基于贝叶斯网络的配电网净负荷预测方法  

Net Load Forecasting of Distribution Network Based on Bayesian Network

作　　者：李小伟 陶毅刚 顾洁[2] 刘书琪 黎敏 陈楚 李镕耀 LI Xiaowei;TAO Yigang;GU Jie;LIU Shuqi;LI Min;CHEN Chu;LI Rongyao(Guangxi Power Grid Co.,Ltd.,Nanning 530023,China;School of Electronic Information and Electrical Engineering,Shanghai Jiaotong University,Shanghai 200240,China)

机构地区：[1]广西电网有限责任公司,广西南宁530023 [2]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《电器与能效管理技术》2020年第9期90-98,共9页Electrical & Energy Management Technology

基　　金：国家重点基础研究计划支持项目(2016YFB0900100)。

摘　　要：配电网中可再生能源渗透率不断提高,对配电网的安全稳定运行造成威胁。通过分析用户侧负荷与可再生能源发电功率的耦合互动特性,提出了基于贝叶斯网络的配电网净负荷直接预测方法。首先根据各影响因素与配电网净负荷的因果关系建立贝叶斯网络结构,然后利用历史数据进行网络参数学习,最后生成用于配电网净负荷预测的贝叶斯网络。采用国外某地区配电网实际数据对模型进行检验,算例结果表明,与支持向量机(SVM)等预测模型相比,所提模型误差小,能有效提高配电网净负荷的预测精度。Increasing penetration of distributed renewable energy poses challenges to the safe and stable operation of the distribution network,which makes load forecasting problem considering renewable energy access scenario become the research focus.This paper analyzed the coupling and interaction between the user side load and the power generated by renewable energy and proposed a direct forecasting method of net load of distribution network based on Bayesian network.Firstly,the Bayesian network structure is established according to the causal relationship between each influencing factor and net load of distribution network,then the network parameters are learned with historical data,and finally the Bayesian network for net load prediction of distribution network is generated.The model is tested with the actual data of distribution network in a foreign country.Compared with the prediction models such as support vector machine(SVM),the prediction error of the model proposed in this paper is small,which can effectively improve the prediction accuracy of net load of distribution network.

关 键 词：配电网 贝叶斯网络 负荷预测 净负荷 

分 类 号：TM744[电气工程—电力系统及自动化]