Reliability Statistical Analysis And Optimum Design Plans Based On Accelerated Life Test Data

In order to realize the rapid evaluation of life and reliability for high reliability and long life products at normal stress level.In this paper,the different reliability statistical models of Weibull distribution constant stress accelerated life test(CSALT)were established.Conduct a comprehensive and deep research for the optimum design and statistical analysis method of test datas for the CSALT.The main research conclusions are as follows:Firstly,by establishing the reliability statistical model of Arrhenius—Weibull.The minimum asymptotic variance of the product median lifetime estimate at normal stress level as optimum object was considered,the size and number of test stress level,samples allocation proporation and censoring time at each stress level are determined as design variables.The optimum design method of confidence sample size for Weibull distribution time censoring CSALT is presented,and the statistical model for evaluating the merits and demerits of the CSALT plan.Through Monte Carlo simulation test,the feasibility study and analysis of the optimum design method are carried out,and the results show that the method is feasible.Secondly,under the reliability statistical model assumption of Weibull CSALT,the simplified Maximum likelihood estimation(SMLE)method for Weibull distribution CSALT datas is put forward.The Fisher information matrix and covariance matrix are obtained,which is used to construct the asymptotic confidence intervals of the model parameters.Based on the asymptotic normality of MLE,the asymptotic confidence intervals of the model parameters and reliability are given.Statistical analysis of CSALT datas at thermal stress and electrical stress is carried out by using the given method,and the results show the feasibility and universal applicability of the given method.Thirdly,by establishing the interactive reliability statistical model of generalized Eyring—Weibull.The optimum design theory and simulation evaluation method for the Weibull distribution double stress CSALT are discussed,taking the minimum asymptotic variance of the median life estimation of the product at the combinations of normal stress level as the optimum criterion,and taking the thermal stress and the non thermal stress as the accelerated stress simultaneously.A statistical model for evaluating the advantages and disadvantages of the test plan is proposed,and the method for determining the sample size is also presented.Under the effect of temperature and vibration double stress on the space electrical connector,Monte Carlo simulation of CSALT is carried out.The results show that the method is feasible and has a certain theoretical reference value.Fourthly,under the no-interactive reliability statistical model assumption of generalized Eyring—Weibull.A simplified MLE method for Weibull double stress SCALT datas is studied.The Fisher information matrix and covariance matrix are obtained,which is used to construct the asymptotic confidence intervals of the model parameters.The asymptotic confidence intervals of the model parameters are also given.The method is used to analyze the Monte Carlo simulation datas of the temperature and vibration double stress CSALT.The analysis results show that the method is feasible and the efficiency compared with the traditional MLE method is significantly improved.