Advanced control and applications

Teacher(s)

Hendrickx Julien;

Language

English

Main themes

Model-based control (pole placement control, predictive control, LQ control, robust control) ; Implementation aspects of digital control.

Aims

At the end of this learning unit, the student is able to :
1	Contribution of the course to the program objectives : AA1.1, AA1.2, AA1.3 AA2.1, AA2.2, AA2.3, AA2.4 AA3.1, AA3.2 AA5.3, AA5.4, AA5.5, AA5.6 AA6.4 The aim of this course is to present different methods of model-based control (pole placement control, predictive control, LQ control, robust control) and to study the implementation aspects of digital control. These methods will be supported by real life cases studies. The course also involves control design exercises (MATLAB), and a set of laboratory sequences during which the students will implement some of these methods on pilot processes at the laboratory. At the end of this course, the students will be able to : Understand the major issues of digital control design. Calculate, with specialized software, digital controllers with specified performances. Implement numerical control laws on real processes (in the laboratory). Present major aspects of a theory or an application in automatic control.

Content

• Discretization of continuous models, Shannon's theorem, choice of sampling periods
• Classical digital control (numerical PID)
• Predictive control
• Prediction compensation of measurable perturbations
• Multivariable control, decoupling, linear quadratic control
• Observers, Kalman filter
• Delay compensation
• Parameterization of Youla Kucera
• Recursive model estimation
• Robust control
• Iterative controller design
• Controller design with different methods using MATLAB and SIMULINK
• Test of different control methods on pilot processes.

The course comprises a set of lectures on theoretical aspects in control design or regarding industrial control applications developed by members of the Automatic Control Lab, as well as a set of compulsory exercises and laboratory sequences. Moreover, each student will have to make an oral presentation on a theoretical topic, or on results obtained in the laboratory or, finally, on an article describing an industrial application.

Teaching methods

Due to the COVID-19 crisis, the information in this section is particularly likely to change.

1. Lectures and exercices: 3-5 lectures and problem-based learning sessions on (i) preliminary notions necessary for the class, (ii) sampling linear systems, and (iii) dealing with constraints on input and output signals.
2. Seminars : Between 6 and 12 seminars prepared by students. Each student/group of student receives several documents on a topic novel for them. Based on these documents and on their own research, they understand the new topic, critically analyze it, prepare a synthesis of its essential aspects, and present this synthesis to the other students. Each group can interact with the professor before their seminar, and a constructive feedback is provided after the seminar. The precise size of the group depends on the number of registred students.
3. Homeworks : One or two homeworks about sampling problems, done alone or by groups of two students.
4. Labs : Two or Three experiments in the laboratory (by groups of 2 or 3). The goal of each lab is to design a controler for a real and nontrivial dynamical system. They also allow students to face realistic (possibly unforeseen) practical problems. Depending on the sanitary conditions and practical constraints, the lab may be raplaced by virtual labs to be performed on computer
5. External activities : these will change every year. They may include :

• presentation of an advanced control method by a researcher
• seminar about practical control issues by someone from working on control problems in the industry
• relevant visit of a plant/other facility where control methods are used
• each student writes a short report after each external activity.

Activities 1 and 2 (lectures, seminar, exercices) normally take place in classrooms, but may be moved partially or fully online depending on the sanitary situation, practical constraints, and the number of registered students.

Evaluation methods

Due to the COVID-19 crisis, the information in this section is particularly likely to change.

The grades will be based on :

• The seminar(s) presented by the students : their ability to convey the main ideas to the other students, their critical thinking and ability to synthetize information when preparing the presentation, and their knowledge of the topic presented.
• The labs
• The homework(s)
• The reports written after every external activity or seminar

The precise evaluation criteria and weights are specified during the first lecture each year, and are available on Moodle.
In case there is a significant difference between the grade obtained for individual activities and group activities, the teaching team may assign a grade reflecting the individual level. A (compulsory) oral exam may be organized for some students to obtain complementary information in case the teaching team has a doubt on the grade to assign.

Other information

Knowledge of basic control techniques (e.g. LINMA 1510) and dynamical systems is expected.

Online resources

http://moodleucl.uclouvain.be/course/view.php?id=7955

Faculty or entity

MAP