MULTIVARIATE STATISTICAL PROCESS CONTROL AND FAULT DETECTION (Applications in industrial film production and paper making)

摘要

This paper reviews Multivariate Statistical Process Control and its application to two industrial processes for performance monitoring - high speed polyester film production and paper making. Univariate statistical process control is widely applied in industry but it effectively only detects disturbances related to individual measurement sources (usually quality) and consequently provides little information about the interactions between variables which are important for the understanding of complex manufacturing processes. Multivariate Statistical Process Control provides a diagnostic tool for the comprehensive monitoring and detection of process malfunctions for both continuous and batch manufacturing processes.