Statistical analysis of accelerated degradation testing for old products with initial working time

Statistical analysis of accelerated degradation testing has been of interest to researchers and developers for some time. Relevant study on accelerated degradation testing mainly assumes that samples subjected to accelerated degradation testing are new products, with samples that have not pre-degraded and possess reliability equal to 1 at the beginning of accelerated degradation testing. Old products, which have a pre-existing period of initial working time under use stress, are occasionally selected for accelerated degradation testing in practice. Old products are usually considered for economical reasons, occasionally when new products are no longer manufactured. For the purposes of this study, samples of old products have been degraded to a reliability less than 1 at the beginning of accelerated degradation testing, thus preventing the use of current accelerated degradation testing methods. Improved accurate life prediction for old products is proposed through our models of degradation path and lifetime distribution for constant stress accelerated degradation testing. Initial working time under use stress is evaluated, and actual circumstances are described. A suggested maximum likelihood estimation approach is then detailed accordingly, by combining accelerated degradation testing data and field data to obtain a more accurate estimation for model parameters. Simulation and case studies prove the proposed method’s validity, effectiveness, and ability to accurately obtain improved life prediction for old products with initial working time using accelerated degradation testing.