摘要
燃气闸井作为城市燃气基础设施的重要组成部分,其有限空间作业环境具有进出口受限、通风条件差、结构复杂等特点,易导致中毒窒息、气体爆炸等安全事故。本研究针对燃气闸井作业中的主要危险源,结合人工智能、激光气体检测、北斗定位等先进技术,构建了一套智能化的安全监管体系。通过研制作业现场安全监管设备,实现了对作业前、作业中(井上及井下)的全流程动态监管。作业前,系统自动检测劳保装备合规性及井内气体浓度(甲烷、一氧化碳等),确保环境安全;作业中,利用AI图像识别技术监控井上人员违规行为,并通过激光气体检测与智能安全帽实时监测井下气体环境及作业人员健康状态。此外,研究采用改进的YOLOv5算法和BoT-SORT目标跟踪技术,结合Mosaic数据增强与FPN+PAN多尺度特征融合,提升了危险行为与气体异常的识别精度。实验表明,该体系显著提高了燃气闸井作业的安全性与监管效率,为有限空间作业的标准化、智能化管理提供了科学依据。未来,可进一步优化算法模型并拓展技术集成应用范围,以应对更复杂的作业场景。
Gas valve wells serve as a critical component of urban gas infrastructure,but its Confined space features restricted access,poor ventilation,and complex structures,posing risks of asphyxiation,poisoning,and gas explosions.This study focused on the primary hazards in gas valve well operations and constructed an intelligent safety supervision system by integrating advanced technologies such as artificial intelligence(AI),laser gas detection,and BeiDou positioning.By developing on-site safety supervision equipment,the system enabled dynamic supervision throughout the entire workflow—before and during(both above ground and within the well)operations.Before operations,the system automatically checked the compliance of personal protective equipment and monitored gas concentrations(e.g.,methane and carbon monoxide)within the well to ensure a safe environment.During operations,AI-based image recognition was employed to monitor safety violations by personnel above ground,while laser gas detection and intelligent safety helmets were used to track gas conditions and the health status of workers within the well.Additionally,the study incorporated an improved YOLOv5 algorithm and BoT-SORT target tracking technology,combined with Mosaic data augmentation and FPN+PAN multi-scale feature fusion,to improve the accuracy of identifying hazardous behaviors and abnormal gas levels.Experiments demonstrate that this system significantly enhances the safety and supervision efficiency of gas valve well operations,providing a scientific basis for the standardization and intelligent management of confined space work.Future work may focus on optimizing algorithm models and expanding the scope of integrated technological applications to address more complex operational scenarios.
作者
张明作
王林
王生平
鲍青
王明星
ZHANG Mingzuo;WANG Lin;WANG Shengping;BAO Qing;WANG Mingxing(Beijing Gas Group Co.,Ltd.)
出处
《城市燃气》
2026年第2期49-54,共6页
Urban Gas
关键词
有限空间
燃气作业
安全监管
AI图像识别
行为检测
激光气体检测
confined space
gas operation
safety supervision
AI-based image recognition
behavior detection
laser gas detection
