The moving vehicle dynamic loads on a bridge deck are one of the most important live loads of bridges. They should be understood, monitored and controlled before the bridge design as well as when the bridge is open for traffic. This paper performs an ill-conditioning study on the identification of dynamic vehicle axle loads from bridge responses by numerical simulation and experiment investigation. Two identification methods, the time domain method (TDM) and the pre- treatment conjugate gradient method (PCGM), are employed and their identified loads are compared with the axle loads of a supposed moving vehicle. In order to prove the PCGM in practice, the experiment investigation is conducted by designing and making some bridge-vehicle models. The illustrated results show that PCGM have higher identification accuracy and robust noise immunity as well as producing an acceptable solution to ill-conditioning cases to some extent when they are used to identify the moving force from bridge responses.
展开▼