Physics-based models, surrogate models and experimental assessment of the vehicle-bridge interaction in braking conditions

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@Article{ALOISIO:2023:ymssp,

• author = "Angelo Aloisio and Alessandro Contento and Rocco Alaggio and Giuseppe Quaranta",
• title = "Physics-based models, surrogate models and experimental assessment of the vehicle-bridge interaction in braking conditions",
• journal = "Mechanical Systems and Signal Processing",
• volume = "194",
• pages = "110276",
• year = "2023",
• ISSN = "0888-3270",
• DOI = "doi:10.1016/j.ymssp.2023.110276",
• URL = "https://www.sciencedirect.com/science/article/pii/S0888327023001838",
• keywords = "genetic algorithms, genetic programming, Bouncing, Braking, Bridge, Fragility curve, Machine learning, Moving load, Neural network, ANN, Pitching, Roughness, Surrogate model, Vehicle-bridge interaction",
• abstract = "The dynamics of roadway bridges crossed by vehicles moving at variable speed has attracted far less attention than that generated by vehicles travelling at constant velocity. Consequently, the role of some parameters and the combination thereof, as well as influence and accuracy of the modelling strategies, are not fully understood yet. Therefore, a large statistical analysis is performed in the present study to provide novel insights into the dynamic vehicle-bridge interaction (VBI) in braking conditions. To this end, an existing mid-span prestressed concrete bridge is selected as case study. First, several numerical simulations are performed considering alternative vehicle models (i.e., single and two degrees-of-freedom models) and different braking scenarios (i.e., soft and hard braking conditions, with both stationary and nonstationary road roughness models in case of soft braking). The statistical appraisal of the obtained results unfolds some effects of the dynamic VBI modelling in braking conditions that have not been reported in previous studies. Additionally, the use of machine learning techniques is explored for the first time to develop surrogate models able to predict the effect of the dynamic VBI in braking conditions efficiently. These surrogate models are then employed to obtain the fragility curve for the selected prestressed concrete bridge, where the attainment of the decompression moment is considered as relevant limit state. Whilst the derivation of the fragility curve using numerical simulations turned out to be almost unpractical using standard computational resources, the proposed approach that exploits surrogate models carried out via machine learning techniques was demonstrated accurate despite the dramatic reduction of the total elaboration time. Finally, the accuracy of the numerical (physics-based and surrogate) models is evaluated on a statistical basis through comparisons with experimental data",

}

Genetic Programming entries for Angelo Aloisio Alessandro Contento Rocco Alaggio Giuseppe Quaranta

Citations