화학공학소재연구정보센터
홈 로그인 로그아웃 연락처 사이트맵
Journal of the Korean Industrial and Engineering Chemistry, Vol.13, No.6, 544-550, October, 2002
Export Citation
퍼지 신경망을 이용한 pH 중화공정의 적응제어 시스템
Adaptive Control System of pH Neutralization Process Using Fuzzy Neural Networks
김종대†, 유민수1
  • 한전전력연구원, 대전 305-380
  • 1전남대학교 공과대학 응용화학공학부/촉매연구소, 광주 500-757
Jong Dae Kim†, and Min Su Ryu1
  • KEPCO Electric Power Research Institute, Daejeon 305-380, Korea
  • 1Faculty of Applied Chemicl Engineering and The Research Institute for Catalysis, Chonnam National University, Kwangju 500-757, Korea
E-mail:
초록
신경망 적응제어의 제어성능을 향상하기 위해 신경망 제어기의 학습율을 공정의 동특성에 따라 조절할 수 있는 퍼지논리를 개발하고, 비선형공정인 0.001 M 초산/수산화나트륨의 pH 중화공정에 선형모델예측과 함께 신경망 적응제어에 적용하였다. 그 결과, 초기 제어상태에서 일반적인 신경망 적응제어보다 큰 오버슈트와 진동이 많이 감소되어 보다 안정된 제어응답을 얻었다. 퍼지 신경망 적응제어의 제어응답이 목표점에 수렴한 안정된 상태일 때 최대오차는 ±pH 0.02 정도였다. 또한, pH 6~10 사이의 목표점에 모두 수렴되어 넓은 제어 범위에서 양호한 제어성능을 보였다. pH 제어 실험에서 제안한 퍼지 신경망 적응제어는 신경망 적응제어의 불안정한 제어응답을 효과적으로 개선할 수 있었다.
In order to improve the control performance of the neural adaptive control, we developed a fuzzy logic which is able to adjust the learning rate of the neural controller for process dynamics and it was applied to the neural adaptive control with linear model prediction for the pH neutralization process of 0.001 M CH3COOH/NaOH as a nonlinear process. As results, we obtained a more stable control response because large overshoots and vibrations were greatly reduced in this process as compare to the general neural adaptive controller at the initial control state. When the control response of the fuzzy neural controller reached at setpoints in stable state, maximum deviation was about ±pH 0.02. Moreover, the pH of process converged at various setpoints from pH 6 to pH 10 and it showed a satisfying control performance in a vast controlling extent. The experiments of pH control have shown that the proposed fuzzy neural adaptive control could improve effectively unstable control response of neural adaptive control.
Keywords:neural networks;fuzzy logic;pH control;neutralization process
  1. Buchholt F, Kmmel M, Automatica, 15(6), 665 (1979) 
  2. Piovoso MJ, Williams JM, Self-Tuning pH Control: A Difficult Problem, An Effective Solution, Intech, May, 45 (1985)
  3. Chan HC, Yu CC, Ind. Eng. Chem. Res., 34(5), 1718 (1995) 
  4. Jayadeva B, Rao SNM, Chidambaram M, Madhavan KP, Comput. Chem. Eng., 14(8), 917 (1990) 
  5. Gustafsson TK, Waller KV, Ind. Eng. Chem. Res., 31, 2681 (1992) 
  6. Pottmann M, Seborg DE, Comput. Chem. Eng., 21(9), 965 (1997) 
  7. Zhao H, Guiver J, Neelakantan R, Biegler LT, Control Eng. Practice, 9(2), 125 (2001) 
  8. Krishnapura VG, Jutan A, Chem. Eng. Sci., 55(18), 3803 (2000) 
  9. Loh AP, Looi KO, Fong KF, J. Process Control, 5(6), 355 (1995) 
  10. Palancar MC, Aragon JM, Torrecilla JS, Ind. Eng. Chem. Res., 37(7), 2729 (1998) 
  11. Syu MJ, Chang CB, Bioprocess Eng., 17(4), 203 (1997) 
  12. Syu MJ, Chen BC, Ind. Eng. Chem. Res., 37(9), 3625 (1998) 
  13. Mamdani EH, IEEE Trans. On Computer, C26(12), 1182 (1997)
  14. Aoyama A, Doyle FJ, Venkatasubramanian V, Eng. appl. Artif. Intell., 8(5), 483 (1995) 
  15. Gupta MM, Fuzzy Logic and Neural Networks, Proc. of the Tenth Int. Conf. Multiple Criteria Decision Making, Japan, 3, 281 (1992)
  16. Tanaka K, Sano M, Fuzzy Sets Systems, 71, 71 (1995) 
  17. tan Y, Cauwenberghe AV, Eng. Appl. Artif. Intell., 12(1), 21 (1999) 
  18. McAvoy TJ, Hsu E, Lowenthal S, Ind. Eng. Chem. Process Des. Dev., 2, 68 (1972) 
  19. Kawato M, Uno Y, Isobe M, Suzuki R, IEEE Trans. System Magaz., 8(2), 8 (1988) 
  20. Lennart Ljung, System Identification Theory for the User, Prentice-Hall, Inc, New Jersey, U.S.A., 303 (1987)
  21. Johansson R, System Modeling and Identification, Prentice-Hall, Inc,. New Jersey, U.S.A., 263 (1993)