湖南电力 ›› 2026, Vol. 46 ›› Issue (3): 153-158.doi: 10.3969/j.issn.1008-0198.2026.03.020

• 电力防灾减灾 • 上一篇    

可越障变电站智能巡检装置的研制与应用

张家豪1, 李琦1, 孙远航1, 韩亮2   

  1. 1.国网吉林省电力有限公司延边供电公司,吉林 延吉 133000;
    2.沈阳嘉越电力科技有限公司,辽宁 沈阳 110136
  • 收稿日期:2025-11-28 修回日期:2025-12-24 出版日期:2026-06-25 发布日期:2026-07-07
  • 作者简介:张家豪(1985),男,学士,高级工程师,研究方向为变电运维技术。李琦(1992),男,学士,工程师,研究方向为变电运维技术。孙远航(1998),男,硕士,助理工程师,研究方向为变电运维技术。韩亮(1988),男,学士,助理工程师,研究方向为电力设备智能巡检。
  • 基金资助:
    国网延边供电公司2024年群创专项(SGJLYB00BYJS2400895)

Research and Application of Intelligent Inspection De‍vices for Overcoming Obstacles in Substations

ZHANG Jiahao1, LI Qi1, SUN Yuanhang1, HAN Liang2   

  1. 1. State Grid Yanbian Power Supply Company, Yanji 133000, China;
    2. Shenyang Jiayue Electric Power Technology Co., Ltd., Shenyang 110136, China
  • Received:2025-11-28 Revised:2025-12-24 Online:2026-06-25 Published:2026-07-07

摘要: 变电站巡检是保障电力设备安全运行的关键环节,目前普遍采用的固定式监测、人工巡检与机器人智能巡检等方式,在实际应用中存在功能重叠、资源浪费、效率低下等问题,亟须进行系统性整合与优化。基于此,研制一种基于智能识别技术的可越障变电站巡检装置。首先,进行智能巡检装置的硬件模块设计。其次,构建智能巡检装置运行原理示意图,采用YOLOv5图像识别算法和融合改进A*(A-star)与动态窗口法 (dynamic window approach,DMA)的运动系统算法分别对智能巡检装置的视觉模块与运动模块进行设计;搭建能够方便操控智能巡检装置行进、图像采集云平台视角调整与焦距、视频流回传等动作的软件系统平台。最后,组装图像采集云台、无线通信模组与智能巡检装置本体,通过现场实际应用验证该智能巡检装置及其操控方法的合理性和有效性。

关键词: 变电站, 智能识别, YOLOv5, A*, 动态窗口算法

Abstract: Substation inspection is an effective measure to ensure the safe operation of power equipment. The existing fixed monitoring, manual inspection and robot intelligent inspection methods have problems such as single monitoring indicators, high inspection work intensity and low efficiency, and limited inspection angles. In response to this, this paper develops an obstacle-crossing substation inspection device based on intelligent recognition technology. Firstly,the study designs the various hardware components that make up the intelligent inspection device.Secondly, the study provides a schematic diagram of the operating principle of the intelligent inspection device and designs the visual module and motion module of the intelligent inspection device using YOLOv5 image recognition algorithm and motion system algorithm that integrates improved A*(A-star) and dynamic window approach(DMA). Then, the paper develops a software system platform that can conveniently control the movement of intelligent inspection devices, adjust the perspective and focal length of the image acquisition cloud platform, and transmit video streams. Finally, the image acquisition pan-tilt, wireless communication module and the body of the intelligent inspection device are assembled. The rationality and effectiveness of the intelligent inspection device and its control method are verified through practical application on site.

Key words: substation, intelligent identification, YOLOv5, A*, dynamic window approach

中图分类号: