EWMA p charts for detecting changes in mean or scale of normal variates

Methods of Statistical Process Control (SPC) are used for detecting deviations from regular processing. SPC is applied in manufacturing implementations where statistical tools are used to monitor the performance of production processes in order to identify and correct considerable changes in the process performance. Today, SPC methods are incorporated by organizations around the world as a suitable tool to improve product quality by reducing process variation. The current method of SPC is the application of control charts which are used to monitor process parameters (e. g., mean μ, standard deviation σ or percent defective p) over time. Well-established control chart schemes are, amongst others, exponentially weighted moving average charts (EWMA), cumulative sum charts (CUSUM) or, of course, the classical Shewhart charts. In this article, an EWMA control chart for variables calculating the percent defective p = f (μ, σ) will be presented where both process parameters are under risk to change. The scheme will be compared to several other control chart applications (EWMA X, EWMA X-S2, and an alternative EWMA p chart). Numerical methods and Monte Carlo simulations are used for computing the average run length (ARL) as the measure of performance.