Equipment Fault Forecasting Based on ARMA Model

The analysis of historical time series data that reflects equipment failures is becoming increasingly important in maintenance policies in manufacturing plant. This paper presents a novel methodology to use auto-regressive moving average (ARMA) model for device down time forecasting based on transformed historical data. The 8 orders moving average method was adopted to obtain mean stationary time series with a defined historical data calculated by an algorithm. ARMA model which is extensively used in trend and future behavior prediction, is used to provide a rigorous prediction of the residual series extracted in 8 orders moving average method. By combining data transformation and ARMA model approaches the proposed method can effectively handle the non-linear situation with equipment of highly complicated and non-stationary nature. Its effectiveness is illustrated by an analysis of real-world data. The proposed method is helpful to reflect the equipment condition and thereby can aid predictive maintenance in manufacturing process and reduce the downtime costs.