Onsite Identification of Moving Vehicle Loads on Multispan Continuous Bridge Using Both Dictionary Expansion and Sparse Regularization

Moving force identification (MFI) plays an important role in monitoring moving vehicle loads acting on a bridge, which has attracted great attention in recent years. Although many MFI methods have been proposed, their effectiveness in real bridges has not been carefully investigated. In this study, focusing on how to identify moving forces from structural responses in a real bridge, some important aspects are conducted and discussed in detail. The effectiveness of a MFI method based on dictionary expansion and sparse regularization is evaluated by a five-span continuous bridge in real engineering. First, a finite-element (1-b) model is established to approximately describe the real bridge. Then, the FE model is updated based on the modal information obtained from structural responses, so the transfer matrices between moving forces and structural responses can be calculated. Meanwhile, data preprocessing is conducted and a suitable dictionary is designed to obtain reasonable MFI results. As a result, the MFI equation by adding reweighted l(1)-norm regularization is used for MFI. Finally, two trucks with known gross vehicle weights are adopted to evaluate its effectiveness. The NM results show that this method can reasonably estimate the moving vehicle loads acting on bridge deck, and the identified moving forces are reasonable. (C) 2021 American Society of Civil Engineers.