Assessing the Impact of Parametric Uncertainty on the Performance of Model-Based PCA (original) (raw)

Abstract

In model-based PCA (MBPCA), principal component analysis is carried out on the residual process measurements that cannot be predicted using a physical model. In principle, this approach can improve the detection and identification of unmeasured disturbances and faults in non-stationary and batch processes. Since process knowledge is required for the implementation of MBPCA, the uncertainty associated with the process model will inevitably affect the achievable performance. This paper presents a method for estimating the impact of parametric uncertainty on the performance of MBPCA and demonstrates its application on the monitoring of a continuous stirred tank reactor.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

References (20)

  1. Allgower F. and I1chmann A., "Multi variable Adaptive A-tracking for Non-linear Chemical Processes," Proc. of 3rd ECC, 3, 1645-1656 (1995).
  2. Berlinet A. and Devroye L. , "A comparison of Ker- nel Density Estimates," Publications de I'Institut de Statistique de I'Universite de Paris, 38(3) 3-59 (1994).
  3. Chen 1., Bandoni A. and Romagnoli J. A., "Robust PCA and Normal Region in Multivariate Statis- tical Process Monitoring," AIChE J., 42(12), 3563-3566 (1996).
  4. Dong D. and McAvoy T. J., "Batch Tracking via Nonlinear Principal Component Analysis," AIChE J., 42(8), 2199-2208 (1996).
  5. Dunia R. and Qin S. 1., "A Unified Geometric Approach to Process and Sensor Fault Identifi- cation and Reconstruction: the Unidimensional Fault Case," Comp. and Chem. Engng., 22, 927-943 (1998).
  6. Gazi E., Ungar L.H. and Seider W.D., "Automatic Analysis of Monte-Carlo Simulations of Dynamic Chemical Plants," Computers Chem. Engng, 20(S), S987-S992 (1996).
  7. Hiden H. G., Willis MJ., Tham M.T. and Montague G. A., "Non-linear Principal Components Analysis using Genetic Program- ming," Computers Chem. Engng, 23, 413-425 (1999).
  8. Kassidas A., MacGregor J. F. and Taylor P. A., "Synchronization of Batch Trajectories Using Dynamic Time Warping," AIChE J., 44(4), 864-875 (1998).
  9. Kosanovich K. A., Dabl K. S. and Piovoso M. 1., "Improved Process Understanding using MuIti- way Principal Component Analysis," lnd Eng. Chem. Res., 35, 138-146 (1996).
  10. Kramer M. A., ''Nonlinear PCA using Autoasso- ciative Neural Networks," AIChE J., 37(2), 233-243 (1991).
  11. Lewin D. R, "Predictive Maintenance using PCA," Control Eng. Prac., 3(3),412-421 (1995).
  12. Martin E. 8. and Morris A. 1., ''Non-parametric Confidence Bounds for Process Performance Monitoring Charts," J. Proc. Cont. 6(6), 349-358 (1996).
  13. Nomikos P. and MacGregor J., "Monitoring Batch Processes using Multiway PCA," AIChE J., 40(8),1361-1375 (1994).
  14. Robertson T., Chen J., Romagnoli J.A. and Newell 8., "Intelligent Monitoring for quality Control in Biological Nutrient Removal Waste water Treat- ment Pilot Plant," Proc. of the 5th IFAC Sympo- sium of Dynamics and Control of Process Sys- tems, Corfu, 590-595 (1998).
  15. Rotem Y., Wachs A. and Lewin D. R, "Ethylene Compressor Monitoring using Model-Based PCA," Submitted to AIChE J. (1999).
  16. Rothwell S.G., Martin E.B. and Morris AJ. , "Comparison of Methods for Handling Unequal Length Batches," Proc. of the 5th IFAC Sympo- sium of Dynamics and Control of Process Systems, Corfu, 66-71 (1998).
  17. Scott D.W., Multivariate Density Estimation: Theory Practice and Visualization, Wiley, New York (1992).
  18. Wachs A. and Lewin D. R. (1998), "Process Moni- toring using Model-based PCA," Proc. of the 5th IFAC Symposium of Dynamics and Control of Process Systems, Corfu, 86-91 (1998).
  19. Wachs A. and Lewin D. R, "Improved PCA Meth- ods for Process Disturbance and Failure Identifi- cation," AIChE Jnl., 45(8), 1688-1700 (1999).
  20. Wo Id S., Geladi P., Esbensen K., "Multi-way Prin- cipal Components and PLS-analysis," J. ofChe- mometrics, I, 41-56 (1987).