摘要
负荷模型的准确度对电力系统仿真的有效性影响很大,但是建立反映实际负荷特性的负荷模型目前仍是一个尚未解决的难题,这主要是由于负荷的组成、大小与特性时刻都处于变化之中,因此根据已获得的负荷数据所建立的负荷模型不一定具备好的泛化能力,即该负荷模型虽然能够精确地拟合已有的负荷数据,但未必能够描述未知的负荷特性。文中应用支持向量机工具,建立了负荷数据的特征样本空间;并以辽宁虎石台变电站2004年所有负荷数据为例,研究对比了以下3类负荷模型:基于某月全部数据所建立的负荷模型;基于随机样本所建立的负荷模型;基于特征样本空间所建立的负荷模型的泛化能力。研究表明:基于文章提出的特征样本空间所建立的负荷模型对整个负荷样本空间内的数据都具有强的解释能力,因此具有很好的泛化能力。
The accuracy of the load model has great effects on power system analysis and control. However, load modeling has been a well-known difficult problem and unsolved so far, which is mainly due to the fact that the load is always changing, both in its amount and in its constitutes. Since the load model can only be built on the recorded measurements, the generalization capability of the model has great effects on its validity. By applying the support vector machine tool, this paper proposes an approach to build the load characteristic space, which is only a small subset of the whole data space. The generalization capability of three types of load model is further researched, namely, the load model built on the proposed characteristic space, the load model built on the data from one month as well as the load model built on the data selected randomly from each month. Case studies on all the load data recorded in Hushitai Substation of 2004 have shown that the model built on the load characteristic space has a very strong generalization capability.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2006年第21期29-35,共7页
Proceedings of the CSEE
基金
国家重点基础研究专项经费项目(2004CB217901)~~
关键词
电力系统
负荷建模
统计分析
泛化能力
power system
load modeling
statistic analysis
generalization capability
