November 19, 2025BuildingsOpen Access

Machine Learning Prediction on Progressive Collapse Resistance of Purely Welded Steel Frames Considering Weld Defects

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Machine learning predicts progressive collapse using finite element model and Latin Hypercube sampling, indicating RF as optimal algorithm.

The RF algorithm predicted progressive collapse resistance with an R2 of 0.988 and sMAPE of 5.096%.
Modeling involved a finite element framework and Latin Hypercube Sampling to generate 700 samples from design features.
Training included five algorithms, with RF identified as optimal through Bayesian optimization and K-fold validation.
The findings emphasize the impact of weld quality and beam height-to-span ratio on structural resistance in design.

Guo et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Machine learning predicts progressive collapse using finite element model and Latin Hypercube sampling, indicating RF as optimal algorithm.