Research On Competitive Failure Reliability Method Of Performance Degradation Products Based On Shift Threshold

Reliability is an important theoretical basis for evaluating product performance and safety.The traditional reliability theory is based on probability theory that is modeled and analyzed through a large number of failure data.However,with the rapid development of modern technology,products are characterized by high reliability and long life recently,so that it is difficult to obtain a large amount of failure data in a short time,which restricts product reliability analysis.Product performance degradation data reflects its reliability to a certain extent.Moreover,product failure is caused by competition among various mode failures.Therefore,this thesis will analyze the reliability of competitive failure of products based on performance degradation data combining with Bootstrap under shift thresholds condition.The main research contents are organized as follows:(1)The basic theory of performance degradation is analyzed and the concept of failure threshold is given while describing reliability statistics knowledge.In the failure mode analysis,this paper focuses on sudden failure and degradation failure.The statistical model of degradation amount and the degradation failure model are proposed and finally the reliability analysis method based on performance degradation data is obtained in this paper.(2)Based on the single performance degradation data,the reliability model is established and the degradation quantity distribution method,the degradation trajectory method and the stochastic process method under shift thresholds condition are adopted.In the study of the degradation quantity distribution method,the normalized degradation distribution and Weibull distribution reliability model based on shift failure threshold are established.In the degradation trajectory method,the concept of pseudo-failure life is proposed.SPSS and MATLAB softwares combined with Bootstrap are used to perform hypothesis test analysis on life distribution.In the stochastic process method,due to the good adaptability of the Wiener stochastic process and the inverse life Gaussian distribution are derived that a reliability model based on the Wiener process under shift thresholds condition is established.(3)A product competitive failure reliability assessment method with both sudden failure and degradation failure is established and both sudden failure and degradation failure are considered independently or related.The Weibull distribution is introduced to describe the reliability model of sudden failure and the Gamma random process is used to describe the performance degradation process.The sudden failure is evaluated from the perspective of the amount of degradation and the reliability model of the competition failure is established according to the conditional probability of the performance failure amount according to the sudden failure.The least square method is used to estimate the parameters in the model.(4)In order to study the competitive failure analysis of multi-performance parameter degradation,a multi-parameter degradation model based on multivariate normal distribution and a competitive failure reliability model based on multivariate Gamma stochastic process under shift thresholds condition are established.In the multivariate normal distribution model,covariance matrix are used to describe the correlation between the various degradation parameters.Correlation coefficients are used to describe correlations in a multivariate Gamma stochastic process model.Both methods consider the independence and correlation between the various parameters.