A New Particle Predictor for Fault Prediction of Nonlinear Time-varying Systems

摘要

Fault prediction (FP) is a new research area in fault detection and diagnosis. It has both safety and economic benefits in industrial systems by preventing future serious process faults and improving process maintenance schedules. Fault prediction differs from fault detection because it deals with future faults, whereas the latter considers faults that occurred in the past. As the future fault is uncertain, a probability to describe it may be needed, however calculating the fault probability remains an open problem. This paper proposes a new particle predictor-based method by which a solution to fault prediction is derived for nonlinear time-varying systems. Simulation results illustrate the effectiveness of the proposed approach.