The Short-term Load Forecasting Of Power Systerm Based On SVM

Accurate short term load forecasting can help to develop the electricity sectorrecentlypower generation plan effectively and make the reasonable arrangements for the startand stop ofthe power plant generating units.It effectively reduces the cost of power generation,and keep the real-time supply-demand balance of the power grid energy, to improve powersystem stability with good social and economic benefits.Especially after the introduction ofmarket competition mechanism in the power industry, it got more and more the concern of theelectrical workers.The support vector machine is a novel machine learning methods.Thesupport vector machine is a novel machine learning methods,can convert the nonlinearproblem to the linear problem, with good performance in pattern recognition and functionregression estimation. But it will runfor a long time when you have massive input data toparticipate in the training model.This article first lists a variety of short-term load forecasting algorithm and explains theirrespective advantages and disadvantages. Then it introduces basic theory of support vectormachines step by step.The input of the prediction model is very important to the predictingresults. according to the social environment and the load characteristics of the Foshan area,this paper presents a method which identifies abnormal data in the historical load data withdeviation rate and then corrects them. According to the auto-correlation coefficient ofstochastic-time-sequence, this paper selects inputs of the SVM, so that the selected inputshave the greatest relevance with expect outputs.Using the prediction model proposing in this paper, we predicted the electric load ofFoShan region according to its historical load data. The predictions have high accuracy,indicating the feasibility and rationality of this article.