Norm-based design of robust FDI schemes for uncertain systems under feedback control: Comparison of two approaches

This paper discusses the robust Fault Detection and Isolation (FDI) filter design problem for LTI uncertain systems under feedback control. Two-design methods involving norm-based fault detection filters are applied to the three-tank system, and compared to each other. The first method focuses on the fault indicating signal to be an optimal estimate, in the H∞-norm sense, of the fault signal. The second one consists of designing a H∞/H--based residual generator that maximizes fault sensitivity performance and simultaneously minimizes the influence of unknown inputs, for a large class of model perturbations. To compare the fault sensitivity and robustness performances of the two FDI schemes, a general FDI-oriented performance measure is proposed. The criterion involves the calculation of the generalized structured singular value over a specified frequency range.