Background: One of the novel methods for evaluating suspension bridges during and after earthquakes is the use of the DEMATEL method. This approach provides an effective framework for identifying the relationships and importance of influencing parameters in the seismic assessment of complex structural systems.
Methods: In this study, the seismic behavior of a suspension bridge equipped with a tuned mass damper (TMD) under far-field earthquake records (Case study: Golden Gate Bridge) was investigated using the DEMATEL method. A three-dimensional finite element model of the bridge was developed in ABAQUS software. Time-history analyses were performed using seven far-field earthquake records. First, modal analysis was conducted to extract the natural frequencies and mode shapes of the bridge. Subsequently, the time-dependent structural responses under earthquake excitations were evaluated.
Results: The results indicate that the main cables and towers have the greatest contribution to seismic energy absorption, while the flexibility of the suspension system reduces the transmission of seismic forces to the deck. Furthermore, it was demonstrated that applying nonlinear analysis in the modeling of suspension bridges can significantly improve the accuracy of seismic response predictions. The results also confirm the effectiveness of the tuned mass damper in reducing seismic displacements. The use of a TMD reduced the average permanent displacements developed in the bridge by approximately 35%.
Conclusion: Based on the proposed neural network approach, with an accuracy exceeding 90%, and using the DEMATEL method, it was determined that structural displacement and vibration modes are the most influential response parameters. In addition, foundation and cable stiffness, as well as the behavior of the tuned mass damper, were identified as the most effective influencing parameters. The long computational time and high complexity associated with modeling and analysis represent the main limitations of this research.behavior of the tuned mass damper, were identified as the most effective influencing parameters. The long computational time and high complexity associated with modeling and analysis represent the main limitations of this research.