Seismic Performance Analysis of Rectangular Concrete-Filled Steel Tube Columns Under Cyclic Loading

Literature Overview

The paper by Li Xueping, Lü Xilin, and Guo Shaochun, published in Earthquake Engineering and Engineering Dynamics in 2005 (Vol. 25, No. 5, pp. 104-111), presents a nonlinear analytical study of rectangular concrete-filled steel tube (CFST) columns subjected to cyclic loading. Funded by the National Science Fund for Distinguished Young Scholars and the National Innovation Research Group Fund, this work was conducted at Tongji University's State Key Laboratory for Disaster Prevention in Civil Engineering and the College of Civil and Hydraulic Engineering at Ningxia University.

Core Analytical Framework

Constitutive Models Developed

The authors developed three key constitutive relationships to capture the complex material behavior under cyclic loading:

The Bauschinger effect in steel represents the reduction of yield strength upon load reversal, a phenomenon critical for accurately predicting the cyclic response of steel tubes. The confinement model for rectangular CFST columns must account for the non-uniform lateral constraint provided by the rectangular cross-section, where corner regions experience more effective confinement than mid-span regions of the flat walls.

Analytical Model Architecture

The analysis model is built upon a fiber-section approach with the following hierarchy:

1. Fiber level: Individual material points characterized by uniaxial stress-strain relationships.
2. Interface level: Bond-slip elements connecting the steel tube fibers to the concrete core fibers, allowing relative displacement at the interface.
3. Section level: Assembly of fibers into a rectangular cross-section with appropriate geometric distribution.
4. Column level: Assembly of sections along the member length, incorporating geometric nonlinearity and P-Δ effects.

The analysis program was developed in Fortran, implementing a displacement-controlled incremental-iterative procedure with Newton-Raphson convergence at each load step. The bond-slip interface elements are critical because they govern the load transfer mechanism between the steel tube and concrete core, which directly affects the overall ductility and energy dissipation capacity.

Numerical Simulation Results

Comparison with Experimental Data

The numerical simulations were validated against experimental test results obtained by the authors on rectangular CFST column specimens. The comparison was conducted at two levels:

• Global response: Load-displacement hysteresis curves showed good agreement in peak strength, post-peak degradation rate, and energy dissipation capacity.
• Local response: Load-strain curves at critical sections captured the progressive crushing of concrete and local buckling of the steel tube walls.

Minor discrepancies were observed in the post-peak region, attributed to the idealized representation of concrete crushing behavior and the simplified treatment of local buckling initiation criteria.

Stress-Strain Evolution Tracking

The analysis program enabled detailed tracking of material-level responses throughout the cyclic loading history:

• Steel tube: The outer fibers of the steel tube experienced early yielding at the column ends, with progressive strain accumulation leading to local buckling at drift ratios beyond 3-4%.
• Concrete core: The confined concrete exhibited progressive strength degradation with each loading cycle, but the confinement effect maintained residual strength even after significant damage.
• Interface: The bond-slip response showed initial elastic behavior, followed by progressive debonding at high drift levels, ultimately leading to partial separation at the corners.

Influence of Interface Shear Strength

A parametric study investigated the sensitivity of column performance to the tangential bond strength at the steel-concrete interface. The results demonstrated that:

• Higher interface shear strength improves initial stiffness and peak load capacity by approximately 5-10%.
• The effect on ductility is more pronounced, with increased bond strength delaying the onset of interface separation and maintaining composite action at higher drift levels.
• Beyond a certain threshold, additional bond strength provides diminishing returns because failure transitions from interface separation to concrete crushing or steel buckling.

Engineering Practice Implications

For the design of rectangular CFST columns in seismic regions, this study provides several practical insights:

1. Interface treatment importance: The bond-slip behavior at the steel-concrete interface is not merely a secondary consideration but a primary factor governing seismic performance. Surface treatments, mechanical interlocks, or chemical adhesives that enhance interface bond strength can significantly improve ductility.
2. Cross-section optimization: The non-uniform confinement effect in rectangular sections suggests that corner regions are more efficiently utilized than mid-wall regions. This has implications for optimizing wall thickness distribution and corner radius selection.
3. Damage assessment: The tracked stress-strain evolution provides criteria for identifying damage states in the field. Strain monitoring at critical locations can indicate the progression from elastic to yielding to crushing/buckling stages.
4. Design code calibration: The analytical results provide benchmark data for calibrating and validating design equations in seismic codes, particularly regarding confinement strength enhancement factors and ductility modification factors.

Study Insights and Reflections

This study represents a methodologically rigorous approach to understanding the seismic behavior of composite columns. The development of a bond-slip hysteretic model for the steel-concrete interface is particularly noteworthy, as many analytical models either ignore this interface entirely or treat it as perfectly bonded. In reality, the interface behavior governs the load transfer mechanism and ultimately determines whether the column fails in a ductile or brittle manner.

The use of fiber-section modeling with interface elements represents a significant advancement over earlier approaches that assumed perfect composite action throughout the loading history. This approach captures the progressive nature of damage accumulation and the eventual loss of composite action, which is essential for predicting realistic post-peak behavior.

One area for further development is the incorporation of local buckling as a distinct failure mechanism. The current model captures the consequences of buckling through the stress-strain response of individual fibers but does not explicitly model the buckling initiation and propagation. A more refined approach would incorporate a local buckling criterion that triggers a degradation in the effective stiffness and strength of the steel tube fibers.