Prediction Of NO_x Emissions From Power Station Boilers Based On Multi-model Fusion

The emission of nitrogen oxides(NO_x)generated during the fuel combustion process of coal-fired units has caused great harm to the human body and the environment.In order to promote the healthy and green growth of my country’s economy and meet the demand for electricity supply,the establishment of an accurate and effective NO_x emission model can provide favorable support for reducing the NO_x emission of coal-fired unit boilers.The particle swarm optimization-based support vector machine model(PSO-SVM)is used as the weak learner algorithm of the integrated model,and the improved Adaboost integration method is used to estimate the boiler NO_x emissions,which has achieved better accuracy and real-time prediction.Estimate the effect.The research content of the thesis is divided into the following aspects:(1)First,the importance of the dynamic estimation of NO_x emissions to the SCR denitrification system and the significance of the optimization of boiler combustion conditions are explained;then,the current research status of the NO_x emissions estimation of boilers at home and abroad are reviewed.(2)Analyze the generation mechanism of boiler NO_x emissions,and combine the operating data to select the variables of NO_x emissions to obtain its main influencing factors;through the analysis of the generation mechanism of boiler NO_x,it can be seen that each variable in the original variable is relative to the boiler outlet NO_x Concentrations have different degrees of lag,and the lag time corresponding to each variable is analyzed by the method of mutual information.(3)The support vector machine modeling method of boiler NO_x emissions based on particle swarm optimization(PSO-SVM)is studied.First,introduce the basic principles of the SVM model and the PSO intelligent optimization algorithm,and optimize the relevant parameters of the support vector machine model through the PSO algorithm;then,add the influencing factors of the delay time as the input of the model,and use the support of particle swarm optimization optimization The vector machine established a model of NO_x emissions from power station boilers.(4)The multi-model integrated modeling method of boiler NO_x emissions based on the improved Adaboost algorithm is studied.First,the model ensemble method based on the Adaboost algorithm and its basic algorithm are explained;an Adaboost ensemble algorithm that optimizes the loss function is proposed,and the regularization factor and prior knowledge parameters are introduced into the algorithm to improve the accuracy and generalization performance of the model.,Effectively reducing the complexity of the model;finally,taking a 350 MW coal-fired unit in a power plant as the research object,based on the actual working conditions and operating data of the unit,using PSO-SVM as the weak learner model and improving the Adaboost integration algorithm Integrated model of boiler NO_x emissions.Compared with the above single model,the boiler NO_x emission modeling method based on the improved Adaboost algorithm effectively improves the accuracy and stability of the model,and its generalization ability is improved.