浏览全部资源
扫码关注微信
基于边缘代理及轻量化深度学习模型的变电站智能运检技术研究
Study on Intelligent Operation and Inspection Technology of Substation Based on Edge Agent and Lightweight Deep Learning Model
- 高压电器 2025年 页码：1-7
- 作者机构：
1.国网山西省电力公司电力科学研究院，太原 030001
2.国网山西省电力公司，太原 030001
- 作者简介：
杨冬冬（1983—），男，研究生，高级工程师，研究方向从事配电网运检及智能配用电工作；长期从事配电网新技术研究、新产品推广、智能运检及配用电工作（E-mail：594122581@qq.com）。
王瑞珏（1983—），男，研究生，高级工程师，研究方问为配电运行维护及设备状态评价。
郑志宏（1991—），男，研究生，工程师，研究方向为供电可靠性评价。
杨罡（1983—），男，博士，高级工程师，研究方向为配电网运检及智能配用电研究。
- 基金信息：国网山西省电力公司科技项目(520530202002)
DOI：
中图分类号：-
纸质出版日期：
- 稿件说明：
移动端阅览
杨冬冬, 王瑞珏, 郑志宏, 等. 基于边缘代理及轻量化深度学习模型的变电站智能运检技术研究[J/OL]. 高压电器, 2025,1-7.
YANG Dongdong, WANG Ruijue, ZHENG Zhihong, et al. Study on Intelligent Operation and Inspection Technology of Substation Based on Edge Agent and Lightweight Deep Learning Model[J/OL]. High voltage apparatus, 2025, 1-7.
摘要
随着物联网和人工智能技术日趋发展，巡检机器人已在变电站开展研究并取得了较多成效，但因受现场地形、电磁等环境影响，以及监测对象安装位置的局限性，实用性存在一定不足，且受限于作业时间窗口，无法对重点设备重点对象实现实时在线监测。为此文中提出了以边缘代理集成轻量化深度学习算法作为变电站重要设备的运检补充手段。文中在基于开源的深度学习模型基础上进行了裁剪优化，采用了百度较为先进的飞桨Paddle Lite框架，实现了在边缘代理装置上集成部署，并以变压器油温表计读数识别为典型应用案例进行研究。实验结果验证了文章研究成果实现了在边缘代理设备有限的硬件资源下的集成部署深度学习模型，并满足了变压器油温表识别的准确率和性能的应用要求。研究成果可应用在变电站内对人员身份识别、违章违规作业、环境安全监控、仪表自动读数等智能运检业务领域，具备较好的先进性和推广性。
Abstract
With the development of Internet of things and artificial intelligence technology
patrol inspection robot has carried out research in substation and achieved many results. Due to the influence of the terrain and electromagnetic environment on the site
as well as the limitations of the installation location of the monitoring object
there are many deficiencies in the practicability of the inspection robot. So we propose to use the edge agent integrated lightweight deep learning algorithm as a supplementary means for the operation inspection of important substation equipment. Based on the open source in-depth learning model
we have carried out cutting optimization
adopted Baidu's advanced paddle Lite framework andtaken the reading recognition of transformer oil thermometer as a typical application case. The researching results can be widely used in the field of power grid intelligent operation inspection
such as personnel identification
illegal operations
environmental safety monitoring
automatic meter reading
etc. in substations. The research results of this paper have good progressiveness and popularization.
关键词
Keywords
references
蒲世亮 , 袁婷婷 . 基于云边融合的物联网智能服务架构探讨 [J ] . 智能物联技术 , 2018 , 1 ( 1 ): 1 - 6 .
PU Shiliang , YUAN Tingting . Research on intelligent architecture for IoT services based on fusion of cloud and edge computing [J ] . Technology of IoT& AI , 2018 , 1 ( 1 ): 1 - 6 .
周知 , 于帅 , 陈旭 . 边缘智能：边缘计算与人工智能融合的新范式 [J ] . 大数据 , 2019 , 5 ( 2 ): 53 - 63 .
ZHOU Zhi , YU Shuai , CHEN Xu . Edge intelligence: A new nexus of edge computing and artificial intelligence [J ] . Big Data Research , 2019 , 5 ( 2 ): 53 - 63 .
李志海 , 张荣华 , 张天兵 , 等 . 基于视频监控的变电站智能巡检系统研究 [J ] . 计算机技术与发展 , 2014 , 24 ( 10 ): 233 - 237 .
LI Zhihai , ZHANG Ronghua , ZHANG Tianbing , et al . Research on intelligence substation routine inspection system based on video surveillance [J ] . Computer Technology and Development , 2014 , 24 ( 10 ): 233 - 237 .
施巍松 , 张星洲 , 王一帆 , 等 . 边缘计算：现状与展望 [J ] . 计算机研究与发展 , 2019 , 56 ( 1 ): 69 - 89 .
SHI Weisong , ZHANG Xingzhou , WANG Yifan , et al . Edge computing: State-of-the-art and future directions [J ] . Journal of Computer Research and Development , 2019 , 56 ( 1 ): 69 - 89 .
