Fault-Recovery Strategy on Critic Fault Tolerant Control Systems under FF Technology
The aim of the study deals with some aspects of the functional and hardware redundancy in fault detection, fault isolation, decision making and system recovery to solve the problem of supplying wrong information to severe or critic control systems, achieving a fault tolerant control system. To get such objectives, back-propagation neural networks are used as universal functional approximation devices which are used as residuals generators. Residuals will be evaluated by means of rule based novelty strategies in a decision-making task. Implementation procedure is carried out with the facilities supplied by a FOUNDATION™ Fieldbus compliant tool, which manage databases, neural network structures and training algorithms under an standard object oriented environment. Experimental results on a heat exchanger pilot plant are satisfactory.