基于桥梁挠度的重载车辆荷载识别方法研究  

Research on Heavy Vehicle Load Identification Method Based on Bridge Deflection

作　　者：谷云秋 钱江 江科 胡洁亮 魏哲 Gu Yunqiu;Qian Jiang;Jiang Ke;Hu Jieliang;Wei Zhe(Ningbo Highway and Transportation Management Center,Ningbo 315100,China;Beilun District Highway and Transportation Management Center,Ningbo 315800,China;Ningbo Langda Technology Co.,Ltd.,Ningbo 315100,China)

机构地区：[1]宁波市公路与运输管理中心,浙江宁波315100 [2]宁波市北仑区公路与运输管理中心,浙江宁波315800 [3]宁波朗达科技有限公司,浙江宁波315100

出　　处：《市政技术》2024年第8期23-29,190,共8页Journal of Municipal Technology

基　　金：科技创新2025重大专项(宁波市重大科技任务攻关项目)(2022Z227)。

摘　　要：以浙江省宁波市某桥梁工程为依托,提出了一种基于双向长短时记忆网络(Bi LSTM)的重载车辆荷载识别方法,通过自主学习并提取结构动挠度时程特征,构建了车辆车重与结构动挠度之间的映射关系,实现了重载车辆荷载参数的反演,验证了所提方法具有很好的鲁棒性与泛化性能,车重识别精度最高可达91%,对多车行驶的抗干扰性也较强。相关研究结论可为类似重载车辆荷载识别提供参考。Based on a bridge project of Ningbo City,Zhejiang Province,a heavy vehicle load identification method was proposed based on bi-directional long short-term memory network(BiLSTM).The time-dependent features of structural dynamic deflection were learned and extracted and the mapping relationship between vehicle weight and structural dynamic deflection was constructed to realize the inversion of heavy vehicle loading parameters.The proposed method is verified to have good robustness and generalization performance.The vehicle weight recognition accuracy can reach up to 91%and the anti-interference ability for multi-vehicle driving is stronger.The relevant conclusions can provide reference for load identification of similar heavy vehicles.

关 键 词：重载车辆 荷载识别 桥梁挠度 动态称重系统 长短时记忆网络 

分 类 号：U441.2[建筑科学—桥梁与隧道工程]