Advanced Computer Control Project 
Processreglering (Until 2005)

 

Educational programs:

Program	Course Code	Points	Reg Code
Engineering physics	1TT601	4	-
Mathematics and natural sciences	1RT160	5	62535
Single course	1RT160	5	62535

 

Course moments:

Moment	Hours	Instructors, spring 2005:
Lectures	15x2	Mikael Sternad
Laboratory work	2x4	Ove Ewerlid
Project		Daniel Aronsson, Mikael Sternad , Ove Ewerlid

 

Course description for year 2005 (in swedish)

Schedule for lectures.

Course homepage and results

 

Goal:

This project-oriented course is aimed to reinforce and deepen the knowledge about control systems. The project involves the control and supervision of a multivariable process with a computer system, using various approaches. The course gives ample opportunity for independent and creative work. It also provides experience in the design of human-computer interfaces, and in the use of Matlab, HTML and Java programming languages.

Prerequisites:

Automatic Control and Control Design, or equivalent courses.

Contents:

1. Implementation: Fault detection and alarm handling. The process interface. Computer systems and program development for control systems. Real-time issues. The use of Matlab and Matlab-C compilers. Real-time control with Matlab. The human-computer interface: Interface design, HTML and browser-based interfaces. The Java programming language.
2. Modelling: Physical modelling. Transient-, frequency- and spectral analysis. System identification in the frequency domain. System identification with the prediction error method. Subspace identification of multivariable systems.
3. Regulator design: Feedforward compensation for noninteracting control of multivariable plants. More on PID control. Gain scheduling. Regulator and servo synthesis with polynomial methods. Infinite horizon LQ control with integration, designed by state space methods and polynomial methods. Analysis of robustness and the synthesis of robust control systems. Adaptive control with direct and indirect algorithms. Practical aspects of adaptive control. Decomposition of multivariable systems.

 

Project:

Groups of 4 students can choose among six project variants, which differ in the way the process is modelled:

1. Physical modelling and multivariable control.
2. System identification in the frequency domain.
3. System identification with prediction error methods.
4. Direct adaptive control.
5. Indirect adaptive control.
6. Multivariable control

All projects work on coupled electric drives laboratory processes. The control systems are developed mainly in Matlab under Linux. The human-computer interface utilizes Web browsers and is written in HTML and Java.
 

Examination:

Written report and oral presentation of the project, and a demonstration of the resulting control system. Each group will also develop their own Webpage, which presents the results.

Literature:

• Introduction to the course (in swedish)
• M Sternad: Modelling and Control, 1994
• A. Ahlén and M. Sternad: An introduction to adaptive control, 1988
• K.J. Åström, B. Wittenmark: Computer Controlled Systems, Prentice-Hall, 1996.
• Recommended additional reference: K.J. Åström, B. Wittenmark: Adaptive Control, Addison-Wesley, Second ed. 1995.

 

Course responsibility:

Department of Engineering Sciences, Signals and Systems Group.