The Optimal Control Of Ammonia Injection Based On The SCR Model Of A Coal-Fired Power Plant

With the large-scale operation of the new energy,thermal power units need to further improve the ability of quick response,so as to reduce the effect of new energy power on the grid.However,the rapid rise and fall of the load will cause a significant change in the concentration of NOx in flue gas.Although traditional SCR system control strategy can work,its effiency and accuracy can not meet the actual requirements of power plants.As the ultra low emission standards was issued,in order to ensure the flue gas of coal-fired power plants reach the standard and improve the economical effiency of SCR system,it is necessary to propose the precise control of annomia injection based on the accurate model of SCR system.Making full use of massive history data of DCS system and combining the mechanism analysis of denitration process,the model of SCR system was establishded and the optimal control strategy of ammonia injection was presented in this article.As a result,the operating cost of SCR is reduced while the discharge standard is reached,the denitration effiency is guaranteed and the ammonia injection is precisely controled.Therefore,the adverse effect of excessive injection of ammonia on the safe and stable operation of SCR downstream equipments is avoided.The main research contents are as follow:1.On the basis of analyzing the reaction mechanism of SCR denitration,the SCR mechanism model was built and the model parameters were identified based on actual operation data and optimization algorithm.In order to verify the accuracy and unniveraslity of the model,the operation data of 600 MW and 1000 MW power plant was used to make simulation and verification separately.Moreover,on the basis of mechanism model,the influence factors of denitration efficiency were analyzed.2.SCR reaction is complex,and there are certain error existence in the calculation results of the mechanism model,so it is necessary to use historical data of power plant DCS system to build data-driven SCR model.Furthermore,optimal input variables selection is an important premise to ensure the accuracy of data-driven model.Hence,on the basis of taking full advantage of the useful information in historical data,partial mutual information(PMI)method which is evolved from information entropy was employed to screen the input variables.In order to enhance the accuracy of variable selection,interpolation method was adopted to improve the calculation of the regression value in PMI method.IPMI method was applied to SCR system after it verified by 3 sets of benchmark.The results show that IPMI method improves the model accuracy effectively while reduces the model complexity.3.Based on the optimal input varable set selected in this article,kernel partial least squares(KPLS)method was used to build the data-driven model of SCR system.In view of the diffierent characteristics of input variables,genetic algorithm was employed to indentify the kernel width of each variable.For the problem of model failure caused by the change of working condition,the model updating strategy was designed.Based on that,self-adaptive multi-scale KPLS was proposed to establish SCR system model and the model validity was verified by actual data.4.By comparing the characteristics of the mechanism model and data-driven model,hybrid modeling method was proposed which used data-driven model to correct the deviation of mechanism model and the hybrid model updating strategy was designed.SCR model was established by hybrid modeling method and verification test was made.The calculation results show that hybrid model combines the advantages of mechanism and data-driven model and has good fitting and generalization ability5.Based on the SCR system model presented in this article,the ammonia injection was precisely controled by model prediction control(MPC)method which ensure the efficiency of denitration and avoid excessive injection of ammonia.The simulation was conducted to verify the control effect of MPC and the comparison bewteen traditional control method and MPC was analyzed.The simulation results show that comparing with traditional method,MPC has better control effect and achieves the fast and precise control of ammonia injection.