Abstract
This paper puts forward a reliability estimation method by the Degradation Amount Distribution (DAD) of products, using a composite time series modeling procedure and grey theory based on a random failure threshold. Product DAD data are treated as a composite time series and described using a composite time series model to predict a long-term trend of degradation. The degradation test is processed for a certain electronic product and the degradation data is collected for reliability estimation. Comparison among the reliability evaluation by DAD composite time series analysis and grey theory, based on a constant and a random failure threshold, reliability evaluation by DAD regression analysis based on a random failure threshold, reliability evaluation by degradation path time series analysis, and real reliability of the electronic product is done. The results show that the reliability evaluation of the product using the method proposed is the most creditable of all.