MEM 355 Performance Enhancement of Dynamical Systems
Introduction to Control System Design
Harry G. Kwatny
Department of Mechanical Engineering & Mechanics Drexel University
Outline
Course practical information Control: open loop and closed loop Short history of control Contemporary applications Technology drivers Summary
What is the course content? What is control? Why should an ME care? Why all the math?
Practical Information
Lectures: Tues & Thurs 2-3:20 pm URL: http://www.pages.drexel.edu/faculty/hgk22.html Text Book: Nise, Control Systems Engineering, 5th ed., J. Wiley & Sons, 2008. Software: The MathWorks, Inc. The Student Edition of MATLAB, Version +Control Toolbox. (15 seats in UG lab, CAD lab). Tutorial @ http://www.engin.umich.edu/group/ctm/basic/basic.html
TA: Jean-Ettiene Dongmo Tel: 215-895-6767, Rm: 4-154, e-mail:[email protected]
Grading:
Homework (6): 20% Midterm (in class): 30% Project (take home): 50%
What you should know going in
Basics of Laplace transform Concepts of state space and transfer function models of a linear system. The meaning of poles & zeros The frequency transfer function and Bode Plots Block diagram manipulation Eigenvalues & eigenvectors, modal analysis and similarity transformations. Stability and Routh table. Basic ability to use MATLAB.
What you should know going out
Understand why automatic control is useful for a mechanical engineer Recognize the value of integrated control and process design Understand the key concepts of control system design Be able to solve simple control problems Recognize difficult control problems Know relevant mathematical theory Have competence in using computational tools
Specific Goals
Define the control system design problem and develop a preliminary appreciation of the tradeoffs involved and requirements for robust stability and performance. Develop concepts and tools for ultimate state error analysis. Develop the relationship between time domain and frequency domain performance specifications, e.g, rise time, overshoot, settling time, sensitivity function and bandwidth. Develop frequency domain design methods, including: the root locus method, Nyquist & Bode methods, and stability margins. Provide an introduction to state space design: controllability and observability, pole placement, design via the separation principle (time permitting). Emphasize computational methods using MATLAB.
What is Control?
Control refers to the manipulation of the inputs to a physical system in order to cause desirable behavior.
Cause output variables to track desired values Impose desirable dynamical behavior, e.g., stabilize an unstable system
Open loop (feedforward) control
Exploit knowledge of system behavior to compute necessary inputs Requires accurate model of system
Closed loop (feedback, active) control
Process information from sensors to derive appropriate inputs Allows compensation for model uncertainty, disturbances, noise Alters system dynamics
Familiar Examples
Household Temperature Control Cruise Control Traction Control Electronic Stabilization Airplane Autopilot
What do Control Engineers Do?
System Design/System Integration
Participate in defining system/subsystem requirements and specifications Develop subsystem/component specifications including cyberstructure Participate in component selection/design Develop math models and simulations of components/subsystems/system Design/implement control systems Participate in testing/validation/verification
Open & Closed Loop Control
Control computer
Feedforward does not alter plant dynamics. Feedback does.
The Magic of Feedback
The adjustment of system inputs based on the observation of its outputs Feedback is a universal strategy to cope with uncertainty In engineering we use feedback: Cause a system to behave as desired Keep variables constant Stabilize unstable system Reduce effects of disturbances Minimize the effect of component variations Another alternative for designers
Origins of Control Engineering
Clocks (escapement) 1200-1400 Windmills 1787 Steam Engines (Watt) 1788 Maxwell ~ Governors 1868 Water Turbines 1893 Wright brothers ~ Airplane 1903 Sperry ~ Autopilot (Gyro) 1914 Minorsky ~ Ship steering 1922 Black ~ Feedback amplifier 1928 Ivanoff ~ Temperature regulation 1934 First real control system analysis. First journal article. Invention of new control paradigm.
Wilber Wright 1901
We know how to construct airplanes. Men also know how to build engines. The inability to balance and steer still confronts students of the flying problem. When this one feature has been worked out, the age of flying will have arrived, for all other difficulties are of minor importance.
Contemporary Applications
Widespread use of automatic control in many fields
Power generation Power transmission Process control Discrete manufacturing Robotics Communications Automotive Buildings Aerospace Medicine Marine Engineering Computers Instrumentation Mechatronics Materials Physics Biology Economics
There is a unified framework of theory, design methods and computer tools that cut across fields of application.
