电梯机房环境监控系统设计及方法研究  

Research on design and methodologies for environmental monitoring system in elevator machine room

作　　者：倪敏敏 路成龙 庆光蔚 Ni Minmin;Lu Chenglong;Qing Guangwei

机构地区：[1]南京市特种设备安全监督检验研究院,南京210002

出　　处：《起重运输机械》2025年第3期44-51,共8页Hoisting and Conveying Machinery

基　　金：国家市场监督管理总局科技计划项目(2022MK156)。

摘　　要：根据GB/T 10058—2023《电梯技术条件》及TSG T7001—2023《电梯监督检验和定期检验规则》中对电梯机房环境的要求,文中设计了一套机房环境检测系统,能够对机房内温湿度、电压、噪声以及火焰烟雾信息进行监测。通过CR3X三相交流电参数模块采集电压信息,SN-ZS-BZ板载噪声模块采集机房主机噪声信息,DS18B20模块采集温湿度信息,通过K210视觉识别模块采集火焰烟雾视频信息,通过植入改进后的YOLOv8算法进行火焰烟雾视觉识别。采集后的信息送入单片机控制系统进行处理,处理后的数据通过无线通信EC800M-CN模块送入远程监控系统。经验证,温湿度、噪声、电压测量能够保持较高的检测精度,同时改进后的算法mAP0.5达到了83.2%,mAP0.5:0.95达到了51.2%,相对于原算法分别提高了2.8%和2.7%,验证了监控系统的有效性。In compliance with national standards and inspection regulations for elevator machine room environments,a machine room environment detection system was developed.This system is capable of continuously monitoring critical parameters such as temperature,humidity,voltage,noise,and flame smoke within the machine room.The voltage data is acquired using the CR3X three-phase alternating current parameter module,while the noise levels of the computer room host are captured by the SN-ZS-BZ onboard noise module.Temperature and humidity readings are obtained through the DS18B20 module.For visual detection of flame and smoke,the K210 visual recognition module is used,and it incorporates an enhanced YOLOv8 algorithm for improved accuracy in visual recognition tasks.The gathered data is transmitted to a single-chip microcontroller-based control system for processing.Subsequently,the processed information is relayed to a remote monitoring system via the wireless communication EC800M-CN module.It is verified that the system maintains high detection accuracy for temperature,humidity,noise,and voltage measurements.The enhanced algorithm achieves an mAP0.5 of 83.2%and an mAP0.5:0.95 of 51.2%,representing improvements of 2.8%and 2.7%over the original algorithm,respectively.These results substantiate the effectiveness of the monitoring system.

关 键 词：机房环境 STM32 YOLOv8 视觉识别 算法改进 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TH166[自动化与计算机技术—计算机科学与技术]