On-line Prediction Of NO_x Emission From Coal-fired Boiler Based On Dominant Factor Analysis

Nitrogen oxides(NO_x)emissions from coal-fired power stations are one of the important pollutants of the environment.In order to meet environmental protection requirements,selective catalytic reduction(SCR)denitration control is an important method for reducing NO_x emissions in most coal-fired power plants.The accurate and timely acquisition of NO_x emissions from coal-fired boilers plays an important role in improving the quality of SCR control systems.The accuracy of the input variable selection of the model is one of the main factors affecting the performance of the model.Due to the many and complex factors affecting the NO_x emissions of the boiler,based on the existing mutual information theory,a variable information based distribution is proposed for the problem of variable selection Based on the characteristic variable selection algorithm,the algorithm is applied to the dominant factor selection of NO_x emission model of coal-fired boilers.Based on the characteristic variables selected by the algorithm as input,a dynamic model of boiler NO_x emissions is established.Finally,based on the established boiler NO_x emission model,an SCR denitration system with a fixed ammonia-nitrogen molar ratio control method was designed,and the actual data was used to verify the effect.The main research contents of this article are as follows:(1)The significance and background of establishing a boiler NO_x emission model are discussed,and the status of research on variable selection methods and modeling methods at home and abroad is analyzed,and the generation mechanism and influencing factors of NO_x are introduced.(2)Aiming at the characteristics of strong coupling,large inertia,and non-linearity of the influencing factors of NO_x emissions from coal-fired boilers,the data was pre-processed,and a distributed characteristic variable selection algorithm was proposed based on mutual information knowledge,and verified with the nominal data set.Effectiveness of this method.(3)Based on the real data of a power plant,use the proposed distributed characteristic variable selection algorithm to select the input variables of the boiler NO_x emission model,and analyze the selected variables in combination with the mechanism.(4)The variable selected by the distributed characteristic variable algorithm is used as an input to build a boiler NO_x emission model,and the results show that the model has higher quasi-determination and strong generalization ability.(5)The application of NO_x dynamic estimation in denitration control was studied.The SCR denitration process is introduced.The SCR denitration control system is designed based on the NO_x emission model of a coal-fired boiler with a fixed ammonia-nitrogen molar ratio control method.Combined with actual data tests,it shows that the control system has achieved good quality results.