Answers are grounded in this paper and cite their sources.
Demonstrates improved identification of dominant failure modes in long-span arch bridges, suggesting enhanced safety and efficiency.
The identification of dominant failure modes (DFMs) of long-span arch bridges is crucial for structural design optimization and safety status assessment. However, existing methods suffer from heavy computational burden when dealing with large-scale systems like arch bridges. To address this issue, a method for identifying DFMs of long-span arch bridges based on deep neural networks (DNN) is proposed. First, Monte Carlo simulation is employed to calculate failed components under specific structural states, load and resistance parameters. A sample library of input parameters and failed components meeting training requirements is thereby constructed to minimize the number of calls to the finite element model. Then, a DNN model is established to capture the nonlinear mapping relationship between input parameters and failed components. Based on the trained model, efficient identification of DFMs of long-span arch bridges is realized through step-by-step prediction of failed components. Finally, a numerical case study of a long-span arch bridge is conducted. The DFMs of the structure under different scenarios are analyzed, and the effectiveness of the proposed method is verified. The results show that under the same computation accuracy as traditional methods, the proposed method has higher computation efficiency, saving 70% of the calculation time. Moreover, the identification results of DFMs are in good agreement with on-site disease detection.
No takes yet. Share an insight, caveat, or question.
Tang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: