Defect diagnostics of gas turbine engine using hybrid SVM-ANN with module system in off-design condition

摘要

A hybrid method of an artificial neural network (ANN) and a support vector machine (SVM) has been used for a health monitoring algorithm of a gas turbine engine. The method has the advantage of reducing learning data and converging time without any loss of estimation accuracy, because the SVM classifies the defect location and reduces the learning data range. In off-design condition, however, the operation region of the engine becomes wide and the nonlinearity of learning data increases considerably. Therefore, an improved hybrid method with the module system and the advanced SVM has been suggested to solve the problems. The module system divides the whole operating region into reasonably small-sized sections, and the advanced SVM has two steps of the classification. The proposed algorithm has been proven to reliably and effectively diagnose the simultaneous defects of the triple components as well as the defects of the single and dual components of the gas turbine engine in off-design condition.