The Study Of Fuzzy Network Generalized Predictive Control In The Unit Coordinated Control

In china, coal-fired power generation is mainly, in spite of unconventional energy of shale gas, combustible ice, but coal is still the most cheap and safe energy. Thermal power unit is the basic unit of power grid, The requirements of basic load, peak shaving and frequency modulation in power grid is done through the thermal power unit, and it affects the stable operation and economic benefits of the power grid. Under the increasingly severe environmental pressure, how to ensure the thermal power plant security economy and high efficiency and stable operation to become the priority.The engine and boiler coordinated control system is a double input two output, strong coupling, nonlinear, and the global is a strongly nonlinear control system. Within the range of conditions, using the conventional control method is very difficult to get a satisfactory control quality. This paper is on the basis of the study of the generalized predictive control and fuzzy neural network, aiming at the characteristics of the generalized predictive control on-line calculation is big, not suitable for the nonlinear object, putting forward fuzzy neural network generalized predictive control base on the fuzzy neural network model. By using the fuzzy neural network system identification method, it can improve the precision of conventional fuzzy system, and it can solve the slow convergence speed and local minimum value problem of the BP neural network learning algorithm. This algorithm greatly, reduces the online computation of generalized predictive control algorithm and improve the real-time property.In the paper, the simulation of coordinated control system for 300 MW units mathematical model was based on the analysis of the algorithm, the simulation results show that the control system can ensure the output power and the main steam pressure tracking set value fast and smooth, when the controlled variables under varying load conditions, it can still maintain good control performance, compared with the traditional PID control, this algorithm shows good adaptability and robustness, and it’s feasible.