Detection of signal drifts by different control charts

AbstractQuality control charts like Shewhart CUSUM, MA and EWMA control charts are widely used in the industry. Control charts are often implemented in statistical process control software packages and are used in conditioning monitoring for fault detection. The functional principle of the control charts is shown and explained through a simple example. The sensitivity of the different control charts to mean value and standard deviation shifts and different parameters are compared. And it is shown that the CUSUM control chart is the best. With some different applications it is shown how easy is it to detect a shift or drift in industrial data. The applications presented are gas concentration measurement and milk production. Also the problem to calculate the correct standard deviation for a sensor system is dealt with.