摘要：A number of processes to which statistical control is applied are subject to various effects that cause random changes in the mean value. The removal of these fluctuations is either technologically impossible or economically disadvantageous under current conditions. The frequent occurrence of signals in the Shewhart chart due to these fluctuations is then undesirable and therefore the conventional control limits need to be extended. Several approaches to the design of the control charts with extended limits are presented in the paper and applied on the data from a real production process. The methods assume samples of size greater than 1. The performance of the charts is examined using the operating characteristic and average run length. The study reveals that in many cases, reducing the risk of false alarms is insufficient.
