Performance Evaluation of an EWMA p Chart Based on Improved Square Root Transformation to detect Small Shift Process Variation

Abstract

A control chart was used to detect whether the unexpected variation due to special causes was happening. Generally, Shewhart control chart was used, while this chart was not a good one to control the process with small variation. EWMA p chart was one of the alternatives used to monitor small shifts in the process mean. Traditionally, attribute Shewhart p chart has been developed for charting the binomial data or using the normal approximation when percent defective was low and sample size was small. In 100% inspection, these assumptions were not valid. For this reason, EWMA p control chart was modified by improving square root transformation. In this research, the implementation of the modified EWMA p control chart based on Improved square root transformation (ISRT p EWMA chart) was studied. This chart was used to monitor the actual production process where the process has shifted. The research found that. The ISRT p EWMA chart detects the shift twice, while Shewhart p chart does not detect any shift. The research also found that at λ = 0.25, the shift is detected 2 times, at λ = 0.5, the shift is detected once while at λ = 0.75, no shift is detected. The larger the value of the λ parameter, the worse the ISRT p EWMA control chart performed.