分布式深度学习框架下配电网无功补偿灰色关联优化规划模型  

Grey Relational Optimization Planning Model of Reactive Power Compensation in Distribution Network under Distributed Deep Learning Framework

作　　者：兰丹阳 王雨舸 陆鑫 曹宁 张世军 陈奎印 LAN Danyang;WANG Yuge;LU Xin;CAO Ning;ZHANG Shijun;CHEN Kuiyin(Xi’an Power Supply Company of State Grid Shaanxi Electric Power Co.,Ltd.,Xi’an 710000,China;State Grid Info-Telecom Great Power Science and Technology Co.,Ltd.,Fuzhou 350003,China;Construction Branch,State Grid Shaanxi Electric Power Co.,Ltd.,Xi’an 710065,China)

机构地区：[1]国网陕西省电力有限公司西安供电公司,陕西西安710000 [2]国网信通亿力科技有限责任公司,福建福州350003 [3]国网陕西省电力有限公司建设分公司,陕西西安710065

出　　处：《微型电脑应用》2024年第5期223-226,共4页Microcomputer Applications

摘　　要：配电网网损较高,投入成本较多,节点功率参数不稳定,由此提出分布式深度学习框架下配电网无功补偿灰色关联优化规划模型。以网络损耗最小、电压品质最优、经济利益最大为目标函数,综合潮流、控制变量、电压约束等约束条件,建立配电网无功补偿规划模型;运用灰色关联分析的方法确定关联度,通过深度学习神经网络以及参数更新后的分布式深度学习框架,求取全局最优解,完成配电网无功补偿优化规划。实验证明,该模型可以有效地规划配电网无功补偿模式,降低网损和投入成本,提升电压质量,保证了配电网无功补偿效果,促进了整个电网的稳定运行。The distribution network loss is high,the input cost is high,and the node power parameters are unstable.Taking the minimum network loss,the optimal voltage quality and the maximum economic benefits as the objective functions,and integrating the constraints of power flow,control variables and voltage constraints,a reactive power compensation planning model for distribution network is established.The grey correlation analysis method is used to determine the correlation degree.Through the deep learning neural network and the distributed deep learning framework with updated parameters,the global optimal solution is obtained to complete the optimal planning of distribution network reactive power compensation.The experiment shows that the reactive power compensation mode of the distribution network can be effectively planed,reduce the network loss and input cost,improve the voltage quality,ensure the reactive power compensation effect of the distribution network,and promote the stable operation of the entire network.

关 键 词：深度学习 无功补偿 分布式框架 灰色关联 

分 类 号：TM7[电气工程—电力系统及自动化]