A robust deconvolution procedure for fault detection and isolation of uncertain linear systems: an LMI approach
| Authors: | Franzè Giuseppe, Università della Calabria, Italy
Casavola Alessandro, Università della Calabria, Italy
Famularo Domenico, Università degli studi di Reggio Calabria, Italy |
|---|
| Topic: | 6.4 Safeprocess |
|---|
| Session: | Fault Detection and Isolation for Linear Systems |
|---|
| Keywords: | Fault detection, Fault isolation, Uncertainty, Linear systems, Convex programming |
|---|
Abstract
A robust Fault Detection and Isolation (FDI) scheme foruncertain polytopic linear systems based on optimal H_\infty deconvolution filters is discussed. The filter must be capable to satisfy two sets of H_\infty constraints: thefirst is a disturbance-to-fault decoupling requirement, whereasthe second expresses the capability of the filter to track thefault signals in a prescribed frequency range. By means of theProjection Lemma, a quasi-convex formulation of the problem isobtained via LMIs. Finally, a FDI logic consisting of an adaptivethresholds scheme based on the on-line rms evaluation of relevantsystem variables is proposed. The effectiveness of the designtechnique is illustrated via a numerical example.
