The use of multivariate SPC methods in the analysis of quality control problems: An application in a textile company

In today's globalized world, quality is a crucial factor for businesses striving to survive in highly competitive markets and meet customer demands. When product quality deviates from set standards during production, it's essential to intervene promptly to prevent potential losses. Deviations can drive up production costs and reduce process efficiency. Therefore, maintaining constant control over production processes, promptly identifying deviations, and implementing necessary corrective measures can deliver significant time and cost savings for businesses. Considering these factors, integrating statistical process control (SPC) methods into production processes is essential for boosting quality and efficiency. In the textile industry, it is observed that multivariate statistical process control (MSPC) methods are not sufficiently implemented and studied in the domestic literature, especially in the context of yarn production processes. The aim of this study is to fill this gap in the domestic literature and demonstrate the effectiveness of MSPC methods in quality control processes in yarn production. In line with this aim, multivariate statistical process control methods such as Hotelling's T2 control chart, the multivariate cumulative sum (MCUSUM) method, and the multivariate exponentially weighted moving average (MEWMA) control chart have been thoroughly examined. The application of Hotelling T2 method have been implemented in the yarn production facilities of a textile company. During the application phase, the quality control laboratory data for the top three types of yarn produced by the company were utilized. These data were initially analyzed using single-variable quality control charts, and subsequently, the quality variables were grouped, and the process was observed using advanced multivariate statistical process control methods. The study's findings revealed that the multivariate statistical process control methods offer a more effective and comprehensive control by considering the interrelationships between multiple quality variables. The calculations and control charts for the study were conducted using the Minitab Statistical Software.