Study On Life Prediction Of Turbofan Engine Based On Weibull Distribution

Research statistics show that the failure of aero-engine is one of the two main reasons leading to aircraft failure.The reliability of the engine is directly related to whether the working state of aero-aircraft meets the mission requirements.However,because of the complex structure system and the severe working conditions,its on-wing reliability is often difficult to be guaranteed,so it is of great significance to make an intensive study of the engine remaining life prediction.Equipment maintenance personnel can judge the operating status of the equipment and the necessity of maintenance or replacement by the results of residual life prediction,so as to develop an optimal maintenance strategy,so that the equipment and even the whole system can consume the least maintenance resources and achieve the longest effective service life on the premise of ensuring reliability.In this paper,the significance of research on the life prediction of turbofan engine is briefly described,and the research status of life prediction technology at home and abroad is investigated and summarized according to the needs of the subject.On this basis,the main research contents and research schemes of this paper are proposed.In order to understand the intermediate data generated in the calculation process better and judge whether the calculation result is reasonable,the structural composition,working principle,important performance parameters,typical failure mode and corresponding parameter changes of the turbofan engine are simply carried out firstly.Then,the relevant information of turbofan engine performance degradation simulation experiment conducted by NASA Forecasting Center and the specific content and form of the published experimental data set are described in detail.Before carrying out life prediction,in order to improve the processing covariate data and the computational efficiency of life prediction,covariate analysis is carried out by Pearson correlation coefficient method and stepwise regression method.Several parameters which can effectively describe the degradation process of equipment are selected from many monitoring parameters to form a comprehensive covariate expression.After the completion of the covariate analysis,an improved dynamic variable-scale Markov model combined with the fuzzy C-means algorithm is used to predict the comprehensive covariate,which is an important basis for the subsequent life prediction simulation calculation.Based on the Weibull distribution fitting test of the experimental data by linear regression method,the life prediction of turbofan engine can be divided into two kinds: the life prediction with covariate and the life prediction without covariate according to the actual demand of the subject.The life prediction without covariate is mainly based on the two-parameter Weibull distribution model,the reliability,failure rate density and residual life probability density of the equipment can be obtained only based on the history data of failure or cut-off life in the life test.The life prediction with covariates is mainly based on the improved Weibull proportional risk model,which can predict the remaining useful life in real time based on the time-varying monitoring data.The effectiveness of the proposed algorithm is verified by comparing the calculated results with the life values given by the experimental data.