The Research On Predicting Blended Coals' Properties And Blending Optimizing Models Based On Intellectual Technology

In recent years, most of fossil-fired power plants have to burn the blended coals due to many factors such as tenseness of coals burning, rising of coal price as well as complicated and variable of coal source etc. However, it is a universal phenomenon that coal-fired devaiates from the coals of boiler designing. In addition, unstable coal properties cause lower boiler efficiency, reduce the security and reliability of combustion equipment and increase pollutant emissions. Therefore, to study the characteristics of blended coals and ignition has great significance in making full use of the local inferior coals and achieving optimization of blended coals combustion in the boiler.In this thesis, various kinds of local coals in Hunan and coals from other provinces are used in the power coal blending study according to the poor quality of local coals. Experiments of net calorific value determination, elemental, industry, and thermogravimetric analysis are done for the single and blended coals selected, characteristics of single coals and ignition are mastered. Comparing linear weighted values to the experimental values of blended coal, we obverse that it is the nonlinear relationship between single and blended coals rather than simple linear weighted.Based on experimental data, the prediction models of quality characteristic, ignition temperature and the parameter of ignition and combustion stabilized character for blended coals are established by using GRNN network. Moreover, the forecast performance of models are investigated respectively, appropriate network parameters and model structures are obtained. In the end, quality characteristic, ignition temperature and the parameter of ignition and combustion stabilized character of blended coals are predicted, the results show that all the average relative errors are less than 5 percents. On the basis of the coals' database in this thesis, power coal blending optimization model is built and the genetic algorithms is adopted in the model. From the constraint conditions of the optimization model in combination with the prediction model of blended coals quality characteristic, we obtain the optimum coal blending plan with the lowest blended coals price in accordance with constraint conditions. It is verified that coal quality characteristic, ignition character and the parameter of ignition and combustion stabilized character as well as price of the optimum blended coals have better superiority than the single coals. The optimization purpose of the blended coals combustion can be achieved.