Research Of Boiler Combustion Model Based On The On-line Detection Of Coal Quality

In recent years, with the development of China power industry, the demand of coal as well as the pollutant of the power state continue to increase. How to realize both improving the efficiency and reducing harmful Emission becomes one of the most important issues in China. Boiler combustion involves complex physical and chemical reaction process, and it has many factors which influence this process. Thus, it is difficult to reach the goal above.This article analysises modeling methods which are commonly used recently to find out that support vector machine method is the most efficient method. SVM treats the boiler combustion process as a black-box operation process which based on a large number of historical data of training so that we finally make the model that can simulate the boiler combustion process.This paper chooses the RBF as the kernel and uses on-line detection technology to obtain the accurate parameters of the calorific value of coal. Further more, it uses the calorific value of coal parameters instead of coal quality data from traditional laboratory to set up the model.The first three chapters of the thesis mainly about the basic research and method, the fourth chapter establish accurate boiler combustion process model. The model uses the MATLAB program language to set the efficiency of the boiler on-line model and the emission of the NOx on-line model.