Fuzzy inductive reasoning model-based fault detection applied to a commercial aircraft

Abstract

This paper concentrates on progress made in fault detection using the Fuzzy Inductive Reasoning (FIR) methodology. The so-called envelope approach is presented and compared with previous research efforts published by the same group in SIMULATION in 1989 and 1994. The comparison is achieved simulating faults in a B-747 commercial aircraft, the same example that had been used in the previous publications. The previously used crisp detection has been replaced by a new approach, based on the computation of an acceptability interval for each predicted variable. The paper demonstrates the improved capabilities of the new approach to detect faults earlier and in a more reliable manner.