A demonstration of artificial neural-networks-based data mining for gas-turbine-driven compressor stations

This paper presents a successful demonstration of application of neural networks to perform various data mining fiunctions on an RB211 gas-turbine-driven compressor station. Radial basis function networks it-ere optimized and were capable of performing the following functions. (a) backup of critical parameters, (b) detection of sensor faults, (c) prediction of complete engine operating health with few variables, and (d) estimation of parameters that cannot be measured. A Kohonen SOM technique has also been applied to recognize the correctness and validity of any data once the network, is trained oil a good set of data, This it-as achieved by examining the activation levels of the winning unit on the output layer of the network. Additionally, it would also be possible to determine the suspicious, faulty or corrupted parameter(s) in the cases which are not recognized by the network by simply examining the activation levels of the input neurons.