Chemical processes monitoring based on weighted principal component analysis and its application

Preliminary-summation-based principal component analysis for non-Gaussian processes

“…The Tennessee Eastman process simulation was developed by Downs and Vogel (1993) 25,26 , which simulates industrial processes containing both non-Gaussian and nonlinear features. As a benchmark simulation, the TE process has been widely used to test the performance of various monitoring approaches.…”

Fault Detection and Diagnosis in Chemical Processes Using Sparse Principal Component Selection

“…Contrarily, if h is too small, sharp peaks will be positioned at the sample points in the density estimator. The number of data points, the data distribution, the choice of the kernel function, and other several factors decide the optimal choice of h. Various approaches on the choice of h have been reported, in which a proper empirical value for each specific case is suggested Jiang and Yan (2012), Jiang et al (2013), Scott (1992), Webb et al (2011).…”

“…With the the monitoring plots of different cases, as well as the ℝOC curves plotted in Figure 14 and Figure 18, it is concluded that the monitoring performance of SPCS usinng T 2 statistic is the best. In order to give a more comprehensive comparison, PCA, WPCA (weighted principal component analysis) Jiang and Yan (2012), SPCA (senstive principal component analysis) Jiang et al (2013) and we propsoed SPCS are performed on each case from Fault 1 to Fault 21. We selected 27 PCs for PCA according to CPV≥85% and set Card(β)=27 for SPCS.…”

“…Many algorithms for selecting principal components have been put forward, such as, cumulative percent variance (CPV) Jiang and Yan (2012), variance of reconstruction error (VℝE) Lu et al (2004) and cross validation (CV) Diana and Tommasi (2002). CPV selects the first several PCs that represent the major variance information of original data.…”

Fault Detection of a Proposed Three-Level Inverter Based on a Weighted Kernel Principal Component Analysis

“…However, for a certain KPC, fault information that is quite different from the others and the KPCs are not of equal importance. If the information is relatively limited, then the useful KPCs could be suppressed by the useless ones, thus submerging the useful information [15], [16]. Therefore, adapting covariance matrix to different classes of information is necessary.…”