%0 Journal Article %1 cho2005fault %A Cho, Ji-Hoon %A Lee, Jong-Min %A Choi, Sang Wook %A Lee, Dongkwon %A Lee, In-Beum %D 2005 %J Chemical Engineering Science %K PCA kernel_PCA methods %N 1 %P 279 - 288 %R http://dx.doi.org/10.1016/j.ces.2004.08.007 %T Fault identification for process monitoring using kernel principal component analysis %U http://www.sciencedirect.com/science/article/pii/S0009250904006013 %V 60 %X In this research, we develop a new fault identification method for kernel principal component analysis (kernel PCA). Although it has been proved that kernel PCA is superior to linear PCA for fault detection, the fault identification method theoretically derived from the kernel PCA has not been found anywhere. Using the gradient of kernel function, we define two new statistics which represent the contribution of each variable to the monitoring statistics, Hotelling's T 2 and squared prediction error (SPE) of kernel PCA, respectively. The proposed statistics which have similar concept to contributions in linear PCA are directly derived from the mathematical formulation of kernel PCA and thus they are straightforward to understand. The main contribution of this work is that we firstly suggest a fault identification method especially applicable to process monitoring using kernel PCA. To demonstrate the performance, the proposed method is applied to two simulated processes, one is a simple nonlinear process and the other is a non-isothermal CSTR process. The simulation results show that the proposed method effectively identifies the source of various types of faults.

@article{cho2005fault, abstract = {In this research, we develop a new fault identification method for kernel principal component analysis (kernel PCA). Although it has been proved that kernel PCA is superior to linear PCA for fault detection, the fault identification method theoretically derived from the kernel PCA has not been found anywhere. Using the gradient of kernel function, we define two new statistics which represent the contribution of each variable to the monitoring statistics, Hotelling's T 2 and squared prediction error (SPE) of kernel PCA, respectively. The proposed statistics which have similar concept to contributions in linear PCA are directly derived from the mathematical formulation of kernel PCA and thus they are straightforward to understand. The main contribution of this work is that we firstly suggest a fault identification method especially applicable to process monitoring using kernel PCA. To demonstrate the performance, the proposed method is applied to two simulated processes, one is a simple nonlinear process and the other is a non-isothermal CSTR process. The simulation results show that the proposed method effectively identifies the source of various types of faults. }, added-at = {2015-02-10T00:32:46.000+0100}, author = {Cho, Ji-Hoon and Lee, Jong-Min and Choi, Sang Wook and Lee, Dongkwon and Lee, In-Beum}, biburl = {https://www.bibsonomy.org/bibtex/2635126c5172ae8f5023d3208aa160d89/peter.ralph}, doi = {http://dx.doi.org/10.1016/j.ces.2004.08.007}, interhash = {9f4dd2d0b55f875649266efd4851da1d}, intrahash = {635126c5172ae8f5023d3208aa160d89}, issn = {0009-2509}, journal = {Chemical Engineering Science }, keywords = {PCA kernel_PCA methods}, number = 1, pages = {279 - 288}, timestamp = {2015-02-10T00:32:46.000+0100}, title = {Fault identification for process monitoring using kernel principal component analysis }, url = {http://www.sciencedirect.com/science/article/pii/S0009250904006013}, volume = 60, year = 2005 }