摘要
为了提高安全性,全世界都在寻求实施无线传感器网络(WSNs)来监测复杂的、动态的和环境恶劣的地下煤矿。文中引入了一种可靠的物联网(IoT)空气质量监测系统,该系统由传感器模块、通信协议和基站组成。基于STM32的传感器模块具有八个不同的参数,安装在可操作的地下煤矿的不同位置。基于感知数据,该系统用煤矿环境指数(MEI)对地下煤矿矿井空气质量进行评价。采用主成分分析法确定了CH 4、CO、SO 2和H 2 S是影响矿井空气质量最主要的气体。将主成分分析的结果输入到RNN神经网络模型中,实现了MEI的预测。结果表明,基于主成分分析的神经网络在MEI预测中具有较好的性能,主成分分析+RNN预测模型的性能指标R 2和RMSE值分别为0.4890和0.1204,提高了线性回归模型对矿井大气污染物的预测精度。因此,提出的基于STM32和Tensorflow平台的人工神经网络可以快速评估和预测矿井空气质量,提高矿井环境安全性。
In order to improve security,the world is seeking to implement wireless sensor network(WSNs)to monitor complex,dynamic and harsh underground coal mines.We introduce a reliable Internet of Things(IoT)air quality monitoring system which is composed of sensor module,communication protocol and base station.The sensor module based on STM32 has eight different parameters and is installed in different positions of operable underground coal mine.This system evaluates the air quality of underground coal mines with the coal mine environmental index(MEI).The principal component analysis method is used to determine that CH 4,CO,SO 2 and H 2 S are the main gases affecting mine air quality.The system inputs the results of principal component analysis into the RNN neural network model and realizes the prediction of MEI.The results show that the neural network based on principal component analysis has better performance in MEI prediction.Principal component analysis combined with RNN prediction model method reduces the error.Its error indices R 2 and RMSE are 0.4890 and 0.1204 respectively,which improve the prediction accuracy of linear regression model for mine air pollutants.Therefore,the proposed neural network based on STM32 and Tensorflow platform can quickly evaluate and predict mine air quality and improve mine environmental safety.
作者
李湘文
周辅杰
崔崴
邓琴秀
张辉雨
LI Xiang-wen;ZHOU Fu-jie;CUI Wei;DENG Qin-xiu;ZHANG Hui-yu(Engineering&Technical College,Chengdu University of Technology,Leshan 614000,China)
出处
《计算机技术与发展》
2020年第8期115-119,共5页
Computer Technology and Development
基金
四川省科技计划项目(2019JDKP0020)
四川省教育重点科研项目(18ZA0066,18ZA0068)
乐山市科技计划项目(18JCXY011)
成理工程科研基金项目(C122016002,C122018009,C122017021)
教育部高等教育司产学合作协同育人项目(201802022027,201802047128)。
