A Novel Approach to Multivariate Process Control Using Sigma-Scaled Metrics

Abstract

This study presents a new methodology for multivariate statistical process control (MSPC) that leverages sigma-scaled metrics to improve detection sensitivity and reduce false alarm rates in manufacturing environments. Traditional MSPC methods often struggle with high-dimensional data and non-linear relationships between process variables. Our proposed approach introduces a novel sigma-scaling technique that normalizes multivariate data while preserving critical variance structures. Through extensive simulation studies and real-world case studies from automotive manufacturing, we demonstrate that our method achieves superior performance compared to conventional Hotelling T² and MEWMA control charts. The results show a 32% reduction in false positives while maintaining 95% detection power for process shifts as small as 1.5 sigma. This research contributes to the growing body of knowledge on advanced statistical process control and provides practitioners with a robust tool for modern manufacturing quality control.

Keywords

Article Timeline