Research On Optimization Methods Of Load Dispatch Under The Condition Of Fickle Coal In Thermal Power Plant

With the continuous development of data acquisition and storage technology, the size and quality of production data of thermal power plant improved, the data contains a large number of information required by operation optimization. The paper discused the applied actuality of data mining in the electric power industry, its application status in the electric power industry has wide development space, and statistical method of data mining study is one of the important ways. The paper studied the correlativity of the variables using the regression pattern of the data mining, and introduced the partial least squares regression method according to the multiple related features of production data of thermal power plant, the idea of abstracting components of the partial least squares regression method is good at solving the problems of multiple correlation between variables effectively. The paper identified several factors on the unit power supply coal consumption selectively, accordting to operation parameters'values of two units of a certain power plant, and analysed the relationship between the power supply coal consumption and these variables, then established a mathematical model with the partial least squares regression method, and made a evaluation to results using patial least squares regression analysis, by contrast with the traditional least squares regression model, the findings show that the partial least squares regression method in plant unit state parameter prediction has better feasibility and reliability.This paper take the coal consume values predicted by the partial least squares sregression model and the corresponding load of the unit as discrete data points, and fitted the quadratic polynomial of the unit's coal consume character by least squares method and established the objective function of the load dispatch through making the quadratic polynomial as the basis of the the load dispatch.The genetic algorithm has no special requirements on the curve of the coal consume character because of its strongly global search performance and adaptability. Take the genetic algorithm as optimization algorithm,using MATLAB producing procedure to solve the model, the optimal solution was got and the solution was analyzed in detail. Considering the coal consume value was the fitted value of partial least squares regression during the process of load dispatch and the computed results were accord with the ture value, the results further shows that the thesis has certain theoretical and practical value and it is propitious to guide the operation optimization of the unit.