Research On Reliability Assessment And Prediction Method Of Airborne Equipment Under Asynchronous And Left-censored

The aircraft is usually put into operation in batches,and asynchronous operation is a typical feature of field failure and maintenance data of airborne equipment.In addition,the field data is also faced with the problem of left censored.Asynchronous left censored data is the difficulty of reliability assessment and prediction of airborne equipment using field data.For this reason,this dissertation first deals with the asynchronous left censored field data,then studies the reliability assessment method of airborne equipment based on stochastic process,and studies the fault prediction method of airborne equipment based on one-sample prediction theory.The field fault records are analyzed,and the fault records are filtered by unifying the aircraft types and selecting the airborne equipment.Then,the fault information that is not the selected equipment from the filtering results is eliminated.At last,the date information of eliminated filtered results is only retained.The asynchronous operation process is transformed into synchronous operation by translating the data points.The data is divided into two groups with or without left censored data.On this basis,the problem of left censored data modeling is solved by modeling separately.Aiming at the problem that the traditional left censored data model can't be applied to the left censored data with unknown fault time and times,this dissertation establishes the non-homogeneous Poisson process model with left censored data and without left censored data respectively on the basis of fault observation data combined with asynchronous operation process,and then synthesizes the two parts to get the fault process model under the condition of asynchronous left censored,which solves the problem of asynchronous operation and left censored data modeling.Based on the non-homogeneous Poisson process model under the condition of asynchronous left censored,the priori information is integrated into the model by Bayes formula,and then Bayes statistical inference is carried out for airborne equipment.Bayes point estimation and confidence limit of reliability parameters of airborne equipment are given.Finally,the life data of airborne equipment is taken as the basic data of reliability prediction.The one-sample prediction research based on Weibull distribution is carried out.By introducing odds ratio,the failure times of future time are predicted,and the confidence interval of the predicted failure times is given.Then,based on the failure observation data,the one-sample prediction of future failure time of airborne equipment is carried out based on the non-homogeneous Poisson process,the prediction of fault time and its confidence limit under different fault numbers in and the future are given.Through the analysis of the one-sample prediction results of the two methods,it is concluded that the one-sample prediction results based on the nonhomogeneous Poisson process are more close to the engineering practice.