Quality-Related Process Monitoring Based on Improved Kernel Principal Component Regression

To date, quality-related multivariate statistical methods are extensively used in process monitoring and have achieved admirable effects. However, most of them contain recursive processes, which result higher time complexity not suitable for increasingly complex industrial processes. Therefore, this paper embeds singular value decomposition (SVD) into the kernel principal component regression (KPCR) to accomplish Quality-related with a lower computational cost. Specifically, technique is devoted map original input dimensional space boost nonlinear ability (PCR), then KPCR employed capture correlation between matrix output matrix. At same time, kernelized decomposed two orthogonal quality-unrelated spaces by SVD, statistics calculated detect faults respectively. Compared other methods, it has following advantages: 1) A analysis (QR-KPCR) algorithm proposed. 2) partial least squares method, omitted training shortened. 3) The model more concise fault detection faster. 4) By contrast monitoring, rate. Experimental results on widespread example an industry benchmark verify effectiveness reliability proposed method.