,gra
vel
1.2-2.2
Steel
7.85
1-7
7-20
100-800
130-200
10-18
cold rolled
annealed
high strength
3401000
5501200
5-25
machine
340-700
550-860
5-25
spring
4001600
7001900
3-15
stainless
280-700
4001000
5-40
tool
520
900
structural
200-700
340-830
10-40
astm-a48
14
12
materials - 37.4
thermal expansion
10-6/deg C
G
(GPa)
(MPa)
E
(GPa)
poisson
Density
(Mg/m3)
Type
Material
Table 1:
astm-a47
astm-a36
250
400
30
12
astm-a572
340
500
20
12
astm-a514
700
830
15
12
wire
2801000
5501400
5-40
astm-a5242
astm-a441
Stainless
Steel
aisi 302
Titanium
typ. alloys
Tungsten
Water
17
4.5
100-120
39-44
0.33
1.9
340-380
140-160
0.2
7601000
9001200
10
8.1-11
14004000
0-4
4.3
fresh
1.0
sea
1.02
Douglas fir
.48-.56
11-13
30-50
50-80
30-50
40-70
Oak
.64-.72
11-12
40-60
50-100
30-40
30-50
Southern
pine
.56-.64
11-14
40-60
50-100
30-50
40-70
Wood (dry)
materials - 37.5
Axial/Normal Stress/Strain
P
= E = --A
Shear Stress/Strain
= G
Poissons ratio
lateral
= -------------------------- longitudinal
PL
= L = ------AE
xy
xy = ------G
E = 2G ( 1 + )
x
y
z
x = ----- ----- ----E
E
E
Torsion
Tc
max = -----J
max
max = ---------G
4
TL
= ------JG
r
J = -------2
(cylinder)
ten = comp =
4 4
J = --- ( r o r i )
2
(hollow tube)
= --2
P = 2fT
materials - 37.6
Beams
bh
I = -------- (for rectangle)
12
My
= -------I
r
I = -------- (for circle)
2
VQ
= -------Ib
L =
c
max = --
My
= -------I
EI
= -----M
EI
P < ----------2 2
k L
Buckling
d/2
3.0
d/2
2.5
2.0
1.5
r/d
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
materials - 37.7
2.5
D
d
P
2.0
D/d = 2.0
D/d = 1.5
D/d = 1.25
D/d = 1.1
1.5
1.0
r/d
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
materials - 37.8
K
T
r
d
D
2.5
2.0
1.5
D/d = 2.0
D/d = 1.33
D/d = 1.2
D/d = 1.1
1.0
r/d
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
bibliography - 38.1
38. BIBLIOGRAPHY
38.1 TEXTBOOKS
38.1.2 VandeVegte
Citation: VandeVegte, John, "Feedback Control Systems", Prentice Hall, Second
Edition, 1990.
