A Bayesian approach to hybrid fault detection and isolation

Abstract

Fault diagnosis is a crucial component in aircraft control. Fast detection and effective isolation of faults is desired for both manned and unmanned aircrafts to take correct actions when a fault has occurred. This paper proposes a hybrid algorithm for helicopter fault detection and isolation (FDI), which systematically integrates the two paradigms in FDI - model based and data based methodologies - in the Bayesian framework. This hybrid FDI approach has been tested against a helicopter model [1] and excellent FDI performance has been observed.