Just‐in‐time reorganized PCA integrated with SVDD for chemical process monitoring

Abstract

Although principal component analysis (PCA) is widely used for chemical process monitoring, improvements in the selection of principal components (PCs) are still needed. Given that the determination of complicated and changing fault information is not guaranteed using offline-selected PCs, this study proposes a just-in-time reorganized PCA model that objectively selects the PCs online for process monitoring. The importance of the PCs is evaluated online by kernel density estimation. The PCs indicating more varied information are then selected to reorganize the PCA model. Given that the most useful fault information is concentrated, support vector data description is used to replace traditional statistics, thereby relaxing the Gaussian assumption of the process data. The monitoring performances of the proposed method are evaluated under three cases. Compared with conventional PCA methods, more varied information is captured online, and the monitoring performances are improved. V C 2014 American Institute of Chemical Engineers AIChE J, 60: 949-965, 2014

GMM and optimal principal components-based Bayesian method for multimode fault diagnosis

“…After the PCA transformation, the obtained PCs are uncorrelated with each other and the bins in evidence equal to the number of the PCs retained. It has been pointed out in literatures that fault has no definite mapping to a certain PC and the importance of PCs is not necessarily in the variance top-down order (Jiang and Yan, 2014a;Jiang et al, 2013). The PCs with smaller variance may carry more important fault information.…”

“…Operation modes of the TE process (Chiang et al, 2001;Ricker, 1996 Table 3 Monitored variables in the TE process (Chiang et al, 2001;Jiang and Yan, 2014a Fig. 8.…”

Nonlinear plant-wide process monitoring using MI-spectral clustering and Bayesian inference-based multiblock KPCA

“…Step IDV (8) A, B, C feed composition (stream 4) Random variation IDV (9) D feed temperature (stream 2) Random variation IDV (10) C feed temperature (stream 4) Random variation IDV (11) Reactor cooling water inlet temperature Random variation IDV (12) Condenser cooling water inlet temperature Random variation IDV (13) Reaction kinetics Slow drift IDV (14) Reactor cooling water valve Sticking IDV (15) Condenser cooling water valve Sticking IDV (16) Unknown Unknown IDV (17) Unknown Unknown IDV (18) Unknown Unknown IDV (19) Unknown Unknown IDV (20) Unknown Unknown IDV (21) The valve for stream 4 was fixed at the steady state position…”

“…As a result, data-based methods, especially the multivariate statistical process monitoring (MSPM), have now become very popular [4][5][6][7][8][9]. Among these MSPM methods, principal component analysis (PCA) usually serves as the most fundamental technique and has been extensively used [4,[10][11][12][13]. PCA projects high-dimensional and highly correlated data onto a lower-dimensional space that contains the most variance of the original data, and simultaneously de-correlates the data.…”

Adaptive database management based on the database monitoring index for long-term use of adaptive soft sensors

“…A local model can be constructed with a nonlinear regression method such as kernel partial least squares regression [21] and kernel principal component regression [22]. JIT modeling can be applied to process monitoring [23]. A TD soft sensor uses the temporal differences between y-variable values and those of the corresponding X-variables and adds a known y-value to the TD of the y-variable to form a prediction [24].…”