Accelerated Degradation Test Plan Based On Inverse Gaussian Process With Measurement Error

With the continuous increase of product requirements from customers,more and more products have the characteristics of high reliability and long life.How to get the product's life in a shorter time is a very important research topic.Studies have found that the product's performance characteristics' degradation includes enormous valuable reliability information.And engineers can predict the life of a product by collecting degradation data over time.The degradation path of a product can be described by a stochastic process model.Due to a better match than Wiener process and Gamma process in some monotonically increasing product degeneration,inverse Gaussian process model has been paid close attention by researchers.In the actual data collection,the measurement error can't be completely avoided,so the degradation process model with measurement error is more real.In addition,under the influence of accelerating stress,the rate of degradation is accelerated and the test time is greatly shortened.Therefore,the accelerated degradation test can obtain the reliability information of products quickly and effectively.This paper focuses on the inverse Gaussian process model with measurement error.The parameter estimation method of the complicated model is studied.The optimal configuration of constant stress and step stress accelerated degradation tests are discussed and illustrated by examples.The main works include:(1)The basic theory of inverse Gaussian distribution and inverse Gaussian process is explained.Based on the theory,an inverse Gaussian process model with measurement error is established.(2)The likelihood functions of unknown model parameters are constructed.Convolution formula and Monte Carlo method are used to calculate the maximum likelihood estimation of the complex model parameters.This paper gives the lifetime estimation method of inverse Gaussian process model with measurement error.(3)The optimal configurations of the inverse Gaussian process with measurement error is discussed respectively in constant stress and step stress accelerated degradation tests.The data of carbon film resistors and the electrical connectors is taken as examples.(4)This paper analyzes parameter estimation and the optimal configuration of accelerated degradation test when the inverse Gaussian process with measurement error is mis-specified as the Wiener process model.The research in this paper shows that the inverse Gaussian process model with measurement error is more suitable in the actual situation.In the constant stress and step stress accelerated degradation tests,the evaluation accuracy of the model improves while the budget of the test increases.In addition,the mis-specification of models greatly affects the accuracy of product life estimation and the optimization of accelerated degradation tests.Therefore,it is necessary to strengthen the research on the mechanism of product failure to accurately select the degradation process model.