中文摘要：

与系统负荷不同，母线负荷水平较低，波动性强，规律性弱，可预测性差，引入虚拟母线技术有利于提高母线负荷预测的工作水平与预测精度。在电网中存在一些紧密联系的连通的局部网络，其内部各母线对关键断面具有相同或相近的发电机输出功率转移分布因子（generation shift distribution factor，GSDF），同时其内部母线的负荷曲线具有一定的相似性，这些母线所组成的连通局部母线组为虚拟母线。根据虚拟母线概念，提出一种新颖的聚类预测方法：虚拟母线负荷预测方法；对虚拟母线负荷预测与独立预测的精度进行了比较，分析了预测误差的形成机理，提炼出预测误差的数学表征；针对虚拟母线聚类判据对于提高预测精度的有效性进行了分析；采用河北南网的实际数据验证了虚拟母线负荷预测方法的有效性。

英文摘要：

Compared to system load, bus load is more difficult to accurately forecast, as there are too many buses within the power system and the bus load is always volatile and with weak regularity. Therefore, introduction of virtual bus technique would be helpful to improve the accuracy of bus load forecasting. Virtual bus is one type of the closely connected local electric networks in the grid. The internal buses in the virtual bus share a similar load curve and have the same generation shift distribution factor to the critical transmission section. Based on virtual bus, this paper presented a novel virtual bus load forecasting （VBLF） method, which is based on clustering method. Forecasting accuracies of VBLF and individual forecasting were compared. Then, the mechanism of forecasting error was analyzed and formulated. On this basis, the effectiveness of VBLF on improving forecasting accuracy was analyzed. A numerical case using real data of Hebei South Power Grid verified the effectiveness of the proposed method.