Geometric-based approach to fault detection for multilinear affine systems
| Authors: | Hamelin Frederic, Henri Poincare University, Nancy, France, Metropolitan
Boukhobza Taha, Henri Poincare University, Nancy, France, Metropolitan |
|---|
| Topic: | 6.4 Safeprocess |
|---|
| Session: | Active fault diagnosis/ Decision system design |
|---|
| Keywords: | Fault detection, residual generation, robustness, multilinear affine system |
|---|
Abstract
This paper deals with the Robust Fault Detection (RFD) problem formultilinear affine uncertain systems and proposes an approach tosolve it. Convex Minkowski combinations are used to compute asprecisely as possible uncertainty sets according to the parametersvariations in a fault-free context. Faults to be detected arecharacterized either by a parameter change outside its definitioninterval or by an additive term. The proposed approach leads togenerate not only a residual function but also an optimalthreshold. The detection filter associated to the residual generation is suchthat it minimises the effects of both parametric uncertainties andunknown input disturbances on residual while maximizing faultsensitivity.
