Title :
Integrated control and fault diagnosis design: a polynomial approach
Author :
Stoustrup, J. ; Grimble, M.J.
Author_Institution :
Tech. Univ., Lyngby, Denmark
Abstract :
A design method is presented which integrates control action and fault detection and isolation. Control systems operating under potentially faulty conditions are considered, and it is demonstrated how to design a single unit which handle both the required control action, as well as identifying faults occurring in actuators and sensors. This unit is able to: 1) follow references and reject disturbances robustly; 2) control the system such that undetected failures do not have disastrous effects; 3) reduce the number of false alarms; and 4) identify which faults have occurred. The method uses a type of separation principle which makes the design process very transparent, and a polynomial H∞ formulation which makes weight selection straightforward. As a consequence of the separation between control and diagnosis, we prove that the controller needs not be detuned in order to get good diagnosis results, in contrast to common beliefs
Keywords :
control system synthesis; control design; control systems; disturbance rejection; fault detection; fault diagnosis design; fault isolation; feedback; identification; polynomial; polynomial H∞ control;
Conference_Titel :
Modeling and Signal Processing for Fault Diagnosis (Digest No.: 1996/260), IEE Colloquium on
Conference_Location :
Leicester
DOI :
10.1049/ic:19961380