Research On Multivariate Accelerated Degradation Test Based On Copula Function

With the development of science technology and the improvement of processing technology,many products have the characteristics of high reliability and long life, so that it is more difficultto assess reliability and predict life based on product failure data. Accelerated degradation test isan effective measure to assess reliability and predict life for the high-reliability and long-lifeproducts, which can take advantage of product degradation data. Most current research on theperformance degradation test is focused on only one performance parameter, but many complexproducts have more than one degradation performance parameters, and there exists correlationbetween the multiple performance parameters data. So it’s necessary to establish a reasonablecorrelation model for the multiple performance parameters data. Copula function as a flexible,robust correlation analysis tool has many advantages in the analysis of the correlation betweenvariables. Therefore, this paper will base on copula function to study the reliability model,statistical analysis method, reliability assessment and life prediction for the product with multipleperformance parameters. During the whole work, the main research includes:Firstly, chapter1states the background and significance of study; gives an overview of thedevelopment history and research status of accelerated degradation test at home and abroad;summaries the problem and trend of accelerated degradation test; introduces the developmentand research status of copula function in reliability field.Secondly, chapter2presents product reliability index and concept of performancedegradation; establishes the reliability model based on copula function for the product withmultiple performance parameters; introduces the basic theory of copula function, includingdefinition, properties, correlation index, parameter estimation and model selection. Theseresearches are the foundation for multivariate statistical analysis of accelerated degradation testdata and product reliability assessment.Thirdly, chapter3analyzes the characteristics of accelerated degradation data; summariesthe general form of degradation data distribution, degradation path and acceleration equations;establishes the statistical model for multivariate accelerated test data; proposes data distributionselection, distribution parameter estimation, degradation path fitting and copula functionparameter estimation and selection, product reliability assessment and life prediction method;puts forward a statistical analysis method for multivariate accelerated degradation data based oncopula function. An example of numerical simulation is given to verify the feasibility of themethod. Fourthly, chapter4analyzes the failure mechanism of nitrile butadiene rubber under storageconditions; establishes the performance degradation path and acceleration equations for o-ringused on electro-hydraulic servo mechanism; designs an accelerated degradation test plan foro-ring by taking temperature as accelerating factor, as well as taking compression set andcompression stress relaxation as degradation performance parameters; obtains reliability andlifetime of o-ring through analyzing multivariate accelerated degradation data by using themethod proposed in chapter3.Finally, chapter5summarizes the research contents and points out the future researchcontent of accelerated degradation test based on multiple performance parameters.