A Coupling Prediction Algorithm for Gas Turbine Remaining Useful Life Based on Health Degree

Abstract

A prediction algorithm for the remaining useful life (RUL) of gas turbine based on the health degree (HD) is proposed in the paper. According to the historical data of the monitoring parameters, the degradation trend of the gas turbine and parameters can be obtained to achieve the purpose of predicting the remaining useful life, and provide the basis for subsequent fault diagnosis and maintenance work. Firstly, the fuzzy analytic hierarchy process (FAHP) is used to construct the calculation model of gas turbine HD. Secondly, the acceleration change point analysis method is combined with the kernel density estimation method to determine the gas turbine fault threshold. On this basis, this paper proposes a new prediction algorithm-- the splicing prediction algorithm based on HD and establishes the RUL prediction model of the gas turbine. Finally, the test data set in C-MAPSS is used for case analysis, and the predicted RUL is compared with the real value to obtain the prediction accuracy. The results show that the proposed prediction algorithm can predict the RUL of some data that meets the degradation detection, and the prediction accuracy is 86.67%, which proves the validity and feasibility of the proposed method.