Search

Concrete-filled steel tubular (CFST) columns are widely utilized in structures owing to their cost-effectiveness, high strength, ductility, and proven fire resistance. Despite these advantages, concerns persist regarding the performance of CFST columns with locally corroded steel tubes, impeding their widespread adoption in offshore structures. This research employs three-dimensional finite element (FE) simulations for circular CFST stub columns subjected to axial compression, validated against experimental data. Verification includes full load-displacement histories, ultimate axial strengths, and failure modes. The validated FE models are then employed to assess and compare the structural behavior of CFST columns with various shapes of localized corrosion, analyzing overall load-deformation curves, interaction stress-deformation responses, and composite actions. To acquire critical data for understanding stub column behavior, 81 specimens underwent non-linear finite element analysis, with local corrosion represented by artificial notches in steel tubes. Test parameters, such as notch orientations, lengths, depths, widths, and concrete compressive strength, are varied. Results elucidate the impacts of these parameters on different aspects of stub column responses with corroded steel tubes. Furthermore, the accuracy of existing design models in predicting the ultimate axial strengths of CFST columns with corroded steel tubes is assessed. Finally, through parametric analysis of test results, a practical model is proposed and verified for computing the peak resistance of a CFST stub column with a locally corroded steel tube. • Finite-element simulation model for circular CFST stub columns with locally corroded steel tubes is developed. • The influence of different shapes of localized corrosion on the performance of circular CFST stub columns is investigated. • The behavior of circular CFST stub columns with locally corroded steel tubes is parametrically studied. • The proposed finite-element simulation and design models provide accurate predictions. [ABSTRACT FROM AUTHOR]