Fault tolerant control scheme based on multi-ann faulty models

Abstract

Industrial process fault tolerant control (FTC) aims at maintaining the process operating conditions, upon fault occurrence, as much as possible, until scheduled fault isolation procedure is carried out. This paper proposes a FTC scheme based on the employment of multi arti>cial neural networks as process faulty models, and the introduction of compensation paths, using ANNs, to enhance operating conditions recuperation. The proposed scheme is applied to a non-linear simulation model and the advantage of this FTC is illustrated via simulation results. Key wordsArtiJicial Neural Network, Fault Tolerant Control, Genetic Algorithms, Fault Detection and Diagnosis, Modeling, Simulation.