Akademik Veri Yönetim Sistemi
Advances in Civil Engineering, cilt.2026, sa.1, 2026 (SCI-Expanded, Scopus)
This study quantifies how beam-to-column joint damage alters the axial resistance and failure modes of cold-formed steel (CFS) uprights used in industrial storage rack systems. Current design-oriented assessments typically neglect the reduction of joint rotational stiffness and moment resistance caused by prior cyclic joint deterioration, which can compromise subsequent compression performance. To address this gap, six commercial perforated upright sections are tested under axial compression in intact and joint-damaged conditions. A unified framework combines experiments, nonlinear finite element analysis (FEA; ANSYS), and machine learning (ML) prediction using geometric descriptors (thickness, area, second moments of area, and radii of gyration) together with experimentally derived joint properties (rotational stiffness and ultimate joint moment). Among the evaluated regressors, the gradient boosting model provides the best predictive performance (cross-validated R2 and lowest root mean square error [RMSE]/mean absolute error [MAE]), showing close agreement with experimental outcomes and numerical trends. Results indicate axial capacity reductions of up to ~30% after joint degradation and a tendency toward coupled distortional and torsional–flexural (TF) instability. The findings highlight the engineering need to account for joint stiffness degradation in residual capacity checks and support rapid, data-informed decisions for inspection and retrofit of rack uprights. The study provides a practical predictive framework for postdamage capacity evaluation in industrial rack systems.