Failure trend analysis using time series model

Abstract

Failure trend analysis has to be based on observed operational failure data. Assume the failure data can be viewed as a series of data over time. And a set of time series will be obtained. So it is perfectly natural to use the time series model to test the failure trend. Then we consider the failure number arranged by time order as a variable. As a result of the effects of seasons and cycles, we found the structural time series model is the appropriate model for modeling the public transport vehicles failure data. The structural time series model used in this paper is added with four components, namely trend, cyclic, seasonal and irregular. The failure number forecasting and correcting are also be given. In order to illustrate the efficiency of the structural time series model, a real-world example will be presented.