Selection of residual generators in structural analysis for fault diagnosis using a diagnosability index

Structural analysis for fault detection and isolation (FDI) is an efficient method for designing structured residual generators for model-based diagnosis. It generates a number of diagnostic test candidates, whose number may be large in case of complex systems. This paper proposes a method for selecting residual generators from a possibly large number of candidate equation sets using a diagnosability index that measures the degree of diagnosability in a diagnostic system. The selection process of residual generators is aimed at reducing the solution sets for achieving the optimal diagnostic performance, i.e, maximizing the detectability and isolability of various faults. The approach for selecting residual generators is applied to a permanent magnet synchronous machine (PMSM) drive system in an electrified vehicle to show that the proposed methodology is effective in automatically downsizing the candidate equation sets for diagnostic tests from a relatively larger number of choices derived from the structural analysis algorithms.