Application of Autoassociative Neural Network on Gas-Path Sensor Data Validation

Abstract

Gas-path analysis holds a central position in the engine condition monitoring and fault diagnostics technique. The success of gas-path analysis, as concluded from previous investigations, depends mainly on the quality of the measurements obtained. For approaches using either a classical Kalman filter method or a contemporary artificial neural network approach, a high success rate of diagnosis can only be guaranteed when a correct set of measurement deltas is available. The objective of the present work is to design a genetic autoassociative neural network algorithm that can perform offline sensor data validation simultaneously for noise filtering and bias detection and correction. The neural network-based sensor validation procedure usually suffers from the slow convergence in network training. In addition, the trained network often fails to provide an accurate accommodation when bias error is detected. To remedy these network training and bias accommodation problems, a two-step approach is proposed. The first step is the construction of a noise-filtering and a self-mapping autoassociative neural network based on the backpropagation algorithm. The second step uses an optimization procedure built on top of these noise-filtering and self-mapping nets to perform bias detection and correction. A nongradient genetic algorithm search is employed as the optimization method. It is shown in the present work that effective sensor data validation can be achieved for noise filtering, bias correction, and missing sensor data replacement incurred in the gas-path analysis. This newly developed algorithm can also serve as an intelligent trend detector. A true performance delta and trend change can be identified with no delay, to assure a timely and high-quality engine fault diagnosis.