Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Structural Engineering and Mechanics Volume 92, Number 1, October10 2024 , pages 65-79 DOI: https://doi.org/10.12989/sem.2024.92.1.065 |
|
|
|
Machine learning surrogate model for reliability analysis of RC columns with reverse curvature |
||
Arthur de C. Preuss and Herbert M. Gomes
|
||
| Abstract | ||
| This work aims to present an analysis of the structural reliability of reinforced concrete (RC) columns designed according to the general method outlined in Eurocode 2 (EN 1992-1-1 2004). Probabilistic analyses are conducted by integrating the Monte Carlo method with metamodels (or surrogate models) generated using Kriging and some machine learning techniques. The study was developed based on an algorithm that verifies the columns subject to biaxial bending, considering the physical and geometric nonlinearities. Columns were analyzed assuming sign inversion of end bending moments (with reverse curvature), which portray the typical situations in conventional structures of RC buildings. The probabilistic results reveal that the typical RC columns in buildings designed according to the design procedures of the studied standard, whether they are located at the center, corner, or edge, exhibit reliability levels surpassing those deemed acceptable within the technical community. Furthermore, the integration of surrogate models proves beneficial by alleviating the computational burden associated with evaluations while preserving accuracy. | ||
| Key Words | ||
| metamodels; RC columns; reliability analysis; reverse curvature; surrogate models | ||
| Address | ||
| Arthur de C. Preuss and Herbert M. Gomes: Graduate Program in Civil Engineering, Federal University of Rio Grande do Sul, Av. Osvaldo Aranha, 99, 3o. Andar, Porto Alegre, RS, Brazil | ||