摘要
针对城市隧道中环境监测存在的监测方案不完善、监测数据量大、数据受影响因素多等问题,提出基于多传感器的隧道环境监测数据协同融合方法以提高环境监测的准确性。首先,使用箱线图法检测环境监测数据中的异常值,并基于均值替代法修正异常值,提高数据的精度;其次,利用自适应加权平均算法实现对同质传感器的数据融合,有效降低系统冗余;最后,基于PSO-BP神经网络算法实现对异质传感器的数据融合。结果表明,经过PSO算法优化的BP神经网络融合模型对隧道环境等级的判断准确性超过80%,优于传统BP神经网络,证实该方法可对隧道整体环境质量得出较可靠的评价,能为隧道内机电设施智能控制提供重要决策信息。
In this study,in view of the problems existing in environmental monitoring of urban tunnels,such as im⁃perfect monitoring schemes,large amount of monitoring data and many factors affecting the data,a collaborative fu⁃sion method for tunnel environmental monitoring data based on multi-sensor is proposed to improve the accuracy of environmental monitoring.Firstly,the box plot method is used to detect the abnormal values in the environmental da⁃ta,and the abnormal values are corrected with the mean substitution method,improving the accuracy of the data;Secondly,data fusion of homogeneous sensors is realized with the adaptive weighted average algorithm,reducing the system redundancy effectively;Finally,data fusion of heterogeneous sensors is realized based on the PSO-BP neural network algorithm.The results show that the accuracy of the BP neural network fusion model optimized by the PSO algorithm in judging the environmental grade of a tunnel is more than 80%,which is better than the traditional BP neural network.It is confirmed that the overall environmental quality of the tunnel can be evaluated more reliably with this method,which provides important decision information for the intelligent control of electromechanical facil⁃ities in the tunnel.
作者
米春
李思颖
牟佳祎
袁宵龙
李涛
MI Chun;LI Siying;MOU Jiayi;YUAN Xiaolong;LI Tao(Faculty of Geosciences and Environmental Engineering,Southwest Jiaotong University,Chengdu 611756)
出处
《现代隧道技术》
CSCD
北大核心
2023年第5期177-185,共9页
Modern Tunnelling Technology
基金
国家重点研发计划(2022YFC3801104)
四川省科技厅项目(2020YFH0045)
国家自然科学基金(52378412)。