Fault diagnosis in nonlinear dynamic systems via neural networks

This paper proposes a new approach for detecting and isolating faults in nonlinear dynamic processes using neural networks. Two stages are involved. The first is to generate residual signals based on a comparison between the actual and predicated states. A multilayer perceptron network is trained to predict the future system states based on the current system inputs and states. The paper shows that a satisfactory accurate state prediction for the nonlinear dynamic system can be achieved in this way. In the second stage of fault detection and isolation, a neural network is trained to classify characteristics contained in the residuals. Hence, based on the classification given by the network, faults can be detected and isolated. The developed techniques are demonstrated in a laboratory 3-tanks system and promising results are described.