Applications residual control charts based on variable limits

Abstract

The main purpose of this paper is to verify the stability of a productive process in the presence of the effects of autocorrelation and volatility, in order to capture these characteristics by a joint forecast model which produces residuals that are evaluated by a control chart based on variable control limits. The methodology employed will be the joint estimation of the residuals by ARIMA – ARCH models and the conditional standard deviation from residuals to establish the chart control limits. The joint AR (1)-ARCH (1) model shows that an appropriate forecasting model brings a great contribution to the performance of residual control charts in monitoring the stability of industrial variables using just one chart to monitor mean and variance together