Self-adaptive stress accelerated life testing scheme

Abstract

The initial stress of step-up stress accelerated life tests is commonly too low which is not very efficient when used in accelerated life testing. Though the efficiency of step-down stress accelerated life tests has improved in general, researchers and engineers may have to deal with the failure mechanism changing due to the high starting stress. In this paper, we develop a new accelerated life test scheme, called self-adaptive stress accelerated life tests (SAS-ALT), with considerations of the relative stress loading criterion. We also develop algorithm steps for implementation of SAS-ALT. We formulate and discuss an optimization function of the proposed SAS-ALT that minimizes the variance of the system life time using the maximum-likelihood method, Fisher matrix, and accelerated model. Some factors of the SAS-ALT, such as stress and failure number, were chosen depending upon the optimum function. We also discuss an application of the computer numerical control system based on real failure data to illustrate step-by-step reliability accelerated life test design policy based on our proposed SAS-ALT scheme.