基于三种模型的昆士兰州短期电力负荷预测对比探究  

The comparative study on short-term electricity load forecasting based on three models

作　　者：智勇[1] 何欣[1] 梁琛[1] 刘巍[1] 

机构地区：[1]国网甘肃省电力公司电力科学研究院,甘肃兰州730050

出　　处：《电子测试》2015年第2X期49-51,共3页Electronic Test

摘　　要：为提高电力负荷预测模型的预测精度本文将基本ARMA模型与Kalman滤波结合建立ARMA-Kalman滤波模型,将ARMA模型与SVM模型以及SVM的优化方法结合,建立ARMA-SVM模型,以试图提高模型的预测精度。同时将前两个模型以及基本ARMA模型应用于澳大利亚昆士兰州短期电力负荷预测中。实例研究表明,ARMA-Kalman模型未能如期较大提高对澳大利亚昆士兰州的电力负荷预测精度,ARMA-SVM模型在一定程度上提高了预测精度。To improve the forecasting accuracy of electricity load, the basic ARMA model and Kalman filter are combined into ARMA-Kalman filter model,ARMA model and SVM model along with its optimization methods are associated into ARMA-SVM model in this paper.The two models as well as basic ARMA model are employed in the forecasting of short-term electricity load in Queensland,Australia.The case study shows that ARMA-Kalman model can’t improve the forecasting accuracy as expected,while ARMA-SVM model improve the forecasting accuracy to some extent.

关 键 词：ARMA模型 KALMAN滤波 SVM模型 电力负荷预测 

分 类 号：TM715[电气工程—电力系统及自动化]