基于EEMD-SVM模型的电网工程设备价格预测  

Price Forecast of Power Grid Equipment Based on EEMD-SVM Model

作　　者：卢艳超[1] 温卫宁[1] 赵彪[1] 郑燕[1] 史雪飞[1] 

机构地区：[1]国网北京经济技术研究院,北京市100052

出　　处：《电力建设》2013年第1期25-30,共6页Electric Power Construction

基　　金：国家电网公司科技项目(KB4410110002)

摘　　要：由于电网工程设备价格具有非线性和非平稳性特征,导致其价格预测难度大、预测精度低,针对这一问题,建立了EEMD-SVM预测模型。利用集合经验模态分解理论(ensemble empirical mode decomposition,EEMD)对经验模态分解理论(empirical mode decomposition,EMD)进行了改进,通过EEMD将历史价格分解为平稳的、周期波动的若干价格分量,并以此作为输入,对各分量进行基于支持向量机(support vector machine,SVM)的价格预测,最后将各预测分量叠加得到预测值。以220kVA柱式断路器的历史数据为样本,通过EMD-SVM与EEMD-SVM的预测结果进行对比及误差分析,证明EEMD-SVM比EMD-SVM的预测精度更高,其预测结果对于工程造价管控和设备招投标具有一定的参考价值。As the price of power grid equipment is nonlinear and non-stable, which leads to big difficulty and low accuracy of forecast, the model of ensemble empirical mode decomposition （EEMD） and support vector machine （SVM） has been built. The empirical mode decomposition （EMD） is improved by EEMD. The history price is divided into several smooth and periodically fluctuated components by EEMD, which is used to forecast with SVM as an input. Finally, the superposition value of forecasted components has been gained as the predictive value of price. Taking the history data of 220 kVA column circuit breaker as samples, the prediction results of EMD-SVM and EEMD-SVM are compared and the errors are analyzed. The results show that the prediction accuracy of EEMD-SVM is better than that of EMD-SVM, which can make references to project cost control and equipment bidding to some extent.

关 键 词：电网工程设备价格 集合经验模态分解 支持向量机 预测 

分 类 号：TM7[电气工程—电力系统及自动化]