Combustion Modeling And Optimization Of Supercritical Units

The thermal power generation will remain the most inpportant form of power generation for a long time in China,which will make energy consumption and environmental pollution more serious while promoting the industrial development.With the increasing of unit capacity,the combustion process becomes more complicated.In this paper,the modeling and optimization of boiler combustion are discussed for 660 MW ultra-supercritical coal-fired units.The combustion system is modeled hierarchically according to the temperature and load of each combustion layer,the primary and secondary air volumes,coal feeding and so on.The model parameters are optimized by using particle swarm algorithm.Based on the combustion model and ant colony algorithm,each combustion sub-model is optimized,and the NOx emission and exhaust gas temperature are reduced.The main research contents are as follows:Firstly,the main factors affecting the boiler thermal efficiency and the generating mechanism of NOx in coal-fired units are analyzed.Based on the temperature of various combustion layers and other related operating parameters,the correlation between parameters and NOx emission and exhaust gas temperature are analyzed.According to the boiler combustion history data of 660MW ultra-supercritical units,the sample space is divided into three parts:the low,medium and high load based on the principle of load change interval trisection.The features of the various load input sample subspace are extracted by using the principal component analysis(PCA).The most important influencing factors of the NOx emission and exhaust gas temperature are selected in each load section,which laying the foundation for the modeling of combustion system.Secondly,whole model of NOx emissions and exhaust gas temperature of three layers of the sample space are built respectively by using the least squares support vector machine(LSSVM)method.The regularization coefficient and the parameters of the kernel function of each layer model are optimized based on the improved particle swarm algorithm.The simulation results show that the combustion model has the characteristics of high prediction accuracy and generalization ability of NOx emission and exhaust gas temperature.Finally,the bi-objectives optimization of NOx emission and exhaust temperature is realized based on LSSVM combustion motel.For boiler combustion system LSSVM models of three layers,the adjustable operation parameters of the boiler are regarded as optimization variables,including the coal feeding for each combustion layers,the secondary air of layers,burning wind,etc.Taking aim at reducing the NOx emissions and exhaust gas temperature of boiler,the global optimization is finished by utilizing ant colony algorithm of multi-objective optimization.The optimization results of each sub-model show that the NOx emissions and exhaust gas temperature have been decreased.