Prognostics of products using time series analysis based on degradation data

Abstract

This paper presents a method to predict degradation path of products using time series modeling procedure based on product performance degradation data. The product performance degradation data are treated as a time series data and stochastic process are utilized to describe the degradation process for predicting long-term trend. A degradation test is processed for miniature bulbs until they failed and the degradation data are collected for prognostics. Degradation path of a miniature bulb is predicted using time series analysis based on short time period degradation data and long time period degradation data respectively. A comparison between the predicted degradation path and the real degradation path of the miniature bulb is processed and the results show that the degradation path prediction of the product using time series analysis is effective.