Fault Detection Approach Based on Weighted Principal ComponentAnalysis Applied to Continuous Stirred Tank Reactor

摘要

Fault detection approach based on principal component analysis (PCA) may perform not well when the processis time-varying, because it can cause unfavorable influence on feature extraction. To solve this problem, a modified PCAwhich considering variance maximization is proposed, referred to as weighted PCA (WPCA). WPCA can obtain the slowfeatures information of observed data in time-varying system. The monitoring statistical indices are based on WPCAmodel and their confidence limits are computed by kernel density estimation (KDE). A simulation example on continuousstirred tank reactor (CSTR) show that the proposed method achieves better performance from the perspective of both faultdetection rate and fault detection time than conventional PCA model.