Explainable boosting machines for web-crippling resistance prediction of cold-formed sigma sections under end-two-flange loading
Author(s)
Simwanda, Lenganji
Kannan, Thushanthan
Thirunavukkarasu, Kajaharan
Sifan, Mohamed
Poologanathan, Keerthan
Type
Journal Article
Abstract
Cold-formed steel (CFS) sigma sections are highly efficient but vulnerable to web crippling under concentrated bearing loads. Current design provisions in AISI S100, AS/NZS 4600 and EN 1993-1-3 are largely calibrated for channel and Z-sections and often give conservative or scattered predictions for sigma sections under end–two–flange (ETF) loading. This study presents an explainable machine-learning framework for predicting ETF web-crippling resistance of sigma sections. A validated nonlinear finite-element study in ABAQUS produced 756 simulations covering variations in geometry, yield strength and bearing conditions. Finite-element predictions were verified against experimental data for carbon steel, stainless steel and aluminium, showing mean Test/FE ratios of 0.95–1.01 and coefficients of variation below 5%. Six ensemble models were trained using Bayesian optimisation, with boosting-based methods achieving the best accuracy (R2 ≈ 1.0; RMSE < 0.1 kN) and outperforming code equations. SHAP analysis identified thickness, yield strength and bearing length as dominant predictors. The proposed models provide fast, accurate and transparent tools for ETF web-crippling resistance estimation.
Date Issued
2026-09-01
Date Acceptance
2026-09-01
Citation
ce/papers, 2026, 9 (2-3), pp.2997-3002
ISSN
2509-7075
Publisher
Wiley
Start Page
2997
End Page
3002
Journal / Book Title
ce/papers
Volume
9
Issue
2-3
Copyright Statement
© 2026 The Author(s). ce/papers published by John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes
License URL
Publication Status
Published
Date Publish Online
2026-09-14