Application of Wavelets and Principal Component Analysis to Process Quantitative Feature Extraction

A two-step feature extraction approach combining wavelet transform and principal component analysis (PCA) is presented. Wavelet transform provides a compact, information-rich expression of process data through a set of coefficients that carry localized transient information of process operating condition. PCA is used to reduce the dimension of correlated coefficients in an optimal way. Case studies on the Tennessee Eastman process illustrate that the proposed method is able to capture the inherent characteristics from process measurements.