摘要
随着城市轨道交通建设规模扩大,施工环境复杂化导致的危险源动态识别与智能预警成为行业重点关注的课题。针对重庆轨道交通18号线北延工程,提出基于“感知层-数据处理层-应用层”架构的智慧工地危险源动态识别与智能预警系统。通过GIS与多Agent协同系统、改进YOLOv5模型及三级预警机制,实现CH4泄漏毫秒级响应、多源数据交叉验证及断网数据高可靠补传。测试表明,系统在30天压力测试中,CPU/内存占用率稳定低于70%,重大危险源预警覆盖率达100%,事故隐患响应时间缩短75%,为《智慧工地建设与评价标准》(DBJ50/T—356—2020)三星级达标提供了关键技术支撑。
The expansion of urban rail transit construction scale and the complexity of construction environment have led to dynamic identification and intelligent warning of hazards becoming industry pain points.A smart construction site hazard dynamic identification and intelligent warning system based on the"perception layer data processing layer application layer"architecture is proposed for the northern extension project of Chongqing Rail Transit Line 18.Through GIS and multi-agent collaborative system,improved YOLOv5 model and three-level warning mechanism,CH4 leakage millisecond level response,multi-source data cross validation,and highly reliable transmission of disconnected data are achieved.Tests have shown that the CPU/memory usage rate of the system remained stable below 70%during the 30 day stress test,the coverage rate of major hazard warning reached 100%,and the response time to accident hazards was shortened by 75%.This provides key technical support for achieving the three-star standard of"Smart Construction Site Construction and Evaluation Standards"(DBJ50/T-356-2020).
作者
贾皞杰
Jia Haojie(China Communications Construction Corporation Rail Transit Branch,Beijing 100088)
出处
《现代工程科技》
2025年第10期173-176,共4页
Modern Engineering Technology
关键词
智慧工地平台
危险源动态识别
智能预警系统
Smart construction site platform
Dynamic identification of hazards
Intelligent warning system
