Load Forecasting of Coal-Fired Unit Based on SVM Model

Abstract

In order to obtain accurate load forecasting of coal-fired unit, a new algorithm based on Support Vector Machine (SVM) method is presented. This algorithm establishes a model to reflect the complicated relation between the load of coal-fired unit and the furnace flame Images. The trained SVM model is applied to a 660MW coal-fired unit to forecast the load with two groups of test samples. The results are compared with that of BP neural network model. It is shown the SVM model is more accurate than the BP NN model. The SVM method can satisfy the demand of engineering applications with the advantages of high forecasting accuracy and more generalized performance.

The command allocation of the combined fire and storage FM system considering the benefit optimization

“…The two-point coal consumption fitting model fits the coal consumption data of many types of thermal power units to obtain a more accurate steady-state coal consumption model [4], and its mathematical model is as follows…”