仿真基混合效应模型加速退化试验方案优化设计研究

加速退化试验技术广泛应用于存在性能指标随时间逐渐退化的高可靠长寿命产品的可靠性及寿命评估.不同的加速退化试验方案得到的可靠性及寿命评估结果有很大区别,试验费用也各不相同.如何设计试验方案使试验评估结果最准确、试验代价最小,是加速退化试验应用中面临的一个重要问题.针对解析优化方法在某些场合最优解难以得到甚至不存在的情况,提出一种新的基于Monte Carlo仿真的混合效应模型加速退化试验优化设计方法,将复杂的优化问题转化为较易的统计分析问题,使之流程化好、便于工程应用.以发光二极管恒定应力加速退化试验方案优化设计为例研究方法的有效性.针对该方法的敏感性分析结果表明,在较小的模型偏差情况下,最优加速退化试验方案具有一定的鲁棒性.在工程应用中,利用相同的方法流程可实现其他类型加速退化试验方案的优化设计.

Simulation-based Optimal Design for Accelerated Degradation Tests with Mixed-effects Model

Accelerated degradation test (ADT) is widely used to evaluate the reliability of highly reliable products with long life if there exists a product quality characteristic whose degradation over time can be related to reliability. Different ADT plans induce very different estimation of reliability and life characteristics with different test cost. It is a main challenge that how to design accelerated degradation test plans which induce precise results with low cost. Aimed at an occasion that analytical solution of optimal plan does not exist, a new method of Monte Carlo simulation-based optimal design for ADT is presented to convert a difficult design problem to a rather simple problem of statistical analysis for application. Optimal design for constant stress accelerated degradation test of light emitting diode is taken as an illustrative example to demonstrate the validity of the proposed method. Results of sensitivity analysis of the optimal test plan show that the plan is robust when the model deviation is small. In application, optimal plans of other types of accelerated degradation test can be obtained by the same method and process as in the example.