基于RBF神经网络的SCR脱硝系统喷氨优化

Ammonia Spraying Optimization of an SCR Denitrification System Based on RBF Neural Network

为实现对电厂选择性催化还原(SCR)脱硝装置喷氨的优化控制,以广东某电厂350 MW锅炉为研究对象,采用径向基函数(RBF)神经网络法,以锅炉负荷、烟气体积流量、SCR烟气温度、脱硝进口NO_x 质量浓度以及喷氨质量流量等为输入变量,以SCR脱硝效率为输出变量,建立输入变量与输出变量之间的关系模型,实现对SCR脱硝效率及脱硝出口NO_x 质量浓度的预测.在满足NO_x 排放标准的前提下,以SCR系统运行成本最小为目标,利用Matlab对该模型进行仿真实验,寻求氨耗成本和电耗成本与NO_x 排放费用的临界点,得到最佳喷氨质量流量.结果表明:最佳喷氨质量流量计算值比实测值或高或低,但在满足NO_x 排放标准的前提下,其SCR系统运行成本呈降低趋势.

To optimize the control on ammonia spraying of the selective catalytic reduction（SCR）denitrification device in a 350 MW power boiler in Guangzhou,a relationship model was established between the input and output variables based on radial basis function（RBF）neural network by taking the boiler load,flue gas flow,SCR inlet flue gas temperature,SCR inlet NOx concentration and the spraying ammonia flow as the input variables,and the SCR denitrification efficiency as the output variable,so as to realize the prediction of SCR denitrification efficiency and outlet NOx concentration.Under the premise of satisfying the requirements of NOx emission and aiming at minimizing the operating cost of the SCR system,Matlab was used to perform a simulation experiment on the model to seek a critical point among the ammonia consumption cost,power consumption cost and the NOx emission fee,thus obtaining an optimal flow of spraying ammonia.Results show that the calculated mass flow of ammonia spraying is either higher or lower than the measurements,but the operating cost of the SCR system always keeps decreasing under the premise of meeting the NOx emission standard.