Decentralized Fault Diagnosis of Large-scale Processes Using Multiblock Kernel Principal Component Analysis

Modified kernel principal component analysis based on local structure analysis and its application to nonlinear process fault diagnosis

“…(18) to (20), and establish the control limits for the monitoring statistics, T 2 and Q, using the kernel density estimators (21) and (22) based on the sample data {T…”

“…A, B and C feed compositions (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 shift IDV (14) Reactor cooling water valve Sticking IDV (15) Condenser cooling water valve Sticking IDV (16)- (20) Unknown Unknown IDV (21) Valve position constant (stream 4) Constant position fault detection time and fault alarming rate. Fault detection time is defined as the first sample number after previous eight consecutive samples have exceeded the confidence limit, while fault alarming rate is defined as the percentage of the alarming samples in all the fault samples.…”

“…Nguyen and Golinval [19] applied the KPCA to detect mechanical system faults by comparing the subspace angle between a reference and the current state. Considering the dynamic property of process data, Jia and co-authors [20] developed a dynamic KPCA method by integrating the kernel PCA and ARMAX time series model, while Zhang et al [21] discussed a multi-block KPCA for large-scale processes. For effective monitoring of nonlinear and nonstationary processes, the work [22] proposed a variable window adaptive KPCA based on a fast block adaptation.…”

Root cause analysis of process fault conditions on an industrial concentrator circuit by use of causality maps and extreme learning machines

“…(Venkatasubramanian et al, 2003b). To this end, qualitative modelbased approaches such as signed digraphs (Iri et al, 1979;Kramer and Palowitch, 1987;Ahn et al, 2008) and hierarchical system decomposition (Douglas, 1985;Finch and Kramer, 1987;Lind, 1991;Bie and Wang, 2009;Zhang et al, 2010) not only allow the causal structure of a system to be captured, but also to reason from it.…”

Fault Detection of Nonlinear Processes Using Fuzzy C-means-based Kernel PCA

“…MONG the existing nonlinear methods, kernel-based techniques have been successfully developed for tackling the nonlinear problem in recent years [1]. They have attracted wide attentions, including support vector machine (SVM) [2], [3], and [4], kernel principal component analysis (KPCA) [5], [6], [7], [8], and [9], kernel partial least squares (PLS) [10], [11], and [12], kernel fisher discriminant analysis (FDA) [13], [14], [15], and [16] and Kernel Independent Component Analysis (KICA) [17], [18], and [19]. The basic idea is that the mapped data are analyzed using conventional linear statistical analysis techniques in high dimensional feature space, which is equivalent to nonlinear analysis in original input space [20].…”