Study On The Application Of Internal Model Control Based On RBF Neural Network In Thermal Power Plant Desulfurization System

In the operation of the thermal power plant,the process will emit a large number of SO2,which will also cause serious environmental pollution,so the reasonable control of SO2 emissions has become an urgent problem to be solved in our country.The PH control process of slurry is a typical nonlinear and large time delay system in the absorption tower of the thermal power plant desulfurization process.The control process has significant characteristics of multi-variable,nonlinear,variable gain and so on.In this paper,an improved RBF neural network is proposed to realize the control of PH value,based on the analysis of the characteristics of PH control.Firstly,the PH control process of slurry is modeled in the absorption tower.The characteristics of PH and the course of neutralization reaction are analyzed through the intensive study on the PH control process of slurry in absorption tower during desulfurization process.The mathematical model,which is used to control the PH of slurry is identified by least squares identification based on Hammerstein model,and obtain the corresponding mathematical model.Secondly,it is difficult to determine the number of hidden layer units,the center vector and the extension parameters of RBF neural network.To solve this problem,subtractive clustering algorithm is used in this paper.Finally,the improved RBF neural network is applied to the internal model control,and the forward model and the reverse model are identified,thus a complete internal model control system is formed.In this way,the adaptive ability of the internal model control as well as the application range can be greatly improved and extended.Therefore,an advanced control method has been developed and applied to the PH control process in absorption tower.The results of MATLAB simulation show that the internal model control method,which is based on improved RBF neural network,has good control performance,tracking ability,adaptive ability and robustness to the PH value control of the desulfurization system.Compared with the traditional PID control,it has a significant advantage.