Research On Degradation Reliability Evaluation Based On Gamma Process

With the development of science and technology,reliability has become a key factor to improve product competitiveness.With the improvement of reliability requirement,a large number of high reliability and long life products are applied.However,for high reliability and long life products,the traditional reliability evaluation method based on life data is very difficult to obtain failure data and carry out degradation experiments under limited time and cost.Considering that the process of product degradation contains abundant reliability information,the reliability evaluation method based on performance degradation data has been paid more and more attention.In the practical application,the reliability evaluation based on degenerate products is not perfect.Taking these problems into consideration,the original model is optimized.The main contents are as follows:(1)Two traditional methods based on degenerate data are introduced,analyzes their applicability and deficiencies in dealing with degenerate data,and proposes a stochastic process model to overcome these problems.(2)Evaluation of the Gamma process reliability taking into account individual differences.Using the Gamma process to describe the degradation process of mechanical products,the original Gamma process model based on performance degradation increment was established,and then taking into account the impact of individual differences,the scale parameters and shape parameters of the Gamma process were randomized,and the optimized Gamma model was established.Taking into account the problem of small samples,the bayesian parameter estimation method is used to combine the historical prior information to carry out the parameter estimation and complete the product reliability evaluation.Taking the gear as the research object,two methods are adopted to model.By using the DIC model selection criteria,the reliability evaluation model considering individual differences is proved to be more suitable for gear performance degradation data.(3)An accelerated degradation evaluation model under constant stress based on Gamma process and an experimental optimization scheme for product degradation under different degradation trajectories are established.By using the relation between Gamma process and BS distribution,the Gamma process is transformed into BS distribution,which simplifies the complexity of the derivation of mean variance of reliability.Taking the mean variance of reliability as the optimization objective,the optimization experiment scheme of products under convex,concave and linear trajectories was analyzed,and the sensitivity of the shape parameters,scale parameters and nonlinear parameters of the Gamma process to the accuracy was analyzed.(4)Reliability assessment considering multivariate degradation and sudden failure.The Clayton Copula function and Gumbel Copula function,which are very sensitive to the fluctuation of the upper and lower tail,were selected and combined into Copula function.This paper describes the multivariate degradation using the optimized mixed Copula function.On the basis of multivariate degradation,the impact of sudden failure is taken into account.Using Wiener process,the burst failure model is established.