基于深度学习的城市配电网多级动态重构决策方法  

Multi-level Dynamic Reconfiguration Method for Urban Distribution Networks Based on Deep Learning Algorithm

作　　者：姜思远 高红均[1] 马望 王仁浚 石铖 刘俊勇[1] JIANG Siyuan;GAO Hongjun;MA Wang;WANG Renjun;SHI Cheng;LIU Junyong(College of Electrical Engineering,Sichuan University,Chengdu 610065,China;Electric Power Research Institute of Hubei Power Grid Corporation,Wuhan 430077,China)

机构地区：[1]四川大学电气工程学院,成都610065 [2]湖北省电力公司电力科学研究院,武汉430077

出　　处：《高电压技术》2024年第4期1468-1477,I0026,I0027,共12页High Voltage Engineering

基　　金：国家自然科学基金(52077146)。

摘　　要：多级动态重构技术相比于传统全局重构技术更适合用以改善大规模城市配电网的运行经济性,但是随之而来的重构级别识别又成了新问题。由此,该文依托现有的多级动态重构数学模型,提出一种基于深度学习算法的城市配电网多级动态重构决策方法,可跨过级别识别过程直接实现输入净负荷数据与最优重构决策方案之间的非线性映射。建立了将特征空间注意力机制、时间序列注意力机制、卷积神经网络与门控循环单元相结合的组合神经网络,针对城市配电网净负荷数据时空分布不均衡的特性,采用特征空间注意力机制与卷积神经网络对整个配电网的净负荷数据进行空间特征的学习以感知其高维特征空间的潜在联系。接着采用门控循环单元与时间序列注意力机制来充分挖掘净负荷数据在长时间尺度下的时序特征,提取其在时间分布上的不平衡特征,充分训练所提组合神经网络以拟合现有数学模型。最后,通过实际的145节点系统、IEEE 33节点系统、PG&E 69节点系统进行算例分析,验证了所提方法的有效性。Compared with the traditional global reconfiguration technology,multi-level dynamic reconfiguration technology is more suitable for improving the operational economics in large-scale urban distribution networks,however,the subsequent reconfiguration level identification has become a new problem.Therefore,relying on the existing mathematical model of multi-level dynamic reconfiguration,this paper proposes a decision-making method of multi-level dynamic reconfiguration for urban distribution networks based on deep learning algorithm,which can be adopted to directly realize the nonlinear mapping between the input net load data and the optimal reconfiguration decision-making scheme without the level identification process.In this paper,a combined neural network combining the feature space attention mechanism,the time attention mechanism,the convolutional neural network and the gated recurrent unit is established.In view of the unbalanced spatial and temporal distribution of the net load data,the feature space attention mechanism and the convolutional neural network are used to learn the spatial characteristics of the net load data in the entire distribution network to perceive the potential relationship of its high-dimensional feature space.Then gated recurrent unit and time attention mechanism are used to fully mine the time-series characteristics of net load data in a long-time scale,extract its unbalanced characteristics in time distribution,and fully train the combined neural network proposed in this paper so as to fit the existing mathematical model.Finally,the effectiveness of the proposed method is verified by numerical examples of 145 real bus system,IEEE 33 bus system,and PG&E 69 bus system.

关 键 词：城市配电网 多级动态重构 时空分布不均衡 组合神经网络 双重注意力机制 

分 类 号：TM73[电气工程—电力系统及自动化] TP18[自动化与计算机技术—控制理论与控制工程]