The Transactions of the Korean Institute of Electrical Engineers D (대한전기학회논문지:시스템및제어부문D)

The Korean Institute of Electrical Engineers (대한전기학회)
권호 표지

Nonlinear PID Controller with Neural Network based Compensator

신경회로망 보상기를 갖는 비선형 PID 제어기

  • Lee, Chang-Gu (Dept.of Electronics Information Engineering, Engineering College, Chonbuk National University)
  • 이창구 (전북대 공대 전자정보공학부)
  • Published : 2000.05.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we present an nonlinear PID controller with network based compensator which consists of a conventional PID controller that controls the linear components and neuro-compensator that controls the output errors and nonlinear components. This controller is based on the Harris's concept where he explained that the adaptive controller consists of the PID control term and the disturbance compensating term. The resulting controller's architecture is also found to be very similar to that of Wang's controller. This controller adds a self-tuning ability to the existing PID controller without replacing it by compensating the output errors through the neuro-compensator. Various simulations and comparative studies have proven that the proposed nonlinear PID controller produces superior results to other existing PID controllers. When applied to an actual magnetic levitation system which is known to be very nonlinear, it has also produced an excellent results.

Keywords

References

  1. Watanabe, K.,Adaptive estimation and control, partitioning approach, Prentice Hall, 1992
  2. Ahmed, M.S. and Tasadduq, I.A.,'Neural-net control for nonlinear plants: design approach through linearization,'IEE Proc. Control Theory Application, Vol. 141, No. 5, pp. 315-322, 1994
  3. Chan, K.C. and Leong, S.S.,'A neural network PI controller tuner,'Artificial intelligence in engineering, Vol. 9, pp. 167-176, 1995 https://doi.org/10.1016/0954-1810(94)00014-V
  4. Jota, F.G.,'Practical automatic tuning methods of PID controllers for a sour water stripper, IEEE international symposium on intelligent control, Columbus, Ohio, USA, pp. 22-26, 1994
  5. Tan, Y and Keyser R, 'Adaptive PID control with neural network based predictor,' CONTROL'94, March 1994 https://doi.org/10.1049/cp:19940357
  6. Wang Fuli and et all, 'Neural network pole placement controller for nonlinear systems through linearization,' Proc. of the ACC, pp. 1984-1988, 1997 https://doi.org/10.1109/ACC.1997.611035
  7. Wang Fuli and et. all, 'A PID-like controller for nonlinear systems,' Proc. of the ACC, pp. 1558-1562, 1997 https://doi.org/10.1109/ACC.1997.610826
  8. Harris, T.J.and MacGREGOR, J.F., 'An overview of discrete stochastic controllers: generalized PID algorithms with dead-time compensation,' The canadian J. of Chemical Eng., Vol. 60, pp. 425-432, 1982
  9. Ishida, Y. and et. all, 'Nonlinear PID controller using neural networks,' IEEE International Conf. on Neural Networks, pp. 811-814, 1997 https://doi.org/10.1109/ICNN.1997.616127
  10. Astrom, K.J.and Hagglund, T., Automatic tuning of PID controllers, ISA, NC, 1995
  11. 이창구, 신동용, '오차 자기순환 신경회로망에 기초한 적응 PID 제어기,' 제어자동화시스템 논문지, Vol. 4, No. 2, pp.209-214, 1998
  12. Scalero, R.S. and Tepedelenliouglu, N., 'A fast new algorithm for training feedforwrd neural networks,' IEEE Trans. Signal Processing, Vol. 40, pp. 202-210, 1992 https://doi.org/10.1109/78.157194