A Fuzzy Bayesian Based Zero-Failure Accelerated Life Demonstration Test Design Method

The zero-failure accelerated degradation demonstration (ADD) test method is employed to verify the product reliability. Generally, the parameter of the acceleration model is assumed to be known in advance. However, the epistemic uncertainty due to the discrepancy of lifetime distribution between the earlier test unit and current products should be taken into consideration. Besides, the epistemic uncertainty due to small sample also leads to an inaccurate estimation. Therefore, it is worth to consider these two epistemic uncertainties above and the random uncertainties caused by random effects in time dimension and unit-to-unit variations. In this paper, a fuzzy Bayesian based zero-failure accelerated degradation demonstration (ADD) test method is proposed. This method takes the three uncertainties into account and is optimized based on the target of minimizing the test cost with a minimum sample size. An electrical power panel zero-failure ADD test is presented to illustrate the proposed method.