uÒFDr.Andres VargasiGerman Aerospace Center,Senior Scientistj
"Robust fault detection and isolation using the Fault Detection Toolbox for MATLAB"
The robust fault detection and isolation (FDI) problem is formulated in a general nonlinear setting and residual generation approaches based on multi-model fault descriptions and explicit actuator/sensor fault modelling are presented. A solution approach is described using linear FDI techniques implemented in the recently developed Fault Detection Toolbox for MATLAB. The main aspect for robust FDI is the determination of appropriate thresholds for residual signals which provide a good trade-off between false alarm and missed detection rates. A robust FDI problem for a single uncertain parameter (mass) is exemplary solved for a Boeing 747 actuator fault detection problem. Preliminary results obtained using the FD Toolbox illustrate both the difficulties of solving robust FDI problems as well as the high potential of employing reliable software tools to analyze and solve complex FDI problems.