基于多目标动态跟踪的桥梁动态称重方法研究与试验验证

Overloaded vehicles have become the main reason for the fatigue damage and failure and the reduction of fatigue life of bridge structures. Developing bridge weigh-in-motion (BWIM) techniques that can fast and accurately identify the weight of moving heavy vehicles is critical for managing overloaded vehicles and bridge safety forecast. However, the difficulty in accurately locating the positions of moving vehicles and separating the multi-source bridge response signals has led to two drawbacks of the current BWIM systems, i.e., low accuracy in multi-vehicle identification and limited bridge types and bridge span length range for application. To solve these problems, this research project aims to develop a novel BWIM method based on the multi-target dynamic tracking technique. To handle the diversity of vehicle characteristics and the uncertainty of vehicle movement, a multi-target identification and tracking algorithm based on self-adapting technique in dynamic background will be developed and used to accurately identify the real-time positions of heavy vehicles on the bridge. Methods for separating bridge responses from different sources and de-noising will be developed. A high-accuracy two-dimensional identification method will be developed for identifying the axle weights of heavy vehicles. The accuracy and reliability of proposed method will be verified through tests. The research results from this project will solve the bottleneck of traditional BWIM methods under complicated cases of multi-vehicle identification and with uncertainties in vehicle movement tracking, expand the types of bridge and span length range for application, and provide a feasible technique for monitoring and managing overloaded vehicles in China.

车辆超载运输已成为桥梁结构疲劳损伤破坏和运营寿命缩短的主要原因。开发能准确快速识别移动重车重量的桥梁动态称重(BWIM)技术对于加强超载车辆监控和桥梁安全预警具有重要意义。然而，移动车辆精确定位和多源桥梁响应分离的难题使当前的BWIM方法存在两大缺陷：多车工况下识别精度低、适用桥梁类型与跨径范围小。为此，本项目拟开发基于多目标动态跟踪技术的新型BWIM方法。针对车辆特征多样性及其运动轨迹不确定性，提出动态背景下特征自适应的多目标同步识别与跟踪方法，准确确定桥上重车的实时位置；研究复杂多源激励下桥梁响应的去噪和分离算法；开发基于二维高精度影响面的轴重识别算法准确计算各辆重车的轴重；最后通过试验验证方法的识别精度和可靠性。研究成果将突破传统BWIM方法在多车同时过桥、车辆行驶轨迹不确定等复杂工况下轴重识别精度低的瓶颈，拓宽BWIM系统的适用桥型和跨径范围，为我国超载车辆监控提供可靠的技术手段。

车辆荷载是公路关键枢纽——桥梁上的主要活载，随着交通流量的持续增长，监测桥上移动车辆行驶轨迹和重量对于桥梁的运营安全、维护管养具有重要的意义。桥梁动态称重（BWIM）方法可以在车辆行驶状态下测量出车辆轴重和总重，无需在桥面安装传感器，具有安装维护不中断交通不破坏路面、使用寿命长的优点。然而由于桥面往往会同时存在多个车辆，复杂的多车—桥耦合作用使得各车辆作用效应的分离存在困难；同时大交通流情况下还存在车辆频繁变速的情况。这些问题导致传统BWIM方法难以获得较高的精度。本项目针对桥梁动态称重方法在大交通流桥梁上的应用难点，结合有限元分析、模型试验和实桥监测，研究开发高精度二维BWIM方法，以解决桥上多车辆随机行驶情况下的多车位置与重量同步识别问题。.本项目的主要研究内容以及研究成果包括以下几个方面：（1）研究提出了用于公路车桥耦合振动数值分析的精细化的三维轮胎数值模型。提出了接触面不均匀形变的计算方法，为更准确地模拟复杂工况（桥头跳车、经过伸缩缝、经过明显凹凸激励）下的车桥动力相互作用提供了可能。（2）研究了桥梁动态称重方法，提出了新颖的桥梁动态称重虚拟车轴（VA）算法。本方法利用桥梁的整体弯曲应变（即称重传感器信号）同时实现了车轴检测和轴重称量，避免了传统BWIM方案对专用车轴探测器的依赖。（3）研究了人工智能方法在BWIM工程中的应用，创新性地提出了半监督学习的BWIM网络框架BwimNet算法，解决了人工智能算法在BWIM领域的应用时最困难的数据集获取问题。（4）研究了全桥面车辆荷载跟踪与识别问题，提出了基于双目视觉和车轴位置的桥面车辆追踪与荷载识别方法。有效解决了全桥面多车辆跟车、并行、变速等复杂工况下的车重识别问题。（5）研究了车辆轴限的合理取值方法。结合桥梁可靠指标变化、桥梁剩余使用寿命及其年均加固费用，建立了轴限值设定曲线。该研究开创了综合考虑考虑公路交通基础设施安全和区域综合经济效益的车辆限载研究先河。

内容获取失败，请点击重试