Topics/Chapters:
Linearized dynamic models
Transfer function models of physical systems
Modeling of feedback systems and controllers
Transient performance and the s-plane
The perfoarmance and dynamic compensation of feedback systems
The root-locus method
bibliography - 38.2
Topics:
Objectives:
ea = KE
Te = KT ia
2
T e T L C d = JD
2
T e = JD + C d + T L
v a = r a i a + l a Di a + e a
va = ia ( ra + la D ) + e a
va = ia ( ra + la D ) + K E
Te
v a = ------ ( r a + l a D ) + K E
KT
2
( JD + C d + T L )
v a = ------------------------------------------------ ( r a + l a D ) + K E
KT
2
( JD + C d + T L ) ( r a + l a D ) + K E K T = K T v a
2
Jr a D + C d r a + T L r a + Jla DD + l a DC d + l a DT L + K E K T = K T v a
( Jla ) + ( Jr a + l a C d ) + ( K E K T ) + ( C d r a ) = v a ( K T ) + T L ( r a ) + T L ( l a )
0 ( Jl a ) + 0 ( Jr a + l a C d ) + ( K E K T ) + ( C d r a ) = v a ( K T ) + T L ( r a ) + 0 ( l a )
2
( KE KT ) + ( Cd ra ) = v a ( K T ) + T L ( r a )
2
( C d ra ) + ( KE KT ) + ( TL ( ra ) va ( KT ) ) = 0
At lower steady state speeds (little drag),
0 ( Jl a ) + 0 ( Jr a + l a C d ) + ( K E K T ) + 0 = v a ( K T ) + T L ( r a ) + 0 ( l a )
( KE KT ) = va ( KT ) + TL ( ra )
ra
( K E ) + T L ------ = v a
KT
The static friction limit,
0 ( Jl a ) + 0 ( Jr a + l a C d ) + 0 ( K E K T ) + 0 ( C d r a ) = v a ( K T ) + T L ( r a ) + 0 ( l a )
KT
T L = v a ------
ra
2
+ A + B + C = v a D + T L E + T L F
TL
vs
1 -------------------La D + ra
Te
ia
Ka
Ka
where,
K a = motor speed constant
L a = motor armature inductance
r a = motor armature resistance
J = inertia of rotor
B m = damping of rotor
v s = voltage applied to armature
i a = armature current
T e = effective motor torque
T L = load torque
r = motor speed
1
-------------------JD + Bm
Motor parameters can be measured by applying a voltage and holding (stalling) the shaft.
The stalled motor torque, and the current are measured. When stalled, the load and effective torques are equal, and the feedback loop is zero, allowing the simplified system
block diagram. This allows the calculation of the motor speed coefficient directly.
vs
1
--------------------La D + ra
ia
Ka
TL
T
K a = -----Lia
The resistance of the windings can be found by direct measurement with a DMM, but can
also be verified by using the steady state current and voltages.
i
1 1
1
---a- = -----------------= --------------------------- = ---vs
La s + ra
L a ( j0 ) + r a
ra
v
r a = ----s
ia
The inductance of the windings can be found using the time constant - phase shift
????????????????
i
1 ---a- = -----------------vs
La s + ra
XXXXXXXXXXXXXXX
If the motor is unloaded (TL = 0) the following differential equation can be derived. If the
system is allowed to reach steady state, the derivatives of motor speed will become
zero.
vs
1 -------------------La D + ra
ia
Te
Ka
1
-------------------JD + Bm
Ka
D r ( L a J ) + D r ( L a B m + Jr a ) + r ( r a Bm + K a ) = Vs K a
2
( 0 ) ( L a J ) + ( 0 ) ( L a B m + Jr a ) + r ( r a B m + K a ) = V s K a
2
r ( ra Bm + K a ) = Vs Ka
2
Bm
Vs K a K a
= ------------ -----ra r r a
D r ( L a J ) + D r ( L a B m + Jr a ) + r ( r a B m + K a ) = V s K a
TL
vs
1 -------------------La D + ra
ia
Ka
Ka
1
i a = ( v s K a r ) --------------------La D + ra
Di a ( L a ) + i a ( r a ) = v s K a r
K a
ra
1
Di a = v s ----- + r --------- + i a --------
La
La
La
1
r = ( K a i a T L ) -------------------JD + B m
D r ( J ) + r ( B m ) = K a i a T L
K
B
1
D r = i a -----a- + T L ------ + r ---------m-
J
J
J
d- i a
---=
dt
r
r K
-------a- --------aLa La
ia
Ka B m r
------ ---------J
J
Te
1---0
La
vs
1 T
0 ------ L
J
1
-------------------JD + Bm
ra Ka
-------- --------La La
ia
Ka B m r
------ ---------J
J
1---0
La
vs
1 T
0 ------ L
J
1
E = ----La
ra
A = -------La
i
vs
= A B a + E 0
C D r
0 F TL
La = E
r a = AL a = AE
Ka
B = --------La
K a = BL a = BE
Ka
C = -----J
Ka
BE
J = ------ = ---------C
C
Bm
D = ---------J
BED
B m = --------------C
1--D
1--D
Ts
1e
G h ( s ) = -----------------s
t
0 T
39.7 MISC
- controller goals
stabilize - to make the system stable
regulate - to control the system about a central design value
follow - to track an input/feedback signal
disturbance rejection - to minimize the effects of noise
- state diagrams can be used with multiple input/multiple output (MIMO) systems
- state based controller designs,
state feedback - in an ideal controller the full state variables can be measured
observer-estimator - determines the system state using limited data.
