A multi-feature fusion model based on BRB of health state prediction for aeroengine gas path system

Abstract

The health of an aeroengine gas path system is essential to the reliable flight of the aircraft. Due to the complexity and coupling of aeroengine gas path systems, the establishment of a dynamic and comprehensive model for the health state prediction is difficult. It is very necessary to establish the prediction model by fusing multiple features instead of using a single feature such as exhaust temperature. A belief rule base (BRB) shows outstanding performance in modeling complex systems. This paper proposes a multi-feature fusion model based on BRB of health state prediction for aeroengine gas path system. In this model, firstly, the health characteristics of the aeroengine gas path system with different physical characteristics is taken. Secondly, a time series prediction model of the health characteristics based on BRB is established. Finally, the evidence reasoning (ER) algorithm is used to fuse these health characteristics to achieve the comprehensive health state prediction of the aeroengine gas path system. The BRB health state prediction model combines both quantitative information and expert knowledge to remedy deficiency of effective data and improve the prediction accuracy. Considering the initial parameters given by experts are subjective and may not be appropriate for engineering practice. The projection covariance matrix adaptive evolution strategy (P-CMA-ES) is selected as the optimization algorithm for training the initial parameters. Finally, a certain type of aeroengine is taken as a case to verify the effectiveness of the proposed model. The results show that the health state prediction model based on BRB with multi-feature fusion can accurately predict the health states of aeroengine gas path system.