Design and Evaluation Method for Multi-Stress Accelerated Life Test with Small Samples
A small-sample multi-stress accelerated life test design and evaluation method is proposed to solve the problems of conventional multi-stress accelerated life tests,such as many combined stress levels,long period,high cost.By establishing the relationship between the one-sided lower confidence limit of reliable life at normal stress level and the confidence limits of reliable life at accelerated stress levels,the one-sided lower confidence limit of reliable life at normal stress level can be assessed with high confidence and high reliability.Firstly,for the most commonly used dual-stress situation,the design and evaluation methods for small-sample dual-stress accelerated life test without and with coupling are established,respectively.Using these methods,only three(without coupling)or four(with coupling)combined stress levels need to be selected in the life tests,and the reliable life at normal stress level of the product can be evaluated with high confi-dence and high reliability.Then,for the general m-stress situation,a small-sample multi-stress accelerated life test design and evaluation method is proposed based on the multi-stress linear acceleration model,which requires conducting accelerated life tests at m+1 combined stress levels,and the one-sided lower confidence limit of reliable life at the normal stress level can be obtained.The method is easy to im-plement and applicable to complete data,incomplete data and various lifetime distributions.Numerous case studies and Monte Carlo simula-tions show that the proposed method not only saves the specimens and test time,but also is more theoretically rigorous and more accurate than the conventional methods.
Multi-stressAccelerated life testSmall-sampleTest designReliable lifeConfidence limit
NETL
NSTL
万方数据
