Assessment of model uncertainty based on model level and application method
The present study explores the influence of different numerical modelling approaches on the determination of model factors for global buckling, adopted in this work as a representative structural stability phenomenon. Beam, shell and solid element models are investigated through both deterministic analyses and stochastic Monte Carlo simulations. The performed simulations take into account the variability in material and geometrical properties, enabling a more realistic representation of the global buckling behaviour. Three reliability-based methods are employed to evaluate the corresponding model factors and to assess the influence of the modelling approach and the selected reliability-based method on the resulting design buckling resistance. The results indicate that the evaluated model factors are generally consistent across the different modelling approaches, suggesting a limited sensitivity to the adopted formulation of finite elements. However, this insensitivity holds true for the studied case, in which all three models can lead to consistent global buckling resistance, with local instabilities prevented. In contrast, the choice of the reliability-based method has a pronounced effect on the resulting design buckling resistance, providing a more improved estimation of the design resistance while reducing the conservative estimation, and maintaining consistency with the Eurocode standards. The provided results aid the designer in the process of selecting an appropriate modelling approach and reliability-based method for the assessment of the resistance.
Authors
- Mohammad Radwan (ORCID: https://orcid.org/0000-0002-8197-2754)
- Balázs Kövesdi
Institutions
- Budapest University of Technology and Economics (HU)
Publication Details
- Journal
- Scientific Reports
- Published
- 2026-11-20
- DOI
- https://doi.org/10.1038/s41598-026-79853-9
- Primary Topic
- Probabilistic and Robust Engineering Design
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Magyar Tudományos Akadémia