- Linear Time Invariant (LTI) systems are linear, having constant coefficients. For
example, state coefficient matrices are all constant values
- Linear systems allow superposition
- leaf springs are non-linear and have a cubic relationship to displacement
- if a system input (setpoint) is zero, or not used, then the control system is called a
regulator.
- stepper motor motion plans should not be approximate, the stepper motor should
follow the path, although trapezoidal profiles are more common.
M
-1
1
G M = -------OC
C
c
A
0dB
25
0.5dB
-355
-5
20
-10
-350
15
3dB
10
-1dB
5
-330
-30
dB
12dB
0
-300
-5
-60
-9dB
-10
-270
-90
-15
-240
-20
-280
-260
-240
-220
-210
-200
-180
-120
-150
-180
-160
phase (deg)
-18dB
-140
-120
-100
The gain and phase margins are measured from the plot as shown below,
-80
180
M
0dB
GM
Where,
M = phase margin
G M = gain margin
These correlate to values on the Bode plot. To find the approximate gain/phase
margins with a Bode plot, locate the point where the gain is 180 to find the gain margin,
and find the phase margin where the gain is 0dB.
In unstable systems the function will circle the center point of the plot.
polynomial
order
roots
1
2
3
4
5
6
7
8
9
10
-4.6200
-4.05302.3400j
-5.0093,-3.96683.7845j
-4.01565.0723j,-5.52811.6553j
-6.4480,-4.11046.3142j,-5.92683.0813j
-4.21697.5300j,-6.26134.4018j,-7.12051.4540j
-8.0271,-4.33618.7519j, -6.57145.6786j,-7.68242.8081j
-4.45549.9715j,-6.85546.9278j,-8.16824.1057j,-8.76931.3616j
-9.6585,-4.569611.1838j,-7.11458.1557j,-8.56925.3655j,-9.40132.6655j
-4.683512.4033,-7.36099.3777j,-8.98986.6057j,-9.96573.9342j,10.42781.3071j
- the roots are used to construct the charateristic equation of the system.
- To increase, or decrease the reposnse time the root values are multiplied by the
desired settling time.
39.11 ITAE
- The Integral of Time Absolute Error (ITAE) integral equation becomes large for
errors that persist for a long time.
J ITAE =
t e ( t ) dt
0
- is used to select poles to minimize the integral value when selecting a characteristics equation.
cos =
The proof begins with a taylor series expansion of a function that has a continuous differential. The second term in the expression is the higher order terms. There is an
assumption that the system starts at equillibrium, so there is no initial value. The
result is a linear approximation.
d
f
f
----- x = f ( 0 ) +
x
+
f
(
t
)
=
h.o.t.
x x = 0
x x = 0 x = Ax
dt
f
A =
x x = 0
df
f
---x =
x +
u + f h.o.t. ( t ) = Ax + Bu
u x = 0, u = 0
dt
x x = 0, u = 0
f
A =
x x = 0, u = 0
f
B =
u x = 0, u = 0
A linearized model of the control law can be written and the linearized model rewritten.
u
u
x = Gx
x x = 0
d---x = Ax + BGx = ( A + BG )x
dt
39.14 XXXXX
- ea.
39.15 XXXXX
- ea.
39.16 XXXXX
- ea.
39.17 XXXXX
- ea.
39.18 XXXXX
- ea.
39.19 XXXXX
- ea.
39.20 XXXXX
- ea.
39.21 SUMMARY
39.25 REFERENCES
39.26 BIBLIOGRAPHY
How, J, "16.31 Feedback Control Course Notes", MIT Opencourseware website.