基于EEMD-ARMA模型的电网工程导线价格预测  

Wire Price Forecasting of Grid Project Based on EEMD and ARMA

作　　者：卢艳超[1] 温卫宁[1] 赵彪[1] 郑燕[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11期39-44,共6页Electric Power Construction

基　　金：国家电网公司科技项目(ND71-10-005)

摘　　要：由于电网工程导线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-ARMA预测模型。利用集合经验模态分解(ensemble empirical mode decomposition,EEMD)对经验模态分解(empirical mode decomposition,EMD)进行改进,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,利用自回归滑动平均模型(auto regressive and moving average model,ARMA)对各分量进行价格预测,最后将各预测分量叠加得到预测值。以630/45导线的历史数据为样本,通过EMD-ARMA与EEMD-ARMA的预测结果进行对比及误差分析,验证了所采用的EEMD-ARMA算法较EMD-ARMA算法的预测精度更高,其预测结果对于工程造价管控和设备材料招投标具有一定的参考价值。The forecast of wire price in grid project is difficult and the accuracy is low due to the nonlinear and non- stationary property of wire price. Therefore, the EEMD-ARMA model was built. The empirical mode decomposition （EMD） was improved by using ensemble empirical mode decomposition （EEMD）. The history price was decomposed to some smooth and periodic fluctuations components by EEMD, which was used as the input of autoregressive moving-average （ARMA） model to forecast the components＇ price. Finally, the forecast price was obtained by the superimposing of forecast components. Taking the history data of 630/45 wire as samples, the forecast results of EMD-ARMA and EEMD-ARMA was compared, as well as the errors were analyzed, whose results verified that the forecast result by EEMD-ARMA was better than the result by EMD-ARMA in forecast accuracy. The forecast results by applying EEMD-ARMA model have a certain referencial value to the project cost control and the equipment and material bidding.

关 键 词：集合经验模态分解 自回归滑动平均模型 电网工程导线价格 预测 

分 类 号：TM73[电气工程—电力系统及自动化]