|
Process Control
Processreglering
An advanced project course in control at Uppsala University
Homepage for the Course, Spring 2005
This project-oriented course, is
given within the Engineering Physics program
by the Signals and Systems Group at
Uppsala University.
The project involves advanced
control and supervision of a multivariable
process with a computer system, using various approaches.
Click here
for an outline of the course.
A more detailed introduction to the course (in Swedish)
Lecture scedule
All project groups except one will work with
CE-8 Coupled Electric Drives laboratory processes,
manufactured by TecQuipment Ltd.
The belt tension and belt velocity of these processes
were to be controlled independently. This process is rather
difficult to control, since the belt tension is lightly damped.
All groups first develop SISO PID controllers for the
velocity and the tension. Then, they explore some form
of more advanced control strategy, such as
model based control based on physical models,
ARMAX models or identified state-space models,
predictive control, LQG,
indirect adaptive control or direct adaptive control.
The control systems
are developed in Matlab under Linux.
The human-computer interfaces are implemented
by the groups as Java
applets.
One group will investigate and implement
nitrogen control in a waste water treatment
pilot plant located in Hammarby Sjöstad,
Stockholm.
Graphical User Interfaces and Reports, 2005
(Groups 2005)
Graphical User Interfaces and Reports, 2004
(Groups 2004)
-
Group 1. LQG based on parametric identification.
-
Group 2. Indirect adaptive control.
-
Group 3. Control in a waste water pilot plant.
Graphical User Interfaces and Reports, 2003
(Groups 2003)
-
Group 1. Direct adaptive control.
-
Group 2. Indirect adaptive control.
-
Group 3. MIMO-LQG control with state space methods.
Graphical User Interfaces and Reports, 2002
(Groups 2002)
Graphical User Interfaces and Reports, 2001
(Groups 2001)
Group Hompages and Results 2000:
(Groups 2000)
-
Group 1. Indirect adaptive control.
-
Group 2. Direct adaptive control.
-
Group 3. Direct adaptive control.
-
Group 5. Indirect adaptive control.
-
Group 6. Physical process modelling.
-
Group 7. Indirect adaptive control.
Group Hompages and Results 1999:
(Groups 1999)
-
Group 1. Indirect adaptive control.
-
Group 2. LQG control based on parametric identification.
-
Group 3. Direct and indirect adaptive control.
-
Group 4. LQG control based on parametric identification.
-
Group 5. LQG control based on parametric identification.
-
Group 6. Subspace identification and multivariable LQG.
-
Group 7. Frequency domain identification.
-
Group 1, 1998. Direct adaptive control.
Matlab m-files:
-
ppp.m
for solving scalar Diophantine equations
- spefac.m
for solving scalar polynomial
spectral factorisation algorithms
iteratively with Newtons method.
- spefac2.m
for solving scalar polynomial spectral factorisation algorithms
by computing zeros inside the unit circle.
- spefac5.m:
a fast non-iterative scalar polynomial spectral factorization algorithm.
Of use in indirect adaptive control.
(Note: use n >10 to obtain sufficient accuracy.)
Uses the function
spefac6.m
Reference Material:
-
Matlab online documentation
-
ONE Studio
(formerly called Forte), NetBeans, Sun's main page.
-
Java Lectures (.jar) with live demos
by Alexander Bottema (2003).
Requires JDK 1.4.
Run the presentation with
'java -jar JavaCrash.jar'
- Java Lectures
from year 2000, Java-Kurs.zip, in the form of a java and
html-based tutorial. By Alexander Bottema.
- java.sun.com,
where Java platforms such as JDK 1.4 can be downloaded.
[List of our courses] ;
[Book used
in the course] ;
[Matlab Conference
paper]
Related project course in adaptive signal processing.
[Forskning/forskarutbildning]
|
