| As a complex mechanical system,aero engines have extremely high maintenance costs and reliability requirements.Engine wing life prediction is an important part of the health management technology PHM,which can provide a reasonable basis for the development of maintenance plans.With the development of monitoring technology,engine life prediction has evolved from the traditional reliability method of analyzing failure data to the prediction method based on a large amount of monitoring data.The data-driven life prediction method describes the degradation process through case-based reasoning,artificial intelligence,etc.,and is especially suitable for aeroengines,complex mechanical systems that are difficult to establish accurate physical models and cannot collect enough failure data.Based on the evaluation of the health status of the engine through multi-source monitoring data,this paper focuses on the prediction of the engine wing life based on the similarity of the degradation trajectory.The performance parameters closely related to the health of the engine were selected,and the original monitoring data was similarly corrected and the outliers were eliminated.The multi-source monitoring data was merged into the health index HI through the state space model.Cluster analysis of health index sequences to improve the efficiency and accuracy of similarity calculations.An engine wing life prediction method based on the similarity of the degraded trajectory is proposed.The similarity of each engine health index degradation trajectory is calculated by the Euclidean distance method,and the remaining life of the weighted reference sample is obtained to obtain the predicted sample remaining life.Aiming at the characteristics of different initial performance and different decay rate of different engines during the actual use of the engine,two improvements of similarity between sequence slip and segmentation calculation are proposed.The accuracy of the method was verified by actual data used by the airline. |
