Soft Sensor Modeling and Application Based on Kernel Function Principal Component Analysis

A novel method is proposed in developing soft sensor based on the nonlinear principal components obtained from kernel principal component analysis (PCA) technique to deal with the nonlinear characteristics of data. Field application shows that the proposed method is effective and superior to traditional methods.