Monitoring of chemical processes using improved multiscale KPCA

Statistical process monitoring charts are critical in ensuring safety for many chemical processes. Principal Component Analysis (PCA) is often used, due to its computational simplicity. However, many chemical processes may be inherently nonlinear, and this degrades the performance of the linear PCA method. Kernel Principal Component Analysis (KPCA) is an extension of the conventional PCA chart, which can help deal with nonlinearity in a given process. Additionally, PCA assumes that process data are Gaussian and uncorrelated, and only contain a moderate level of noise. These assumptions do not usually hold in practice. Multiscale wavelet-based data representation produces wavelet coefficients that possess characteristics that are able to handle violations in these assumptions. A multiscale kernel principal component analysis (MSKPCA) method has already been developed to tackle all of these issues, but it usually provides a high false alarm rate. In this paper, an improved MKSPCA chart is developed in order to deal with the false alarm rate issue, by smoothening the detection statistic using a mean filter. The advantages brought forward by the improved method are demonstrated through a simulated example in which the developed fault detection method is used to monitor a continuous stirred tank reactor (CSTR). The results clearly show that the improved MSKPCA method provides lower missed detection and false alarm rates as well as ARL1 values compared to those provided by the conventional methods.

[1]  Bernhard Schölkopf,et al.  Kernel Principal Component Analysis , 1997, ICANN.

[2]  Hazem N. Nounou,et al.  Enhanced monitoring using PCA-based GLR fault detection and multiscale filtering , 2013, 2013 IEEE Symposium on Computational Intelligence in Control and Automation (CICA).

[3]  Stéphane Mallat,et al.  A Theory for Multiresolution Signal Decomposition: The Wavelet Representation , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  C. Yoo,et al.  Nonlinear process monitoring using kernel principal component analysis , 2004 .

[5]  Tian Xuemin,et al.  A fault detection method using multi-scale kernel principal component analysis , 2008, 2008 27th Chinese Control Conference.

[6]  Hazem Nounou,et al.  Statistical fault detection using PCA-based GLR hypothesis testing , 2013 .

[7]  J. E. Jackson,et al.  Control Procedures for Residuals Associated With Principal Component Analysis , 1979 .

[8]  Mu Zhu,et al.  Automatic dimensionality selection from the scree plot via the use of profile likelihood , 2006, Comput. Stat. Data Anal..

[9]  Mohammed Ziyan Sheriff.  Improved Shewhart Chart Using Multiscale Representation , 2015 .

[10]  Mohamed N. Nounou,et al.  Enhanced performance of shewhart charts using multiscale representation , 2016, 2016 American Control Conference (ACC).

[11]  Hazem Nounou,et al.  Statistical Fault Detection of Chemical Process - Comparative Studies , 2015 .

[12]  T. McAvoy,et al.  Nonlinear principal component analysis—Based on principal curves and neural networks , 1996 .

[13]  Hazem Nounou,et al.  Improving the prediction and parsimony of ARX models using multiscale estimation , 2007, Appl. Soft Comput..

[14]  B. Bakshi.  Multiscale PCA with application to multivariate statistical process monitoring , 1998 .

[15]  Bhavik R. Bakshi,et al.  Multiscale Methods for Denoising and Compression , 2000 .

[16]  Giancarlo Diana,et al.  Cross-validation methods in principal component analysis: A comparison , 2002 .

[17]  Tapas K. Das,et al.  Wavelet-based multiscale statistical process monitoring: A literature review , 2004 .

[18]  Bart De Ketelaere,et al.  A systematic comparison of PCA-based statistical process monitoring methods for high-dimensional, time-dependent processes , 2016 .

[19]  Hiromu Ohno,et al.  Dimensionality reduction for metabolome data using PCA, PLS, OPLS, and RFDA with differential penalties to latent variables , 2009 .

[20]  Hazem N. Nounou,et al.  Detecting abnormal ozone levels using PCA-based GLR hypothesis testing , 2013, 2013 IEEE Symposium on Computational Intelligence and Data Mining (CIDM).