Improved performance of MCUSUM control chart with autocorrelation

Abstract

In recent years, the importance of quality has become increasingly apparent, and quality control in manufacturing has moved from detecting nonconforming products through inspection to detecting quality abnormalities in the process using statistical process control [1]. where it is used effectively, SPC plays an important role in reducing variation in manufactured items and in increasing the competitiveness of the manufacturer by improving product quality while at the same time decreasing production costs. Charts like the Shewhart X and R charts have found wide use in industry because of their ease of use for technicians and others with minimal training in statistics, since the calculations and plotting can be done by hand. An MCUSUM control chart was constructed with autocorrelated data at different levels of autocorrelation and found to be ineffective in detecting shifts as it occurs. In this article, we have proposed new techniques that can improve the performance of the MCUSUM with autocorrelation using run rule schemes. The techniques was evaluated using ARL measures of performance with 10000 iterations to simulate. The results showed that the performance of MCUSUM with autocorrelation has improved significantly with the new technique which was compared to the existing conventional MCUSUM control chart.