CNN-LSTM在桥梁预警机制的研究与应用  

STUDY AND APPLICATION OF CNN-LSTM IN BRIDGE HEALTH MONITORING AND EARLY WARNING MECHANISM

作　　者：潘浩 李富年[1] 余兴盛 秦寰宇[2] 陈志丹 Pan Hao;Li Funian;Yu Xingsheng;Qin Huanyu;Chen Zhidan(College of Information Science and Engineering,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China;China Railway Siyuan Survey and Design Group Co.,Ltd.,Wuhan 430081,Hubei,China;Huazhong University of Science and Technology,Wuhan 430081,Hubei,China)

机构地区：[1]武汉科技大学信息科学与工程学院,湖北武汉430081 [2]中铁第四勘察设计院集团有限公司,湖北武汉430081 [3]华中科技大学,湖北武汉430081

出　　处：《计算机应用与软件》2025年第3期29-33,40,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(51778258,51922046);中铁第四勘察设计院集团有限公司科研项目(2020K006,2019D001,2020D002);中铁二院工程集团有限责任公司科研项目(KYY2019029(19-21))。

摘　　要：为了更好地满足桥梁健康监测系统的需求,提高桥梁健康监测系统的性能,将深度学习与时序数据库InfluxDB结合起来构建新型桥梁健康监测系统的预警机制,提高现代桥梁健康监测系统的危险感知能力。以赣江特大桥为背景,将卷积神经网络CNN与长短时记忆网络LSTM结合起来构建CNN-LSTM模型,对桥梁的挠度数据进行预测。通过对实验结果分析发现CNN-LSTM模型能够有效预测出桥梁的挠度数据,在置信区间为±0.1 mm的情况下,准确率达到92.8%,在预测未来十分钟的挠度数据中,均方根误差RMSE为0.1097。实践表明时序数据库InfluxDB与CNN-LSTM模型的融合增强桥梁健康监测系统对潜在威胁的感知能力,有效提高桥梁健康监测系统的预警报警机制。In order to better meet the needs of the bridge health monitoring system and improve the performance of the bridge health monitoring system,this paper combines deep learning with the time series database InfluxDB to build a new early warning mechanism for the bridge health monitoring system and improve the risk perception ability of the modern bridge health monitoring system.This paper took the Ganjiang River Bridge as the background and combined the neural network CNN with the long short-term memory network to construct a CNN-LSTM model to predict the deflection data of the bridge.Through the analysis of the experimental results,it is found that the CNN-LSTM model can effectively predict the deflection data of the bridge.When the confidence interval is±0.1 mm,the accuracy rate reaches 92.8%.In predicting the deflection data of the next ten minutes,the root mean square error(RMSE)is 0.1097.Practice shows that the fusion of the time series database InfluxDB and the CNN-LSTM model enhances the bridge health monitoring system's perception of potential threats,and effectively improves the early warning and alarm mechanism of the bridge health monitoring system.

关 键 词：桥梁工程 长短时记忆网络 卷积神经网络 CNN-LSTM模型 InfluxDB 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]