基于改进支持向量机的湖北电网特高压规划研究  

UHV planning of hubei grid based on improved SVM

作　　者：王枫[1] 上官安琪[2] 夏俊丽[2] 

机构地区：[1]国网湖北省电力公司,湖北武汉430077 [2]武汉大学电气工程学院,湖北武汉430072

出　　处：《机电工程》2015年第8期1141-1145,共5页Journal of Mechanical & Electrical Engineering

摘　　要：针对湖北电网未来可能面临的电力紧张问题,分析了湖北电网电源装机现状和未来建设规划,提出了一种基于改进支持向量机的中长期电力负荷预测方法,该方法利用遗传算法对支持向量机进行改进,克服了支持向量机模型参数需要人工试验选取的缺陷,利用所提出的预测方法对"十三五"期间湖北省全社会最大负荷进行了预测,在此基础上对湖北电网进行了电力平衡。结果表明:该方法的预测精度比BP神经网络提高了3.66%,并且湖北电网将在2017年出现全年性电力缺口,2020年缺口将达到相当严重的1.533×107k W,因此需在"十三五"期间加快特高压建设进程。Aiming at the shortage of electricity Hubei power grid may face in the future, the present power supply situation and the future construction plan of were analyzed. A kind of mid-long term load forecasting method based on improved support vector machine（SVM） was put forward. Genetic algorithm （GA） was used to improve SVM, so that the optimal parameters can be automatically selected. The maximum load of Hubei grid during the ＂Thirteenth Five Year Plan＂ was predicted using the method proposed. On this basis, the power balance of Hu- bei grid was analyzed. The results indicate that the prediction accuracy of this method is 3.66% higher compared with BP neural network. And there will be a year-round power gap in 2017 in Hubei power grid and the gap will reach to a serious cases of 1. 533 ×107 kW in 2020. Therefore, the construction of UHV must be accelerated during the period of ＂Thirteenth Five Year Plan＂.

关 键 词：支持向量机 遗传算法 中长期负荷预测 电力平衡 特高压规划 

分 类 号：TM71[电气工程—电力系统及自动化]