YANG Jianguo, YE Lingyun, ZHAO Min, ZHAO Hong, LI Yijin, LI Min, YU Yu, DENG Furong
2017, 37(11): 870-875.
Taking into account the category and quality of the coal fired, the combination way of coal mills, the air distribution mode and the operating parameters of a 660 MW boiler, a NOx prediction model was established using support vector machine (SVM), which was used to carry out application simulation and to research the effects of following factors on the NOx concentration, such as the oxygen concentration, OFA rate, air distribution mode, coal category and the combination way of coal mills, etc. Results show that the model has good accuracy and high generalization ability as well as ideal regulation performance, which comprises information on coal category and combination way of coal mills, and therefore is of great significance for multi-coal units in improving the NOx prediction accuracy, reducing the NOx concentration, carrying out the adjustment and optimization for stable combustion, especially in guiding the coal procurement and blending, and even for purchasing of "intelligent fuel".