Journal of the Korean Institute of Intelligent Systems (한국지능시스템학회논문지)

Korean Institute of Intelligent Systems (한국지능시스템학회)
권호 표지
DOI QR코드

DOI QR Code

Modeling and Comparison for Auto-association using Support Vector Regression (SVR) and Partial Least Square Regression (PLSR) in Online Monitoring Techniques

상시감시기술에서 SVR과 PLSR을 이용한 Auto-association 모델링 및 성능비교

  • 김성준 (강릉원주대학교 산업공학과) ;
  • 서인용 (한국전력공사 전력연구원)
  • Received : 2010.02.17
  • Accepted : 2010.07.31
  • Published : 2010.08.25
Download PDF
⟨ Previous Next ⟩

Abstract

An online monitoring based upon sensor system is essential to assure both efficient operation and safety in the power plant. Of great importance is modeling for auto-association (AA) in online monitoring technique. The objective of auto-associative models lies in predicting true values of plant operation parameters from sensor signals transmitted. This paper presents two AA models using Support Vector Regression (SVR) and Partial Least Square Regression (PLSR). The presented models are useful, in particular, when there are many parameters to monitor in the power plant. Illustrative examples are given by using a real-world plant dataset. AA performances of SVR and PLSR are finally summarized in terms of accuracy and sensitivity. According to our results, SVR shows much higher accuracy and, however, its sensitivity is relatively degraded.

센서시스템을 이용한 상시감시는 발전소의 효율적인 운전과 안전을 담보하는 데 필수적이다. 상시감시기술을 구현하기 위해서는 우선 센서로부터 전송된 신호로부터 발전소 운전파라미터의 참값을 예측하는 모델 즉 Auto-association (AA) 모델을 확보하는 것이 중요하다. 이를 위해 본 논문에서는 Support Vector Regression (SVR)과 Partial Least Square Regression (PLSR)을 이용하는 방안을 각각 제시한다. 이렇게 해서 구축된 모델은 모니터해야 할 파라미터가 많을 때에도 쉽게 적용할 수 있다. 실제 발전소에서 수집된 데이터셋을 이용하여 AA 모델링의 정확도 및 민감도를 비교한 결과, 정확도 면에서는 SVR이 우수한 반면 민감도 면에서는 PLSR이 다소 나은 것으로 나타났다.

Keywords

References

  1. J. W. Hines, D. Garvey, R. Seibert and A. Usynin, Technical Review of On-line Monitoring Techniques for Performance Assessment, Vol. 2, Technical Report, U. S. Nuclear Regulatory Commission, 2007.
  2. D. Wrest, J. W. Hines and R. Uhrig, "Instrument Surveillance and Calibration Verification Through Plant Wide Monitoring Using Autoassociative Neural Networks," Proceedings of The 1996 American Nuclear Society International Meeting on Nuclear Plant Instrumentation, Control and Human Machine Interface Technologies, University Park, PA, 1996.
  3. J. W. Hines and A. Usynin, "MSET Performance Optimization Through Regularization," Nuclear Engineering and Technology, Vol. 38, No. 2, pp. 177-184, 2005.
  4. V. Vapnik, Statistical Learning Theory, Wiley, New York, 1998.
  5. L. Zheng, H. Zhou, K. Cen and C. Wang, "A Comparative Study of Optimization Algorithms for Low NOx Combustion Modification at a Coal-fired Utility Boiler," Expert Systems with Applications, Vol. 36, pp. 2780-2793, 2009. https://doi.org/10.1016/j.eswa.2008.01.088
  6. 김성준, “SVM을 이용한 웨이블릿기반 프로파일 분류에 관한 연구,” 한국지능시스템학회논문지, 18권, 5호, pp. 718-723, 2008.
  7. F. J. Martinez-de-Pison, C. Barreto, A. Pernia and F. Alba, "Modelling of an elastomer profile extrusion process using support vector machines," Journal of Materials Processing Technology, Vol. 197, pp. 161-169, 2008. https://doi.org/10.1016/j.jmatprotec.2007.06.025
  8. 전성해, “Support Vector Regression을 이용한 이상치 데이터분석,” 한국지능시스템학회논문지, 18권, 6호, pp. 876-880, 2008.
  9. R. Rosipal, M. Girolami, L. Trejo and A. Cichocki, "Kernel PCA for Feature Extraction and Denoising in Nonlinear Regression," Neural Computing & Applications, Vol. 10, pp. 231-243, 2001. https://doi.org/10.1007/s521-001-8051-z
  10. 김성준, 서인용, "센서기반 상시감시를 위한 Auto-association 모델," 대한산업공학회 추계학술회의논문집, 한양대학교, A07-3, 2008.
  11. 이대종, 박상영, 정남정, 이혜근, 박진일, 전명근, “PLS기반 c-퍼지 모델트리를 이용한 클로로필-a 농도 예측,” 한국지능시스템학회논문지, 16권, 6호, pp. 777-784, 2006. https://doi.org/10.5391/JKIIS.2006.16.6.777
  12. B. S. Dayal and J. F. MacGregor, "Improved PLS Algorithm," Journal of Chemometrics, Vol. 11, pp. 73-85, 1997. https://doi.org/10.1002/(SICI)1099-128X(199701)11:1<73::AID-CEM435>3.0.CO;2-#
  13. C. Chang and C. Lin, http://www.csie.ntu.edu.tw/~cjlin/libsvm/, 2000.
  14. M. G. Na, H. Y. Yang and D. H. Lim, "A Soft-Sensing Model for Feedwater Flow Rate Using Fuzzy Support Vector Regression," Nuclear Engineering and Technology, Vol. 40, No. 1, pp. 69-76, 2007. https://doi.org/10.5516/NET.2008.40.1.069