Model-Based Estimation Methods
|Lecturer:||Prof. Dr.rer.nat. Arnold Reusken, Dr.-Ing. Adel Mhamdi|
|Course schedule:||Please get them from the course calendar RWTHOnline|
|Lecture notes:||Available in Moodle|
It is a common problem in industrial applications that necessary process data can not be obtained directly using any standard measurement technique. In addition, valid process models should include equipment sizing, process optimization, or advanced control strategies such as model predictive control.
The first part of the lecture consists of an introduction to inverse problems. Here, the focus lies on the ill-posedness of inverse problems. The reasons and consequences of ill-posedness are discussed and mathematical methods for the solution of ill-posed problems are presented.
In the next part, these methods are applied to state, input and parameter estimation problems. In addition, classical methods as the Luenberger-Observer are introduced and their relation to regularization techniques is shown.
Finally the concept of model based optimal experimental design is presented. Approximately half of the accompanying exercises are to be performed using a computer. Here, the methods for the solution of ill-posed problems presented in the lectures are implemented in the mathematical programming environment MATLAB. The quality of the different methods is judged with the aid of example problems. In addition to the computer-exercises, theoretical exercises help in gaining a deeper understanding of the underlying mathematical concepts.