Principal Component Chart for Multivariate Statistical Process Control

Abstract

Multivariate statistical process control technique (Hotelling T2 chart) was used to monitor four correlated quality characteristics (active detergent, moisture content, bulk density and ph level) of detergent produced by a company which indicated out-of-control signal. Principal Component Chart is used as a follow-up to out-of-control signal of the Multivariate Control Chart, to identify the quality characteristic(s) that contributed to the signal. The component scores obtained from the principal component analysis of the four quality characteristics measured were used to identify the quality characteristic(s) that contributed to the out-of-control signaled by the Hotelling T2 chart. The chart of the first component which accounted for 96.7% of the total variability and has moisture content highly loaded in it is out-of-control, which implied that moisture content of the detergent produced by the company is out-of-control.