基于深度学习的发电厂设备智能巡检系统设计  

Design of intelligent inspection system of power plant equipment based on deep learning

作　　者：范厚良 FAN Houliang(Jiangsu Datang Lvsi Port International Power Generation Co.,Ltd.,Nantong 226200,China)

机构地区：[1]江苏大唐国际吕四港发电有限责任公司,江苏南通226200

出　　处：《电子设计工程》2022年第1期108-111,116,共5页Electronic Design Engineering

摘　　要：以往使用的传统人工巡检方法、无人机巡检方法缺少对关键设备的图像特征分析,导致巡检效果较差,在该情况下,提出了基于深度学习的发电厂设备智能巡检系统设计。其特点在于机器人通过服务器控制系统智能巡检,同时配备RFID读卡器,可识别电子标签数据信号,使用变磁阻式转速传感器,在线圈中检测故障位置的磁阻变化。采用深度学习方法,分析巡检原理,确定关键设备图像特征,并设计机器人巡检流程。由实验结果可知,该系统与预期执行效果一致,为用户提供了安全的巡检设备。Traditional manual inspection methods and unmanned aerial vehicle inspection methods used in the past lack the analysis of image characteristics of key equipment,resulting in poor inspection effect.Under such circumstances,an intelligent inspection system design of power plant equipment based on deep learning was proposed.The robot is characterized by intelligent inspection through the server control system and equipped with RFID card reader,which can recognize electronic tag data signals.It uses variable reluctance type speed sensor to detect the reluctance change of fault position in the coil.The deep learning method is adopted to analyze the inspection principle,determine the image characteristics of key equipment,and design the robot inspection process.The experimental results show that the system is consistent with the expected implementation effect and provides users with safe inspection equipment.

关 键 词：深度学习 发电厂设备 智能巡检 机器人控制 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]