Optimal Control Of Coal-fired Thermal Power Plants Based On Support Vector Machine

This paper took a certain thermal power plant of Xinjiang as a background, aimed atstudying the problems in the yard of Xinjiang thermal power industry such as impropermanagement of coal yards、reducing the boiler accidents、optimization of boiler combustionefficiency and so on, developed a combustion optimization technology based on the SupportVector Machine by using Support Vector Regression as a modeling tool. This technology apply tothe coal’s quantity of heat forecast,Research and studied the effect of the changes of thepredicting ability of penalty factor c of Support Vector Regression model and parameter g of radialbasis kernel function in such situation that kernel function chose radio basis kernel function andsolve the practical problems by using the scientific ways such as mathematical modeling andoptimal control.The status of the heating surface of the coal burning boiler and the slagging condition of theboiler have the extremely important influence on the safety as well as the economical efficiency ofthe power plant. Thus the Support Vector Regression model was set based on the coal ash and theboiler operational data. Combination the optimal and The Support Vector Regression model madethe slagging predicting model possess good predicting ability and generalization ability and itpredicted comparatively accurately on the slagging tendency of the coal. The result showed that ithas its certain science that we can improve the predicting accuracy of the model throughoptimization and then improving the combustion efficiency and reducing the boiler accidents.As to the improper management of coal yards, considering the influencing factors of climate:the natural evaporation that exists in the open-air coal yards、loss brought by the natural disasterssuch as rain and snow as well as the wind damage, this paper analyzed the loss of coal storage byusing the Support Vector Regression. It also proposed new methods of managing coal yards byuse the Grey-correlation to calculating the correlation coefficient of coal loss which is due to theweather factors and combining sufficiently with the actual needs of management process and thepower coal production.