摘要
随着大量的分布式电源和柔性负荷并网,配电网面临的线路损耗增加和电压越限问题给制定无功优化策略带来挑战。配电网的无功优化模型需要详细的拓扑结构、线路参数和负荷的历史数据,建模过程复杂且优化的准确性较差。为此,提出一种基于深度信念网络(deep belief network,DBN)的配电网无功功率优化的新型数据驱动方法,通过挖掘现有先验知识和海量数据制定无功优化策略,利用卷积层将负荷参数映射到特征向量,完成负荷数据的全面提取。此过程无需明确构建物理模型,且数据分析层能精准分析负荷的输入数据与调度策略之间的非线性关系,通过训练DBN形成新的数据集。最后,在Linux操作系统的PyCharm平台对IEEE 33节点配电网进行仿真验证,结果表明DBN的准确性和鲁棒性高于传统的优化算法,能够满足配电网无功优化的需求。
With the integration of a large number of distributed power sources and flexible loads into the grid,distribution networks are confronted with issues such as increased line losses and voltage violations,which pose challenges to the formulation of reactive power optimization strategies.Given that the reactive power optimization model for distribution networks requires detailed topological structures,line parameters,and historical load data,and that the modeling process is complex with relatively poor optimization accuracy,this paper proposes a novel data-driven method for reactive power optimization in distribution networks based on deep belief networks.By mining existing prior knowledge and massive data to formulate reactive power optimization strategies,the convolutional layer maps load parameters to feature vectors,achieving comprehensive extraction of load data.This process does not require the explicit construction of physical models,and the data analysis layer can precisely analyze the nonlinear relationship between the input load data and the dispatching strategy.Then,the deep belief network is then trained to form a new dataset.Finally,the IEEE 33-bus network is simulated and verified on the PyCharm platform under the Linux operating system.The simulation results show that the accuracy and robustness of the deep belief network are higher than those of traditional optimization algorithms,and that it can meet the requirements of reactive power optimization in distribution networks.
作者
钟毅鸿
覃福明
李梓熔
ZHONG Yihong;QIN Fuming;LI Zirong(Wuzhou Power Supply Bureau of Guangxi Power Grid Co.,Ltd.,Wuzhou,Guangxi 543099,China)
出处
《广西电力》
2025年第1期14-21,共8页
Guangxi Electric Power
关键词
配电网
无功优化
深度信念网络
配电网设备编码
distribution network
reactive power optimization
deep belief network
distribution network equipment code
作者简介
通信作者:钟毅鸿(1998),男,广西梧州人,助理工程师,工学学士,研究方向为配电运维技术、配电自动化技术,645975075@qq.com;覃福明(1995),男,广西梧州人,助理工程师,工学学士,研究方向为配电网运行与管理、配电自动化技术,1053026773@qq.com;李梓熔(2001),女,广西梧州人,工学学士,研究方向为配电自动化技术,13878493220@163.com。
