An application of time series analysis methods in process control

Methods for monitoring level shifts exist in time series since the sixties. Triff-Leach, Kirkendall and Jones provide three different points of view about the same subject. According to the characteristics of the presented methods, they can be easily applied to control processes involving a single productive line, as an alternative to quality control techniques. In this work, we study the behaviour of these methods with observations that follow three time series models, AR(1), MA(1) and ARMA(1,1) with level shifts artificially created of magnitudes considered significant for process control.