考虑联络线负荷匹配特性的供电分区深度迁移学习负荷预测

Deep transfer learning load prediction for power supply zones considering the load matching characteristics of contact lines

合理有效地进行供电分区划分及负荷预测既是挖掘供电潜力、降低电网投资的有效方法,也是新型电力系统规划与运行的重要基础。针对当前供电分区方法对负荷同时率参数选择的依赖较大,且未考虑站间联络线的负荷匹配,以及负荷预测学习方法的精确性仍需进一步提高等问题,提出了考虑联络线负荷匹配特性的供电分区深度迁移学习负荷预测方法。首先,构建了供电分区及负荷预测整体技术框架;其次,构建了考虑联络线负荷匹配特性的供电分区划分模型,并提出了相应的求解方法;在此基础上,提出了基于深度迁移学习的分区负荷预测方法,并进一步提出了供电分区划分及负荷预测一体化流程;最后,通过实际算例仿真分析验证了所提理论方法的有效性与适用性。

The rational and effective zoning of power supply and load prediction is an effective way to exploit the power supply potential and reduce the investment in the power grid,and is also an important basis for the planning and operation of modern power systems.To address the problems that the current power supply zoning method relies heavily on the selection of load simultaneous rate parameters and does not consider the matching of interstation contact lines,and the accuracy of the load prediction learning method still needs further improvement,a deep transfer learning load prediction method for power supply zones considering the load matching characteristics of contact lines is proposed.Firstly,the overall technical framework of power supply partitioning and load prediction is constructed.Secondly,a model of power supply partitioning considering the matching characteristics of contact line loads is constructed and the corresponding solution method is also proposed.On this basis,a partitioned load prediction method based on deep transfer learning and an integrated process of power supply partitioning and load prediction are proposed.Finally,the validity and applicability of the theoretical method proposed are verified through the simulation analysis of practical cases.