| 基于改进卷积神经网络拓扑特征挖掘的配电网结构坚强性评估方法 |
| |
| 引用本文： | 林君豪,张焰,赵腾,苏运. 基于改进卷积神经网络拓扑特征挖掘的配电网结构坚强性评估方法[J]. 中国电机工程学报, 2019, 0(1): 84-96,323 |
| |
| 作者姓名： | 林君豪 张焰 赵腾 苏运 |
| |
| 作者单位： | 上海交通大学电气工程系;全球能源互联网发展合作组织;国网上海市电力公司 |
| |
| 基金项目： | 国家863高技术研究发展计划项目(2015AA050203)~~ |
| |
| 摘 要： | 提出一种基于拓扑特征分析和深度卷积神经网络的配电网网架结构坚强性评估方法。将考虑分布式电源出力不确定性的配电网运行状态与配电网拓扑特性相结合,构建涵盖可靠性、运行裕度和结构鲁棒性3个配电网结构坚强性要点的拓扑指标集,提升评估指标刻画配电网实际状态的能力;通过多通道融合和多级池化改进卷积神经网络,解决了传统方法无法自主挖掘评估指标数据特征,以及难以直接分析不同维度的评价指标和同一指标不同尺寸的数据这两方面问题。通过对华东地区某中压配电网进行算例分析,说明所提出的评估方法的有效性和优越性。
|
| 关 键 词： | 配电网结构坚强性 分布式电源 卷积神经网络 空间金字塔池化 多卷积通道融合 |
Structure Strength Assessment Method of Distribution Network Based on Improved Convolution Neural Network and Network Topology Feature Mining |
| |
| LIN Junhao,ZHANG Yan,ZHAO Teng,SU Yun. Structure Strength Assessment Method of Distribution Network Based on Improved Convolution Neural Network and Network Topology Feature Mining[J]. Proceedings of the CSEE, 2019, 0(1): 84-96,323 |
| |
| Authors: | LIN Junhao ZHANG Yan ZHAO Teng SU Yun |
| |
| Affiliation: | (Department of Electrical Engineering,Shanghai Jiaotong University,Minhang District,Shanghai 200240,China;Global Energy Interconnection Development and Cooperation Organization,Xicheng District,Beijing 100031,China;State Grid Shanghai Municipal Electric Power Company,Pudong District,Shanghai 200122,China) |
| |
| Abstract: | LIN Junhao;ZHANG Yan;ZHAO Teng;SU Yun(Department of Electrical Engineering,Shanghai Jiaotong University,Minhang District,Shanghai 200240,China;Global Energy Interconnection Development and Cooperation Organization,Xicheng District,Beijing 100031,China;State Grid Shanghai Municipal Electric Power Company,Pudong District,Shanghai 200122,China) |
| |
| Keywords: | structure strength of distribution network distributed generation convolutional neural network spatial pyramid pooling multi-convolution channel merging |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
