摘要
为提高母线负荷预测的准确性,提出一种基于小波分解和支持向量机的母线负荷预测方法。该方法利用小波分解算法将目标负荷序列分解为若干个不同频率的子序列,通过分析各个序列的特征规律,构造不同的支持向量机模型对各分量分别进行预测,再将各分量预测值进行重构得到最终预测值。对某一区域内15条母线进行预测,采用平均日母线负荷准确率进行评价。与单独使用支持向量机方法相比,应用所提方法提高了962点的预测效果,占总预测点数的66.8%;全系统的准确率由93.5%提高到了95.1%。
A bus load forecasting model based on wavelet transform and SVM(Support Vector Machine) is proposed to improve its accuracy,which applies the wavelet transform to decompose the target bus load sequence into the components of different frequencies,analyzes their characteristics to build SVM model for each component,forecasts separately the load of each component,and reconstructs them to obtain the final forecast. The loads of 15 buses in an area are studied and the average daily bus load accuracy is used as the index for comparison. Compared with the model only based on SVM,66.8 % of point forecasting accuracies,i.e. 962 points,are improved,and the overall forecasting accuracy increases from 93.5 % to 95.1%.
出处
《电力自动化设备》
EI
CSCD
北大核心
2012年第4期88-91,共4页
Electric Power Automation Equipment
关键词
小波变换
小波分解
支持向量机
母线
负荷预测
wavelet transforms
wavelet decomposition
support vector machines
bus
electric load forecasting