汤鹏杰 , 王瀚漓 , 左凌轩 . 并行交叉的深度卷积神经网络模型 [J ] . 中国图象图形学报 , 2016 , 21 ( 3 ): 339 - 347 .
TANG Pengjie , WANG Hanli , ZUO Lingxuan . Parallel cross deep convolution neural networks model [J ] . Journal of Image and Graphics , 2016 , 21 ( 3 ): 339 - 347 .
卢宏涛 , 张秦川 . 深度卷积神经网络在计算机视觉中的应用研究综述 [J ] . 数据采集与处理 , 2016 , 31 ( 1 ): 1 - 17 .
LU Hongtao , ZHANG Qinchuan . Applications of deep convolutional neural network in computer vision [J ] . Journal of Data Acquisition & Processing , 2016 , 31 ( 1 ): 1 - 17 .
CARRLLL J . Deep learning continues growth in machine vision [J ] . Vision Systems Design , 2020 , 25 ( 3 ): 20 - 23 .
CARROLL J . Deep learning: Finding the optimal hardware for deep learning inference in machine vision [J ] . Vision Systems Design , 2019 , 24 ( 8 ): 8 - 9 .
黄斌 , 卢金金 , 王建华 , 等 . 基于深度卷积神经网络的物体识别算法 [J ] . 计算机应用 , 2016 , 36 ( 12 ): 3333 - 3340 .
HUANG Bin , LU Jinjin , WANG Jianhua , et al . Object recognition algorithm based on deep convolution neural networks [J ] . Journal of Computer Applications , 2016 , 36 ( 12 ): 3333 - 3340 .
RIFAIOGLU A S , ATAS H , MARTIN M J , et al . Recent applications of deep learning and machine intelligence on in silico drug discovery: Methods, tools and databases [J ] . Briefings in Bioinformatics , 2019 , 20 ( 5 ): 1878 - 1912 .
谢景洋 , 王巍 , 刘婷 . 基于YOLOv3算法的不同主干网络对织物瑕疵检测 [J ] . 测控技术 , 2021 , 40 ( 3 ): 61 - 66 .
XIE Jingyang , WANG Wei , LIU Ting . Fabric surface defect detection based on YOLOv3 with different backbone networks [J ] . Measurement & Control Technology , 2021 , 40 ( 3 ): 61 - 66 .
吴凤和 . 基于计算机视觉测量技术的图像轮廓提取方法研究 [J ] . 计量学报 , 2007 , 28 ( 1 ): 18 - 22 .
WU Fenghe . A study on contour extraction method in computer vision measurement technology [J ] . Acta Metrologica Sinica , 2007 , 28 ( 1 ): 18 - 22 .
曾爱林 . 基于改进神经网络的图像生物特征识别方法 [J ] . 科技通报 , 2015 , 31 ( 2 ): 224 - 226 .
ZENG Ailin . Image biometric identification method based on improved neural network [J ] . Bulletin of Science and Technology , 2015 , 31 ( 2 ): 224 - 226 .
CAMILLO J . Deep learning machine vision [J ] . Assembly , 2019 , 62 ( 5 ): 48 - 50 .
马艳军 , 于佃海 , 吴甜 , 等 . 飞桨：源于产业实践的开源深度学习平台 [J ] . 数据与计算发展前沿 , 2019 , 1 ( 1 ): 105 - 115 .
MA Yanjun , YU Dianhai , WU Tian , et al . PaddlePaddle: An open-source deep learning platform from industrial practice [J ] . Frontiers of Data& Computing , 2019 , 1 ( 1 ): 105 - 115 .
陈玫玫 , 玄玉波 , 李兆玺 , 等 . 基于百度飞桨AI Studio的机器学习教学新模式实践与探索 [J ] . 计算机教育 , 2021 ( 9 ): 46 - 50 .
CHEN Meimei , XUAN Yubo , LI Zhaoxi , et al . et al practice and exploration of a new machine learning teaching mode based on baidu PaddlePaddle AI studio [J ] . Computer Education , 2021 ( 9 ): 46 - 50 .
麻兴东 . 计算机智能图像识别算法研究 [J ] . 科技资讯 , 2015 , 13 ( 20 ): 24 - 25 .
MA Xingdong . Research on computer intelligent image recognition algorithm [J ] . Science & Technology Information , 2015 , 13 ( 20 ): 24 - 25 .
卢思缴 , 张智宏 , 刘颖 , 等 . 基于计算机视觉技术的电力指针表自动读数识别研究 [J ] . 自动化技术与应用 , 2021 , 40 ( 11 ): 177 - 179 .
LU Sijiao , ZHANG Zhihong , LIU Ying , et al . Research on automatic reading recognition of electric pointer based on computer vision technology [J ] . Techniques of Automation and Applications , 2021 , 40 ( 11 ): 177 - 179 .
0
浏览量
1
下载量
0
CSCD
- 网络PDF下载
- Meta-XML下载
- BibTeX下载
- Endnote下载
- RefMan 下载
- NoteExpress下载
- NoteFirst下载
关联资源
相关文章
相关作者
相关机构
- 技术支持由北京北大方正电子有限公司提供 陕ICP备05005324号-4
陕公网安备 61010402000197 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies