Short Term Load Forecasting Of Power System Based On Clustering Analysis And Support Vector Machine

Power system short-term load forecasting is not only one of the most important work of power grid enterprises,but also the basic work of power network planning and power system stability operation.As an important basis for the distribution of electricity,with the deepening of the reform of the power market,the accuracy of power system load forecasting is getting higher and higher;and because of the new load and the large access of distributed power,It is urgent to develop a kind of short-term load forecasting technology which can adapt to the development of the power system.In this paper,the load forecasting method based on the traditional support vector machine is improved.Firstly,the original load data are cleaned and the original data are matched by two time series models.Then,the abnormal values are detected,And then use fuzzy C-means clustering algorithm to cluster analysis,this algorithm is one kind of data set is considered as a sample data set in a similar fuzzy subset.The criterion of classifying a class is fuzzy membership degree.It turns the problem of clustering into a nonlinear programming problem with constraints,and then obtain the membership matrix U;then through judging the similarity degree formula between the data sets,The Euclidean distance is used to test the similarity degree,and then the historical data set with the highest degree of similarity is selected.Finally,the short-term load forecasting model is combined with the regression model of support vector machine.The simulation results of the actual example show that the prediction results are within the allowable error range,and the prediction results are more accurate than those based on the support vector machine.