Bayesian Optimal Design for Step-Stress Accelerated Degradation Testing Based on Gamma Process and Relative Entropy

Abstract

AbstractAccelerated degradation testing (ADT) technology for long-life and high-reliability products has become one of the key technologies in life and reliability field. The scientific and reasonable testing program can not only provide correct basis for decision making, but also make full use of resources and reduce the cost of product development. Hence, how to make full use of products’ historical information to develop a short-term efficient pilot program has become a key-problem to be solved in ADT technology. This chaptere proposes the Bayesian optimal design of step-stress accelerated degradation testing (SSADT) based on Gamma process and relative entropy. Firstly, we briefly describe the applicability of Gamma process and the relative entropy in ADT, and the degradation model and relative entropy’s application method are given. Secondly, under the framework of Bayesian theory, we study the Gamma degradation process based SSADT optimal design method by using maximize the relative entropy as the optimization goals and test variables as the optimization design constraints. Finally, we use a 3-steps bulb’s SSADT to verify the effectiveness of the proposed method. The example shows that the method of this chapter is fast and efficient which can comprehensively use the prior information to work out the optimal pilot program.KeywordsPosterior DistributionRelative EntropyDegradation ModelOptimization GoalBayesian TheoryThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.