Monitoring and Evaluation of Clinker Quality Using T2 Hotelling-Generalized Variance Control Chart

Abstract

The cement industry is an important sector in infrastructure development, where the quality of clinker determines the final quality of the product. This study evaluates the application of T² Hotelling's and Generalized Variance (GV) multivariate control charts to clinker data based on three main variables: FCaO, C₃S, and C₃A at PT XYZ. The results show that C₃S has the highest variance in phase I and II (2.61 and 2.53), while FCaO has the lowest variance (0.10 and 0.06). All three variables had mean values within the specification limits, although there were still extreme values outside the limits. Assumption tests showed that the data was not multivariate normally distributed, but it was still assumed to be normal for control analysis purposes. In the wet season, the standard deviation decreased from 1.552 to 1.252, and in the dry season from 1.170 to 1.029, indicating a decrease in variability although the process is not yet fully under statistical control. Capability analysis shows that the dry season process is more stable, with most parameters having multivariate values that exceed the threshold. Compared to the wet season, the dry season process showed more consistent performance and was able to meet production quality standards.