Fault Diagnosis of a Nuclear Fuel Process with Different Operating Points Using Neural Networks

Abstract

The difficult task of developing a fault diagnosis method for large scale industrial plants can be achieved by artificial neural networks, provided that good process data is made available. Industrial processes often operate at different operating points, and it is important to be able to diagnose faults at any operating point for the safety and reliability of processes. This paper investigates the application of the same neural network for the diagnosis of non-catastrophic faults at different operating points on an industrial nuclear processing plant. Data conditioning methods, including statistical data scaling and principal component analysis, are investigated to facilitate fault classification and to reduce the complexity of neural networks. Results are presented to illustrate the performance of trained neural networks for classifying process faults using simulation and real industrial data.