A general solution to the identification of moving vehicle forces on a bridge

Bridge weigh-in-motion systems measure bridge strain caused by the passing of a truck to estimate static axle weights. For this calculation, they commonly use a static algorithm that takes the bridge influence line as reference. Such a technique relies on adequate filtering to remove bridge dynamics and noise. However, filtering can lead to the loss of a significant component of the underlying signal in bridges where the vibration does not have time to complete sufficient number of cycles and in cases of closely spaced axles traveling at high vehicle speeds. In order to overcome these limitations and also to provide additional information on the dynamics of the applied forces, this paper presents an algorithm based on first-order Tikhonov regularization and dynamic programming. First, strain measurements are simulated using an elaborate three-dimensional vehicle and orthotropic bridge interaction system. Then, strain is contaminated with noise and input into the moving force identification algorithm. The procedure to implement the algorithm and to derive the applied forces from the simulated strain record is described. Vehicle axle forces are shown to be accurately predicted for smooth and rough road profiles and a range of speeds. Copyright (C) 2008 John Wiley & Sons, Ltd.