基于支持向量机和互联网信息修正的空间负荷预测方法

为提高电网规划阶段的空间负荷预测精度，提出了一种基于支持向量机和互联网信息修正的空间负荷预测（spatial load forecasting，SLF）方法，该方法分为3个步骤：一是基于k-均值聚类分析和支持向量回归模型得到地块负荷初始预测值；二是基于地块负荷历史数据计算负荷实际值与初始预测值之间的偏差；三是针对这些偏差，利用搜索引擎获取互联网信息，识别造成偏差的不确定事件，包括元胞中新增大负荷事件和元胞中企业营收增长率突变事件。定性分析事件对空间负荷的影响，并建立这两类事件与其造成的影响之间的分类事件影响定量模型，基于该模型对地块负荷初始预测值进行修正，得到规划区域内的地块负荷预测值。通过对北京某地区进行算例验证，结果表明该方法可以提高预测精度，可用于配电网以及能源互联网规划中的空间负荷预测。

Spatial Load Forecasting Method Based on Support Vector Machine and Internet Information Correction

In order to improve the spatial load forecasting accuracy of power system planning, a spatial load forecasting (SLF) method is proposed based on support vector machine (SVM) and Internet information correction. The method can be divided into three steps: firstly, based on k-means clustering analysis, the SVM model is used to get the initial prediction value of the block load; secondly, the deviation between the actual load value and the predicted load value can be calculated based on the historical data of the block load; and thirdly, those uncertain events that causes deviation, including the newly increasing load events in the cells and the mutation of revenue growth rate of the enterprises in the cells, are identified by Internet information that is obtained by search engine. The impact of the uncertain events on spatial load is qualitatively analyzed, and a quantitative impact model is established for classifying the events between these two events and their effects. Based on the model, the initial forecasting value of the load is corrected, and the final forecasting load values of the blocks in the planning area is obtained. A case study of an area in Beijing shows that this method can improve the prediction accuracy and can be used for spatial load forecasting in distribution network and energy Internet planning.