Engine Performance Degradation Prediction Based on Unscented Particle Filter Tracking
Aiming at the performance degradation prediction problem of aero-engine,a model for predicting the residual life of an engine based on untraceable particle filter combined with multi-stage effective sensor fusion is proposed.In view of data fusion,effective sensors at each degradation stage are dynamically updated based on cumulative variance contribution rate to obtain health indicators of degradation state that are more consistent with the actual situation.Then,the time-varying parameters of the model are updated by the untraceable particle filter algorithm to realize the advance prediction of engine health indicators and remaining life.the data set of engine performance degradation simulation experiment is used for data analysis and verification.The results show that the proposed multi-stage nonlinear fusion method based on variance contribution rate can reduce the influ-ence of redundant data in the degradation.The proposed algorithm effectively solves the problems of particle degradation and re-sult divergence in traditional filter algorithms.Compared with the general filtering method,the absolute error in the final predic-tion results is reduced by 7 cycles,and the accuracy is improved by 3.6%,which strengthens the tracking ability of particles and improves the accuracy of target tracking.It provides a reference for the health management and reliability evaluation of aeroen-gine.
万方数据
维普
