The University of British Columbia



 
 
 
 
 
 
 

ROBUST MULTIVARIABLE CONTROL

Postgraduate Control Engineering Course: EECE 571 (3 Credits)


The goal of this course is to provide students with an understanding of the important problems associated with multivariable feedback control. The use of advanced mathematics will be introduced in a control engineering framework, having as final goal the design and analysis of robust multivariable controllers. The interlocking nature of theory and application is a central theme, and this course contains an overview of the existing approaches and tools available to a working control engineer when faced with a multivariable control problem.

Extra effort will be made to address the needs of students from industry. The lectures are scheduled outside of regular work hours in order to accommodate industrial participation.

DETAILS

Course Leaders:

Dr. Mihai Huzmezan
Assistant Professor 
Electrical and Computer Engineering , UBC huzmezan@ece.ubc.ca
Adjunct Professor 
Electrical and Computer Engineering, UBC Honeywell Industrial Control greg.stewart@honeywell.com

Timing and Place: September 2001, 5:30 pm every Monday, MCLD 228, UBC
(during the first lecture a vote will be held with respect to the best day of the week for all students enrolled)

Enrollment for UBC students

In order to take this course you will need to enroll for EECE 572D (106) Cat #72285 (the course) and at the same time for EECE 571D (106) Cat #62879 (the project). In this fashion you will still have the opportunity to take another EECE 571 course that accounts for 3 credits.

Enrollment for non UBC students

In order to take this course you will need to enroll for EECE 571L (109) Cat #35419. Please refer to the UBC registration procedures advertised in the UBC 2001 calendar.

Enrollment for IEEE members

The course, as taken through the IEEE as non-credit, will be $375CDN for IEEE members and $450CDN for non-members. A cheque or money order should be made payable to IEEE Vancouver Section and forwarded to the attention of:

Bruce Sunga BChydro A01 6911 Southpoint Drive Burnaby, BC, V3N 4X8 Phone: 604-528-2383 Email: bruce.sunga@bchydro.com

In order to take this course you will need to enroll for EECE 571L (109) Cat #35419

Prerequisites: It is recommended that the candidates have basic knowledge on Linear Algebra, Modeling, State Space Methods, Continuous and Discrete Time Systems and Classical Control Systems Design.

On completion of the module you will be able to:

COURSE WORK ? ASSESSMENT
 LECTURE SYLLABUS
Analysis of multivariable systems (7 Lectures)
 
  1. Classical feedback control in the multivariable case.
  1. Introduction to multivariable control.
  1. Elements of linear systems theory .
  1. Limitations on performance in SISO and MIMO systems.
  1. Uncertainty and robustness for SISO systems .
  1. Model uncertainty and robust stability for MIMO systems .
  1. Structured uncertainty and robust performance .

Design of multivariable systems (5 Lectures)


  1. H-infinity control, problem formulation, assumptions, solution.
  1. H-infinity control, example of a loop shaping controller design for a wind tunnel model.
  1. Project 1 presentation, assignments discussions, student performance.
  1. An industrial example: two dimensional loop shaping for cross-directional control of a paper machine.
  1. Project 2 presentation, assignments discussions, student performance.

* Students can choose between either project 1 or project 2.

Project objectives:

These projects take students through the simulate/analyze/design/test cycle for an industrial control system. Also exposing students to state-of-the-art software for control. Students can work in pairs. All pairs will produce similar simulation models, but each pair will have the option of using a different approach for designing the control system. Modern simulation and analysis software such as Matlab and Simulink will be used throughout.

The project will involve the modeling and MIMO robust control design of an 'evaporator', a process used in many industries. A simulation model will be built and tested. Following this an H -infinity loop shaping controller will be designed for the process, and its performance checked by simulating its operation with the evaporator. Modern simulation and analysis software such as Matlab and Simulink will be used throughout.
This project will involve the modeling and control of an industrial cross-directional control system for a paper machine. The challenge in controlling such a system lies in its large scale – up to 250 actuators control a single paper property. The second challenge is the ill-conditioned nature of the process – in many industrial systems over half of the modes of the process are uncontrollable. As such, these systems are very sensitive to model uncertainty, and require the design of a robust controller.
REFERENCES

Required text (available at UBC bookstore):
 

The link to the book's web page includes:

Recommended Books and Articles:
 
K. Zhou, J.C. Doyle, and K. Glover, Robust and Optimal Control, Prentice Hall, 1996.
G.J. Balas et. al., Mu-Analysis and Synthesis Toolbox, The MathWorks Inc., 1993 .
The Mu-Tools Software (30 day trial) directly from MathWorks
J.M. Maciejowski, Multivariable Feedback Design , Addison Wesley, 1989.
K. Zhou and J.C. Doyle, Essentials of Robust Control, Prentice Hall, 1999.
M. Green and D.J. Limbeer, Linear Robust Control , Prentice Hall, 1995.
C.J Ogden and T. Huff, "SVD and its applications in image processing "
An issue of Transactions in Control Systems Technology with a lot of "meat" for our lectures (IEEE members only)
Time delay design 1
Time delay design 2
H-inf Controller Structure