面向配电网终端设备的数字孪生映射方法研究
Digital Twin Mapping Method for Distribution Network Terminals
韩璟琳 1冯喜春 2侯若松 2刘洋2
作者信息
- 1. 国网河北省电力有限公司经济技术研究院,河北 石家庄 050000;国网河北省电力有限公司,河北 石家庄 050000
- 2. 国网河北省电力有限公司经济技术研究院,河北 石家庄 050000
- 折叠
摘要
提出了一种面向配电网终端设备的数字孪生映射方法,推动电网企业数字化转型和电力系统自动化,降低成本与安全风险.方法利用虚拟数字孪生空间映射电网物理设备,实现对区域配电网多类型设备故障的实时诊断.首先,对配电网终端设备数据进行压缩转换和级联映射,提取感知终端监测信息与物理拓扑连接关系.其次,采用改进的轻量级Yolov4 网络模型进行设备类型识别与分割,结合物理拓扑图进行数字孪生映射,形成设备孪生模型.最后,设计基于卷积注意力机制的状态评估模型,充分考虑区域设备关联与故障特征,实现对设备孪生模型的故障评估及物理空间设备的状态反馈.上述创新方法有望为电力行业带来更高效、安全的运行模式,推进电网智能化发展.
Abstract
This paper proposes a digital twin mapping method for distribution network terminal equipment,aiming to promote the digital transformation and automation of power grid enterprises while reducing costs and security risks.The method utilizes virtual digital twin mapping of the physical distribution network equipment to achieve real-time diagnosis of multiple types of equipment faults in the regional distribution network.Firstly,the data of distribution net-work terminal equipment is compressed,transformed,and cascaded to extract perception terminal monitoring informa-tion and physical topology connections.Next,an improved lightweight Yolov4 network model(Yolov4-IM)is used for device type recognition and segmentation,combined with the physical topology to form the device twin model.Lastly,a state evaluation model(SEM-CBMA)based on a convolutional attention mechanism is designed,taking into account the correlation between regional equipment and fault characteristics to achieve fault evaluation of the device twin mod-el and feedback on the physical space equipment status.This innovative approach is expected to bring more efficient and secure operating modes to the power industry,promoting the development of intelligent power grids.
关键词
数字孪生/轻量化网络/状态感知/智能电网/深度可分离卷积Key words
Digital twin/Lightweight network/State awareness/Smart grid/Depthwise separable convolution引用本文复制引用
出版年
2024
NETL
NSTL
万方数据