Identification of moving vehicle forces on bridge structures via moving average Tikhonov regularization

Abstract

Traffic-induced moving force identification (MFI) is a typical inverse problem in the field of bridge structural health monitoring. Lots of regularization-based methods have been proposed for MFI. However, the MFI accuracy obtained from the existing methods is low when the moving forces enter into and exit a bridge deck due to low sensitivity of structural responses to the forces at these zones. To overcome this shortcoming, a novel moving average Tikhonov regularization method is proposed for MFI by combining with the moving average concepts. Firstly, the bridge-vehicle interaction moving force is assumed as a discrete finite signal with stable average value (DFS-SAV). Secondly, the reasonable signal feature of DFS-SAV is quantified and introduced for improving the penalty function (||x|| 2 2 ) defined in the classical Tikhonov regularization. Then, a feasible two-step strategy is proposed for selecting regularization parameter and balance coefficient defined in the improved penalty function. Finally, both numerical simulations on a simply-supported beam and laboratory experiments on a hollow tube beam are performed for assessing the accuracy and the feasibility of the proposed method. The illustrated results show that the moving forces can be accurately identified with a strong robustness. Some related issues, such as selection of moving window length, effect of different penalty functions, and effect of different car speeds, are discussed as well.

A revised time domain force identification method based on Bayesian formulation

“…In the 2% noise case, the amplitude drop of TRI happens at both the entrance and exit moment. This observation is consistent with those in the works of Nordström and Pan et al When the NL becomes larger, the results of TRI become even more inaccurate. In subfigure (D), where the NL is high (8%), the TRI result does not even center around the true vehicle weight.…”

“…It will be shown in this paper that the traditional regularization techniques, from a Bayesian perspective, is equivalent to applying a zero‐mean prior distribution on the force history, which is sometimes inconsistent with real‐life situations. A few examples are the wind loads on civil structures, the cutting force on cutting tools, and the vehicle loads on bridges . In these cases, the force history is consisted of a trend component and dynamic components.…”

Generalized algorithms for the identification of seismic ground excitations to building structures based on generalized Kalman filtering under unknown input

“…To reduce the influence of measurement noises in using the Newmark-β method for estimating seismic ground acceleration, smooth technique (Pan et al, 2017) is applied to the estimated unknown seismic acceleration by equation (30) as follows…”

Traffic Sensing Methodology Combining Influence Line Theory and Computer Vision Techniques for Girder Bridges

“…All these advantages have made BWIM a preferable tool to weigh vehicles, especially heavy trucks, attracting many follow-up research and engineering applications. Up to date, this research topic has progressed significantly in aspects as diverse as the identification results, such as time-history moving load identification [9][10][11], or the types of sensors, such as portable accelerometers [12].…”