变电站二次设备全向巡检机器人精确定位与作业辨识

Precise positioning and identification of omnidirectional inspection robot for substation secondary equipment

通过巡检机器人进行变电站二次设备的监测是提升电力设备自动化、智能化管理的重要方式,有利于保障电力工程设备的安全运行。文中开发了一种用于变电站二次设备自动化巡检的Mecanum轮式全向移动机器人,其具备自主导航定位与作业辨识的能力,可极大提升设备巡检效率及保护压板状态识别准确性。通过Mecanum轮的驱动方式实现巡检机器人在狭窄作业环境下的灵活运动与姿态调整,通过多轨道升降平台实现对350~1 800 mm高度范围内的二次设备及压板的图像采集与状态辨识。机器人采用基于激光雷达的即时定位与地图构建(simultaneous localization and mapping,SLAM)导航方法进行自主定位导航,并结合基于视觉的路径提取与跟踪算法进行姿态位置修正,实现在待测点位置的精确定位。同时,提出了基于颜色辨识的图像排列与状态辨识方法,针对二次设备保护压板连通状态进行识别和判断。实验结果表明,研制的变电站二次设备巡检机器人可以实现自主导航与位置精确定位,在路径跟踪过程中最大偏角和偏距分别为±3°和±8 mm。结合机器视觉与颜色辨识的压板辨识方法可以准确识别压板状态,识别准确率大于95.80%,有助于提升机器人自动化的电力巡检作业水平。

Substation secondary equipment is monitored through inspection robots,providing an important means to enhance power equipment automation and intelligent management,thereby ensuring the safe operation of power engineering equipment.In this paper,a Mecanum wheeled omnidirectional mobile robot is developed for automatically inspecting secondary equipment in substations.It possesses autonomous navigation,positioning,and identification capabilities,significantly enhancing inspection efficiency and the accuracy of protection plate state identification.The Mecanum wheels enable the inspection robot to move flexibly and adjust its attitude within narrow working environments.Meanwhile,a multi-track lifting platform facilitates image acquisition and state identification of the secondary equipment pressure plate,covering a height range of 350-1 800 mm.The robot employs the lidar-based simultaneous localization and mapping(SLAM) method for autonomous positioning and navigation,supplemented by a vision-based path extraction and tracking algorithm for precise position correction at set points.Moreover,a color recognition-based image arrangement and state recognition method is proposed to accurately identify and assess the connection state of the secondary equipment protection plate.Experimental results demonstrate that the substation secondary equipment inspection robot,utilizing Mecanum wheels,achieves successful autonomous navigation and precise positioning,with maximum deflection angles and distances during path tracking processes being ±3° and ±8 mm,respectively.Additionally,the plate recognition method,combining machine vision and color recognition,achieves an outstanding recognition accuracy rate exceeding 95.80%,thereby elevating the level of robot automation in inspection operations.