Study of a multivariable coordinate control for a supercritical power plant process

The paper presents the recent research work in study of a novel multivariable coordinate control for a 600MW supercritical (SC) power plant. The mathematical model of the plant is described in the first part of the paper. Then, a control strategy is designed based on Model Predictive Control (MPC) theory. It is noticed that the linear MPC alone performs well only within limited small load changes under a constant level of disturbances and measurement noises generalized from the prediction algorithms. So, a dynamic compensator is proposed to work in parallel with the MPC to track large load changes. Because the model has been identified with on-site closed loop response data, the multivariable optimal control signals have been used as a correction to the reference of the plant local controls instead of direct control signal applications. The simulation results show the good performance of the controller in response to the large load changes. Furthermore, it has been proved that the plant dynamic response can be improved by increasing the coal grinding capability and pulverized coal discharging through implementation of suitable coal mill controllers.

[1]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[2]  Shyan-Shu Shieh,et al.  Developing a robust model predictive control architecture through regional knowledge analysis of artificial neural networks , 2003 .

[3]  James A. Momoh,et al.  Electric Power System Dynamics and Stability , 1999 .

[4]  Q.H. Wu,et al.  Development of a Multisegment Coal Mill Model Using an Evolutionary Computation Technique , 2007, IEEE Transactions on Energy Conversion.

[5]  Sung-Ho Kim,et al.  Neural Network-Based Modeling for A Large-Scale Power Plant , 2007, 2007 IEEE Power Engineering Society General Meeting.

[6]  J. A. Rovnak,et al.  Dynamic matrix based control of fossil power plants , 1991 .

[7]  Hideo Nakamura,et al.  Statistical identification for optimal control of supercritical thermal power plants , 1981, Autom..

[8]  B. P. Gibbs,et al.  Application of nonlinear model-based predictive control to fossil power plants , 1991, [1991] Proceedings of the 30th IEEE Conference on Decision and Control.

[9]  Kwang Y. Lee,et al.  Boiler-turbine control system design using a genetic algorithm , 1995 .

[10]  Kumar Rayaprolu,et al.  Boilers for Power and Process , 2009 .

[11]  Alberto Herreros,et al.  Design of PID-type controllers using multiobjective genetic algorithms. , 2002, ISA transactions.

[12]  Wei-Der Chang,et al.  A multi-crossover genetic approach to multivariable PID controllers tuning , 2007, Expert Syst. Appl..

[13]  G. Oluwande,et al.  Coal Mill Modeling by Machine Learning Based on on-Site Measurements , 2002 .

[14]  Sung-Ho Kim,et al.  Controller Design for a Large-Scale Ultrasupercritical Once-Through Boiler Power Plant , 2010, IEEE Transactions on Energy Conversion.

[15]  Weihua Gui,et al.  Robust MPC Based on Multivariable RBF-ARX Model for Nonlinear Systems , 2005, Proceedings of the 44th IEEE Conference on Decision and Control.

[16]  Sung-Ho Kim,et al.  Modified Predictive Optimal Control Using Neural Network-based Combined Model for Large-Scale Power Plants , 2007, 2007 IEEE Power Engineering Society General Meeting.

[17]  T. Inoue,et al.  A model of fossil fueled plant with once-through boiler for power system frequency simulation studies , 2000, 2000 Power Engineering Society Summer Meeting (Cat. No.00CH37134).

[18]  G. Oluwande,et al.  Coal Mill Modeling by Machine Learning Based on on-Site Measurements , 2002, IEEE Power Engineering Review.