POWER SYSTEM SHORT-TERM LOAD FORECASTING BASED ON SUPPORT VECTOR MACHINES

Abstract

Using the radial base function (RBF) as kernel function, the research of applying the Support Vector Machines (SVM) method to power system short-term load forecasting is presented. At first, the expression of regression estimation function is established by SVM based regression estimation algorithm and the structure of SVM network is given. Adopting the actual data from the distribution network of a certain domestic city, the samples are chosen according to different attributes of daily power loads and historical load data, and then the load is forecasted by use of LIBSVM algorithm and proper kernel function. The forecasted results are compared with those from time series method and BP artificial neural network (ANN) method, and it is shown that the presented forecasting method is more accurate.