机电工程技术2026，Vol.55Issue(14)：55-60,66,7.DOI:10.3969/j.issn.1009-9492.2025.00153
基于UNet检测的核电冷源取水隧洞清扫机器人设计
Design of a Cleaning Robot for Nuclear Power Plant Cooling Water Intake Tunnels Based on UNet Detection
摘要
Abstract
To address safety hazards in nuclear power plant cooling water intake tunnels—including silt deposition,marine organism blockages,and crack damage—along with the low efficiency and high risk of traditional manual cleaning,a tunnel cleaning robot is designed integrating a tracked mobile platform with an extendable robotic arm.Three-dimensional modeling of the robot is completed using SolidWorks.Lightweight,corrosion-resistant materials like modified resin are employed,with weight reduction and heat dissipation achieved through hollowed-out structural optimization.An adjustable water spray device and a three-finger gripping mechanism are configured to enhance the removal capability of complex deposits.The control system utilizes a Raspberry Pi 4B as the main controller,integrating BlueOS and Arduino to build a collaborative architecture.Stable long-distance communication is enabled via a power line carrier module.A crack detection model is developed using the UNet neural network.After 80 training iterations,the model achieved an 80%recognition accuracy in field tests.Experimental validation confirmed the robot's structural strength meets tunnel operational requirements and its control response remains stable.A feasible technical solution for automated and safe tunnel cleaning in nuclear power plants is provided.关键词
隧洞清理机器人/可伸缩机械臂/UNet神经网络/裂缝检测Key words
tunnel cleaning robot/telescopic manipulator/UNet neural network/crack detection分类
信息技术与安全科学引用本文复制引用
钟耀鹏,黄健威,李光鸿,谭文就,陈彦宇,陆万松,黄杰翰..基于UNet检测的核电冷源取水隧洞清扫机器人设计[J].机电工程技术,2026,55(14):55-60,66,7.基金项目
广东大学生科技创新培育专项资金资助项目(pdjh2024a189) （pdjh2024a189）
