Multivariable Disturbance Rejection Predictive Control Strategy With Application To Power Plant Flue Gas Desulfurization System

Large-scale coal-fired power plants basically use wet limestone flue gas desulfurization technology,but the desulfurization process is complicated.The desulfurization system is a typical multi-variable and large-delay controlled object.It is difficult to obtain satisfactory control results using conventional PID control.Therefore,by establishing a dynamic mathematical model of the desulfurization system,mastering the dynamic characteristics of the desulfurization system and the coupling relationship between the loops,and studying the advanced control strategy of the desulfurization system accordingly,it is of great significance to effectively improve the regulation quality of the desulfurization system,improve the desulfurization efficiency,optimize the operation mode and reduce the working intensity of the operating personnel.Therefore,the multivariable disturbance predictive control method of the desulfurization system is studied,and simulation verification and application development are conducted.The main research contents include:1.The existing transfer function form of the desulfurization system model is used to generate closed-loop data under the predictive controller based on the augmented state space model.Based on this,two closed-loop subspace identification methods are studied and compared by simulation.The mathematical model of the desulfurization system in the form of state space is established and the accuracy of the model is verified by simulation,and then the dynamic characteristics of the system is analyzed.This lays a model foundation for the design of multivariable disturbance rejection predictive controller.2.A multivariable disturbance rejection predictive control for desulfurization system is proposed.The disturbance sub-model set is established on the basis of the augmented state space model,the state is estimated by Kalman filtering,and the Bayesian probability weighting method is used to obtain the average prediction model.Not only can the control of the p H value of the circulating slurry and the SO₂ concentration after desulfurization be realized,but also the influence of undetectable disturbances on the desulfurization control system can be rejected quickly and effectively.The simulation is carried out in the presence of step output,step input,ramp input,period input and multiple disturbances,respectively.The effectiveness of the proposed control method is verified by simulation experiments.3.Desulfurization advanced control software is developed,DCS control logic design is conducted and a solution to the actual problems that may occur during debugging the process is provided,which lays the foundation for the final successful implementation of engineering applications.