Abstract
Bridge life cycle management is a collaborative work process involving multiple groups, and finite element analysis is a popular method used in bridge management. The frequent establishment of finite element models by diverse groups directly reduces the efficiency and quality of bridge management. To address this issue, a large model of bridge model transformation was developed by integrating domain knowledge based on existing general large models, and data transformation and data recombination were proposed. Simulation investigation demonstrates that the proposed bridge modeling approach may generate finite element models that fulfill the various needs of participating organizations based on their specific criteria. The domain knowledge base supports the data transformation algorithm, which calls the Pangu Large Model's scientific computing model to enable autonomous data translation, minimizing the amount of time spent manually participating in the modeling process. The seismic study of the bridge reveals that the finite element model generated using data recombination differs by 5.93% from the analytical results of the manually established finite element model. Overall, data transformation and recombination modeling methods have greatly increased bridge modeling efficiency, reducing the need for duplicate modeling and improving bridge management efficiency.