The SPC Evaluation Of Auto-Correlated Process

Traditional statistical process control (SPC) theory is based on a fundamental assumption that the process data are statistically independent. However, in many conditions the process data are auto-correlated, not independent. Under these circumstances, because of serial correlation, the monitoring capacity of quality control chart will be affected heavily, and it increases the probability of false alarm in normal process. Even worse, it will leave out the real alarm in abnormal process. This paper researches in the serial correlation of quality control charts and will improve the theory of quality control chart. Furthermore, it provides guidance in using quality control charts in reality directly.The basic theory and knowledge of the statistical process control (SPC), traditional control chart, auto-correlation and time series are first introduced in this paper. And then gives two methods of dealing with auto-correlated process: modified chart and residual chart.The advantage of the modified chart is that it doesn't need time series models to fit the measured value, as long as the process in a stationary process. Based on the auto-covariance function and auto-correlation function, the ultimate control limits can be got, then gain the control chart of the auto-correlated process.Residual control chart is the method of using the residuals of time series models to control the quality of the auto-correlated process. A case study shows that the residual control chart is more suitable for auto-correlated process than the traditional control chart. The residual control chart can detect out-of-control processes more quickly when the auto-correlated process parameter is larger than 0.At last, the actual production data is used to verify the correctness of residual control chart. For auto-correlated process, a residual control chart is suggested instead of traditional control charts.