This paper presents a novel damage identification technique for bridges based on the dynamic response of a moving vehicle. A quarter car vehicle model instrumented with two accelerometers and an inertial profilometer is used for this purpose. The method involves the coupling of Tikhonov regularization scheme with signal averaging technique to handle the problem of measurement noise and road roughness. In the first stage, a damage‐dependent road roughness profile is estimated from measured acceleration response by minimizing a Tikhonov regularized least squares cost function. The second stage involves the minimization of the profile roughness residual function that depends on the location and magnitude of damage. This objective function compares the roughness profile computed from the first stage and that measured by the inertial profilometer. It is proved that efficient damage detection ensues from considering multiple runs of the vehicle and choosing the appropriate regularization parameter. The present study considers various aspects, such as the uniqueness of results, robustness of results to measurement noise, effect of modelling error, effect of vehicle speed, and uncertainty in estimation. Numerical results show that the approach is capable of detecting the magnitude as well as the location of damage. In addition, the problem of road roughness and measurement noise is handled competently.

中文翻译：

---

使用Tikhonov正则化方案和信号平均技术的桥梁间接健康监测

本文提出了一种基于移动车辆动态响应的桥梁损伤识别新技术。为此，使用了配备有两个加速度计和一个惯性轮廓仪的四分之一汽车模型。该方法涉及将Tikhonov正则化方案与信号平均技术相结合，以解决测量噪声和道路不平整问题。在第一阶段，通过最小化Tikhonov正则化最小二乘成本函数，从测得的加速度响应中估算出与损坏有关的路面粗糙度曲线。第二阶段涉及最小化轮廓粗糙度残留函数，该函数取决于损坏的位置和大小。该目标函数比较从第一阶段计算出的粗糙度轮廓和由惯性轮廓仪测量的粗糙度轮廓。事实证明，通过考虑车辆的多次行驶并选择适当的正则化参数，可以实现有效的损坏检测。本研究考虑了各个方面，例如结果的唯一性，结果对测量噪声的鲁棒性，建模误差的影响，车速的影响以及估计的不确定性。数值结果表明，该方法能够检测出损伤的大小和位置。此外，道路粗糙度和测量噪声的问题也得到了妥善处理。建模误差的影响，车速的影响以及估计的不确定性。数值结果表明，该方法能够检测出损伤的大小和位置。此外，道路粗糙度和测量噪声的问题也得到了妥善处理。建模误差的影响，车速的影响以及估计的不确定性。数值结果表明，该方法能够检测出损伤的大小和位置。此外，道路粗糙度和测量噪声的问题也得到了妥善处理。