基于模糊C均值聚类算法和最小二乘支持向量机的母线负荷预测研究
|
摘要
电力系统中的母线具有数量众多、负荷波动性大等特点,为提高预测精度,提出一种基于模糊C均值(fuzzy c-means,FCM)聚类算法和最小二乘支持向量机(least squares support vector machine,LSSVM)的母线负荷预测方法。利用改进的数据横向比较法对异常负荷进行修正。在此基础上,结合FCM算法将供电区域中具有相似负荷变化规律的母线进行聚类,选取聚类后的母线负荷数据对LSSVM模型进行训练和预测,并通过配比公式进行负荷分配。经数据验证,方法有助于提高母线负荷预测精度。
The bus in the electric power system is characterized by a large number and high load volatility. To improve forecasting accuracy, this paper proposed a bus load forecasting method based on fuzzy c-means(FCM) clustering algorithm and least squares support vector machine(LSSVM). An improved data transverse comparison method was applied to correct abnormal load. Then, buses with similar load variation in the specific power supply area were clustered by FCM algorithm, and clustered bus load data were selected to train and forecast the LSSVM model. Furthermore, load distribution was completed by using a matching formula. Data evaluation showed that the proposed approach could help to raise the accuracy of bus load forecasting.
The bus in the electric power system is characterized by a large number and high load volatility. To improve forecasting accuracy, this paper proposed a bus load forecasting method based on fuzzy c-means(FCM) clustering algorithm and least squares support vector machine(LSSVM). An improved data transverse comparison method was applied to correct abnormal load. Then, buses with similar load variation in the specific power supply area were clustered by FCM algorithm, and clustered bus load data were selected to train and forecast the LSSVM model. Furthermore, load distribution was completed by using a matching formula. Data evaluation showed that the proposed approach could help to raise the accuracy of bus load forecasting.
引文
[1] 康重庆,夏清,刘梅.电力系统负荷预测[M].北京:中国电力出版社,2007.
[2] 杨理才,张文磊,周勇,等.间接预测法在母线负荷预测中的应用[J].电网技术,2011,35(12):177-182.
[3] 童星,康重庆,陈启鑫,等.虚拟母线技术及其应用(一):虚拟母线辨识算法[J].中国电机工程学报,2014,34(4):596-604.
[4] 周开乐,杨善林,王晓佳,等.基于自适应模糊度参数选择改进FCM算法的负荷分类[J].系统工程理论与实践,2014,34(5):1283-1289.
[5] 吴茵,岳菁鹏,罗欣,等.基于自适应FCM算法的智能母线负荷聚类特性研究[J].四川电力技术,2015,38(6):68-71.
[6] 刘文颖,门德月,梁纪峰,等.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
[7] 杨波.基于改进模糊聚类法和CPSO-LSSVM的母线负荷预测[J].宁夏电力,2017,47(5):25-33.
[8] 董瑞.基于聚类分析的母线负荷预测研究[D].浙江:浙江大学,2015.
[2] 杨理才,张文磊,周勇,等.间接预测法在母线负荷预测中的应用[J].电网技术,2011,35(12):177-182.
[3] 童星,康重庆,陈启鑫,等.虚拟母线技术及其应用(一):虚拟母线辨识算法[J].中国电机工程学报,2014,34(4):596-604.
[4] 周开乐,杨善林,王晓佳,等.基于自适应模糊度参数选择改进FCM算法的负荷分类[J].系统工程理论与实践,2014,34(5):1283-1289.
[5] 吴茵,岳菁鹏,罗欣,等.基于自适应FCM算法的智能母线负荷聚类特性研究[J].四川电力技术,2015,38(6):68-71.
[6] 刘文颖,门德月,梁纪峰,等.基于灰色关联度与LSSVM组合的月度负荷预测[J].电网技术,2012,36(8):228-232.
[7] 杨波.基于改进模糊聚类法和CPSO-LSSVM的母线负荷预测[J].宁夏电力,2017,47(5):25-33.
[8] 董瑞.基于聚类分析的母线负荷预测研究[D].浙江:浙江大学,2015.