Study On Data Conversion Used In Accelerated Life Test

Currently in the engineering test, in order to get the life time of the products under normal working conditions, we usually use accelerated life test. Acceleration equation is common used in accelerated life test to calculate the life time of products. However, due to measurement error, model error and other factors exist, there will be a larger acceleration equation error, thus affecting the final calculation results. In some cases, it is difficult to determine accelerate the equation. In addition, the stress designing in accelerated life test is very important, and certain restrictions exist.In this paper, the data conversion method is used to determine the life time of products under normal working conditions. Set aside the restrictions of model and stress design, regression conversion or geometric mean ratio conversion is used to convert data under high stress levels to the normal stress level, and then execute statistical analysis according to the nature of the different distribution. In this paper, through simulation, the feasibility and effectiveness of data conversion used in exponential distribution and lognormal distribution are discussed; In addition, the effect of small sample, number of accelerated stress, and the simulation parameters are also discussed. At the same time, taking into account that censored test is often used in the accelerated life test, another main content of the paper is to consider how to compute life time of the products use censored samples.