考虑光伏发电出力不确定性的年度最大负荷概率预测  

Probabilistic Forecast of Annual Peak Load with consideration of Photovoltaic Generation Output Uncertainties

作　　者：晁颖 金烨 朱晶亮 孙峰 张沛[1] CHAO Ying;JIN Ye;ZHU Jingliang;SUN Feng;ZHANG Pei(Beijing Jiaotong University,Beijing 100046,China;State Grid Jiaxing Power Supply Company,Jiaxing,Zhejiang 300010,China)

机构地区：[1]北京交通大学,北京100046 [2]国网嘉兴供电公司,浙江嘉兴300010

出　　处：《广东电力》2018年第9期83-89,共7页Guangdong Electric Power

摘　　要：为了在年度最大负荷预测中计及光伏、气象等因素的不确定性,提出了一种基于概率理论的年度最大负荷概率预测方法。首先,针对各影响因素的不同影响方式,利用传统预测法结合贝叶斯网络(Bayesian network,BN)加以气象修正预测,得到全社会最大负荷的概率分布;其次,利用BN理论基于历史数据建立光伏发电出力概率预测模型,提出了扩容后光伏发电出力概率预测的方法,利用联合概率分布理论实现了光伏发电出力的概率预测。最后,利用概率理论将全社会最大负荷"减"去光伏发电总出力得到年度最大负荷的概率分布预测结果。通过某地市一个配电公司的案例研究证明了这种方法的可行性。This paper presents a probabilistic forecasting method for the annual peak load taking into account uncertainties of photovoltaic, weather and other factors. First, this paper combines the traditional prediction method with Bayesian network （BN） as well as meteorologic correction to obtain probability distribution of the maximum load of the whole society. Second, this paper applies BN theory to establish the probabilistic forecasting model for photovoltaic generation output using historical data. The joint probability distribution theory is then applied to develop probability prediction functions on expanded photovoltaic generation. Finally, the probability distribution of the annual peak load of the whole society minus the probabilistic distribution of total photovoltaic generation output to obtain the annual peak load provided by power grid. Case study on a local distribution company has proved feasibility of this approach.

关 键 词：概率预测 贝叶斯网络 光伏出力不确定性 光伏扩容 年最大负荷预测 

分 类 号：TM715.1[电气工程—电力系统及自动化]