Study On Emissions Characteristics Of Coal-blending Combustion

Coal plays a very important role in application of primary energy. Nowadays in China, with the cost of coal become higher, the cost of thermal power generation increase quickly. The economic impact has become a big problem. However, the cost of environmental pollution is not easy to be ignored. Thermal power plants in China burn off billions of tons carbon every year, producing ten millions tons of pollutions, including SO2and NOx, some of this pollution can be reduced before emission to atmosphere.In this paper, it study SO2and NOx emission characteristic curves by burning lignite, bituminous coal, lean coal, anthracite coal and their blended coals. The text study the changes in the factors which affecting the pollutant emission characteristics by analysis the changes in emission characteristic curve through controlling the change of the coal, mixed ratio, combustion temperature, combustion oxygen.As the mixed coal combustion affected by various factors, emissions characteristic curve and the number of burned coal is non-linear relationship, it needs a mechanism to simulate non-linear prediction model to calculate pollutant emissions, This paper used the artificial neural network to achieve this purpose. In this paper, it select the appropriate neural network hidden layer, hidden layer nodes, transfer functions and training algorithms to inspect and process the raw data. It use MATLAB to establish the BP neural network model. It train the neural network model to form a complete model which can predict the emission of coal-blending combustion though the training sample. According to examine the results of the test sample, the results show that the error between predicted result and test samples achieve the desired level.