Modeling And Optimal Control Of SCR Denitration System In Thermal Power Plant

With the increasing total installed capacity of thermal power units, the pollution to the environment from the nitrogen oxides in flue gas is more and more serious, so the denitration has become an indispensable process in thermal power plant. Among many types of flue gas denitration technology, SCR is the maturest and most widely applied. In recent years, the power plant simulation system has become one of the principal means of employee training, but most of them don’t have complete model algorithms of denitration system. Therefore, the modeling and optimal control of SCR for thermal power plant is of great practical significance.In this paper, we used mechanism and test method in the modeling of SCR, and then we verified the two models respectively. In the process of mechanism modeling and verification, we firstly established the mechanism models of SCR denitration reactor and the other five equipments or processes; then we wrote algorithms according to the mechanism models and imported the algorithm into the algorithms library of STS simulation system support platform. With the original model algorithms, we developed a complete set of mechanism model of SCR denitration system. Then we constituted a set of simulation system of denitration system with the control logic and human-computer interface which are same to the real power plant. Finally we verified the accuracy of the model on the simulation system. During the process of test modeling and verification, we established BP neural network model of denitration reactor based on the field data and verified the accuracy of this model based on the test data, through data selection, data preprocessing, the determination of input and output, the determination of model structure, model training and model test. It shows that the two models are accurate, which can correctly reflect the static and dynamic characteristics of denitration system.The part of optimization control for denitration system mainly completed the PID parameters optimization of the master and deputy controller of ammonia adjustable valve based on the simulation system which was composed by mechanism model. First, we analyzed the control logic and observed control effect due to initial parameters. Second, according to the engineering practice, we used the step response method to calculate the transfer function for the corresponding link. And then we got the most optimal value of PID parameters by the normal particle swarm optimization. Finally, we replaced the initial parameters of the master and deputy controller of ammonia adjustable valve of the simulation system with the most optimal parameters, and then we found that the control effect was greatly improved,which is full of reference value to the optimization control of the actual denitration system.