| With the continuous development of human science and technology,there are more and more products with long life.The reliability assessment of long life products is usually based on the analysis of accelerated life test data.Accelerated life test usually means to accelerate the failure of samples under the condition that the failure mechanism of products remains unchanged and the test samples are placed under more severe conditions than normal service conditions(collectively known as working stress).The accelerated life test usually takes the time to complete the test,the cost required to complete the test,and the sample configuration at each accelerated stress level as the leading targets.In this paper,the optimum sample configuration for constant stress accelerated life test is studied.The product life is subject to the usual exponential distribution,Weibull distribution and lognormal distribution,and the limited test duration and the expected number of failures at each stress level are taken as the constraint conditions.With the minimum test sample size as the goal,the optimal configuration of the sample size at each stress level is discussed.Sample size configuration of the basic idea is to make full use of a small amount of sample preparation test information,at a given confidence level to ensure the test duration limit,under various stress levels can get expected number of failure data as constraint condition,in order to test sample size at least as the goal,to optimize the allocation of samples under the accelerated stress level.The technical key of this paper is to give the analysis and processing methods of the test data for three different distribution types: exponential distribution,Weibull distribution and lognormal distribution.All the iterative algorithms and optimization processes are realized by programming,and the feasibility,robustness and effectiveness of the algorithm are verified under various experimental constraints. |
PDF Full Text Request