Research On The Predictive Model For SCR De-NO_X Efficiency Based On BP Neural Network

In recent years, with the intensity of electric power environmental protection, the flue gas denitrification technology in coal-fired power plant will be widely used. At present, the selective catalytic reduction (SCR) technology is the most mature process. By analyzing the network structure of BP neural network, the selection of training parameters and the learning rules, a new prediction model of SCR De-NOX has been established in this paper, based on the small-scale experimental research for denitrification by SCR, Meanwhile, Levenberg-Marguardt training rules were used for data processing, and the forecasting results of denitrification efficiency were obtained under the given conditions. Comparing the prediction efficiencies with the experimental values, it shows that the prediction model of SCR De-NOX has faster convergence rate, smaller prediction error and higher precision. It can work out a better prediction on the denitrification efficiency.