Application of Two Concrete Models in Finite Element Analysis of Concrete Filled Steel Tube Columns

Literature Overview

This paper by Chen Xi and Zhou Deyuan from Tongji University investigates the applicability of two constitutive models for concrete in the finite element analysis of concrete filled steel tube (CFST) short columns subjected to axial compression. The study compares the Smeared Cracking model and the Damaged Plasticity model, evaluating their performance through both constrained and unconstrained tie configurations for square and rectangular CFST members. The authors benchmark their numerical predictions against existing experimental data to assess the accuracy and limitations of each model.

Core Technical Content

The Smeared Cracking model relies on the smeared crack approach to represent concrete cracking behavior in a continuum framework, whereas the Damaged Plasticity model employs a damage-plasticity framework that combines a scalar damage variable with plasticity theory. A critical distinction lies in the flow rule adopted: the Smeared Cracking model employs an associated flow rule, which inherently links the plastic potential function to the yield surface, while the Damaged Plasticity model uses a non-associated flow rule, allowing independent definition of the yield surface and plastic potential.

Constitutive Model Comparison

Technical Interpretation

The associated flow rule in the Smeared Cracking model forces the plastic strain increment to be normal to the yield surface, which means the dilation angle of confined concrete is implicitly determined by the geometry of the yield surface. In practice, confined concrete under multi-axial compression exhibits a dilation angle that depends on the confining pressure, the triaxial stress state, and the material properties. The associated flow rule cannot adequately represent this behavior, leading to an underestimation of lateral expansion and, consequently, the confining pressure exerted by the steel tube on the core concrete.

The Damaged Plasticity model, by contrast, decouples the yield surface from the plastic potential function through the non-associated flow rule. This allows the engineer to independently calibrate the plastic potential to reflect the actual dilation behavior of confined concrete. The result is a more accurate representation of the steel-concrete interaction, which is the fundamental mechanism governing the enhanced load-bearing capacity of CFST columns.

Engineering Practice Implications

From a welding and pipe fabrication standpoint, the steel tube in a CFST column is typically manufactured from ERW or HFW welded pipe, or seamless pipe for critical applications. The welding quality of the steel tube directly influences the confinement effectiveness. A poorly welded longitudinal seam can act as a stress concentrator and reduce the effective confining pressure, particularly at the weld toe where micro-cracks may initiate under cyclic or sustained loading.

For engineers selecting a constitutive model for CFST member analysis, the following recommendations emerge:

1. For preliminary design and parametric studies where computational efficiency is paramount, the Smeared Cracking model may be acceptable, provided that the predicted capacities are conservatively interpreted.
2. For detailed design of critical structures, particularly those involving rectangular CFST columns where the corner confinement effect is significant, the Damaged Plasticity model is strongly preferred.
3. When using either model, the material parameters should be calibrated against local concrete test data, as the default parameters in commercial finite element software are often based on standard concrete rather than the specific mix design used in the project.

Key Questions and Reflections

A question that arises from this study is whether the observed differences between the two models would persist for circular CFST columns, where the confinement effect is more uniform and the dilation behavior may be less sensitive to the flow rule. The authors focused exclusively on square and rectangular sections, which are common in building construction but represent a more complex stress state than circular sections.

Additionally, the study does not address the effect of welding-induced residual stresses on the numerical predictions. In practice, longitudinal welded tubes carry significant residual stresses near the weld seam, which can alter the local stress distribution and potentially affect the steel-concrete interaction. Incorporating these residual stresses into the finite element model would provide a more realistic assessment of the member behavior, particularly under combined loading conditions.

Study Insights and Conclusion

This paper provides a valuable methodological comparison for engineers working with CFST structures. The key insight is that the choice of flow rule is not merely a numerical detail but has direct physical consequences on the predicted confinement effect and load-bearing capacity. The Damaged Plasticity model's superior agreement with experimental results validates the use of non-associated flow rules for confined concrete analysis. For steel pipe manufacturers and welding engineers, this reinforces the importance of maintaining high welding quality in CFST applications, as the numerical models that accurately predict member behavior rely on the assumption of effective steel tube confinement, which in turn depends on the structural integrity of the tube itself.