Wavelet Neural Network based Fault Detection and Diagnosis in an Industrial Gas Turbine

This study presents the use of Wavelet Neural Networks (WNN) to design a fault detection and diagnosis system for a gas turbine generator. It also demonstrates the use of firefly algorithm for training WNN based models. In the fault detector block, an adaptive threshold is introduced that makes it more reliable. In all the test cases related to model identification, WNN models showed close agreement with actual data. In the fault detection and diagnosis test, the proposed method was found capable of providing true detection and diagnosis percentages, each higher than 95%.