Reliability modeling and a statistical inference method of accelerated degradation testing with multiple stresses and dependent competing failure processes

• Published in: Reliability engineering & system safety Vol. 213; p. 107648
• Main Authors: Liu, Yao, Wang, Yashun, Fan, Zhengwei, Bai, Guanghan, Chen, Xun
• Format: Journal Article
• Language: English
• Published: Barking Elsevier Ltd 01.09.2021; Elsevier BV
• Subjects: Accelerated degradation testing; Accelerated tests; Degradation; Dependent competing failure processes; Failure; Mathematical models; Model accuracy; Multiple stresses; Parameter estimation; Reliability; Reliability analysis; Reliability engineering; Reliability modeling; Statistical inference; Statistical inference method; Stresses; Wear rate; Accelerated degradation testing; Multiple stresses; Statistical inference method; Reliability modeling; Dependent competing failure processes
• ISSN: 0951-8320; 1879-0836
• DOI: 10.1016/j.ress.2021.107648

•A reliability model with multiple stresses and multiple failure processes is proposed.•A statistical inference method of ATD with multiple stresses is proposed.•Practical example is used to demonstrate accuracy of the model and method. In this paper, a multiple stresses reliability model with dependent competing failure processes (DCFPs) is proposed, which includes constructing the multiple stresses acceleration model and deriving the degradation-shock dependences competing model. The multiple stresses coupling are considered in multiple stresses acceleration model, and the degradation-shock dependence is considered and modeled by Facilitation model. Then, a statistical inference method of accelerated degradation testing with multiple stresses and dependent competing failure processes is proposed. Finally, a practical example is used to demonstrate accuracy of the proposed model and method. It is shown that the reliability model without considering multiple environment stresses is a special case (40°C, 65%RH and random shocks) of that considering multiple environment stresses. We also explain the phenomenon that the reliability at same time is lower with the larger temperature, larger humidity and more random shock due to that the mean of wear rate is larger with the larger stress, and the reliability is lower with the larger mean of wear rate. Moreover, the maximum MSE of the parameter estimation result obtained by the statistical inference method is 0.34%.