Examples
Flight control systems
Commercial & military fly-by-wire Autopilot, auto-landing UAV
Automotive
Mercedes Benz SLR
Robotics
Precision positioning in manufacturing Remote space/sea environments Minimally-invasive surgery RPVs for surveillance, search and rescue
Engine Transmission Cruise, climate control ABS, Traction control, ESP Active suspension
Power plants
Various temps/pressures Power output Emissions control
Heating, ventilation, air conditioning (HVAC)
Examples
Materials processing
Rapid thermal processing Vapor deposition
Smart engines
Compression systems stall, surge, flutter control Combustion systems lean air/fuel ratio for low emissions, improved efficiency
Noise and vibration control
Active mounts Speaker systems
Intelligent vehicle highway systems
platooning for high speed, high density travel Automatic merge Obstacle avoidance
GE 7000
Evolution of the Control Discipline
Classical control
Frequency-domain based tools for linear systems Mainly useful for single-input single-output (SISO) systems WWII years saw 1st application of optimal control Still the main tools used in practice
1940
Modern control
State space approach for linear systems Useful for SISO and multi-input multi-output (MIMO) systems Relies on linear algebra computations rather than Laplace transform Performance and robustness measures not always explicit Just in time for space exploration
1960
Optimal control
Find the input that optimizes some objective function (e.g., min fuel, min time) Used for both open loop and closed loop design
1970 1980 1990 2000
Robust control
Generalizes classical control to MIMO case Enabled by modern control development Geometric theory of affine systems, variable structure Self-tuning and adaptive control
Nonlinear, adaptive, hybrid Discrete Event & Hybrid Systems
Key Technology Trends
Computation/microprocessors
Cheap and powerful microprocessors opened the door to widespread control applications from 1970s onward
Sensors and actuators
Sensors continue to get smaller, cheaper, faster Macro/micro scale actuation evolving (power electronics, piezo-electric, EM-rheological fluids)
Communications and networking
Networks replacing point-to-point communication in large systems (e.g., electric power systems) and small (e.g. automotive)
Active Control in Automobiles
A typical automobile has 200-300 feedback controllers. Here are a few examples in a contemporary Mercedes.
Cruise Control ABC-active body control ABS-anti-lock braking system ASR acceleration skid control ESP electronic stabilization program SBC sensotronic brake control BAS brake assist system Proximity controlled cruising
http://www.mercedes-benz.com/e/innovation/rd/sicherheitspecial/default.htm
The Mercedes A-class
Automatic control gives extra freedom to the designer
Unstable behavior improved by Electronic Stabilization Program (ESP) Now standard on all MB models to be introduced on Jeep GC
Mercedes Benz Electronic Stabilization Package
Active Body Control
ABC continuously matches the stiffness and damping characteristics to current driving conditions. It is possible, for example, to compensate for the rolling motion of the body when taking a bend in the road. Hydraulic cylinders in series with the coil springs, generate forces that counteract wheel load. This is performed with sensors that measure yaw rate, longitudinal and transverse acceleration, vertical acceleration.
Biology/Biomechanics
Feedback governs how we grow, respond to stress and challenge. Feedback regulates factors such as body temperature, blood pressure, and cholesterol level. Feedback makes it possible for us to stand upright. Feedback enables locomotion. Feedback operates at every level, from the interaction of proteins in cells to the interaction of organisms in complex ecologies. Feedback control is used to design drug treatment strategies for diseases like HIV/Aids, Cancer
Biologically inspired control
Research Applications in MEM
Automotive Aircraft/Flight Safety Power Plants Robotics Autonomous Vehicles Mechatronics Biology/Biomechanics Electric Power Systems
State Space & Transfer Function Models ~ remarks on notation
East Coast Notation MATLAB West Coast Notation
Ax + Bu x= y= Cx + Du
Fx + Gu x= y= Hx + Ju
State space models are also referred to as time-domain models
Y ( s ) = G ( s )U ( s ) ,
Transfer Function
U ( s ) = L u ( t )
Transfer function models are also Referred to as frequency domain models.
Y ( s ) = L y ( t )
Summary
Course content. What is a control system?
Open loop/closed loop (feedforward/feedback)
Why is control relevant to ME?
Applications! Applications! Applications!
Why so much math?
Abstraction to accommodate many applications in a common framework Explicit design approaches to meet (optimize) specific performance goals.
