Optimization Of SCR Denitration Control System Based On Stepwise Generalized Predictive

With the improvement of environmental protection awareness in today’s society,the control of NOx emission in the process of power generation of thermal power units has attracted people’s attention.Nowadays,SCR flue gas denitration technology is a more mature method to control NOx emission.But in the real control operation of the SCR flue gas denitration technology in process with time delay,great inertia,often caused by ammonia spraying a quantity to exceed bid time and other issues will seriously affect the denitration efficiency,not in time will directly lead to the amount of ammonia injection ultimately denitration efficiency is low,ammonia injection quantity and can lead to excessive air preheater plugging,therefore,Reasonable and effective optimization of SCR flue gas denitration control system can relatively improve the denitration efficiency,achieve accurate control of export NOx emissions,and better control of NOx emissions.Therefore,the SCR denitration control system was optimized in this paper.The main research contents are as follows:(1)In the first part,the SCR flue gas denitration control system is introduced and the mathematical model of the controlled object is established.Firstly,the SCR flue gas denitrification control system was introduced from two aspects of process flow and reaction mechanism.Secondly,particle swarm optimization(PSO)is used to identify the parameters of the mathematical model of the controlled object of the system controller.The identified model shows a high computational accuracy and has certain practical and theoretical application value.(2)The second part mainly studies the establishment of NOx emission prediction model.Firstly,KPCA was used to extract the auxiliary variables of NOx emissions,which reduced the degree of redundancy among variables,and thus reduced the complexity of the model.Secondly,the relevant principles in the process of establishing the prediction model are introduced.Due to the delay and inertia of the system,the input variables of the NOx prediction model are delayed selected by K-nearest neighbor mutual information,so as to improve the asynchronous problem between the historical data of auxiliary variables.Finally,the auxiliary variable with delay information is used as the input of NOx emission model,and the NOx emission prediction model based on NARX neural network is established.The prediction model has high accuracy.(3)The third part,in order to improve the control accuracy of SCR flue gas denitration control system existing in the traditional PID control problems,an optimization control strategy of SCR denitration control system based on stepped generalized predictive control is proposed.Firstly,the parameters of the stepped generalized predictive controller(SGPC)are set,and the performance characteristics of the stepped generalized predictive controller are analyzed.The performance of SGPC is better than that of PID,and it is more suitable to be used in complex systems.Secondly,using the recursive least squares algorithm of prediction model for online identification of the object,and gives the optimal control structure of the system,the PID controller with prediction model by stepwise generalized predictive controller to replace,it compared with the traditional PID control simulation,the improved PID controller of the poor control precision.Finally,based on feedforward control of ammonia injection volume prediction,the amount of ammonia injection based on NARX neural network predictive value into the optimization control structure,constitute feedforward + feedback denitration optimization control strategy,and the simulation results compared with the traditional PID control,optimal control strategy based on feedforward + feedback can not only reduce exports NOx concentration fluctuations,And it can reflect the change of the modulated quantity in time,so as to overcome the influence of large delay link.