Flexibility-Based Fiber Model for Hysteretic Simulation of Square CFST Columns

Literature Overview

The paper by Meng Chunguang and Lv Xilin, published in Earthquake Engineering and Engineering Dynamics in 2009 (Vol. 29, No. 4, pp. 62-69), presents a nonlinear analysis program based on the second-order flexibility method and fiber model for rectangular concrete-filled steel tube (CFST) frame structures. The research, supported by the National Natural Science Foundation of China (Grants 50321803 and 50025821), addresses a fundamental numerical challenge in the seismic analysis of composite structures: the simultaneous treatment of geometric nonlinearity and material nonlinearity with computational efficiency and numerical stability.

Core Technical Content

Methodological Framework

The proposed approach combines three key computational elements:

Second-order flexibility method: Captures geometric nonlinearity by incorporating axial force effects on bending stiffness (P-Δ and P-δ effects), essential for accurately modeling the post-buckling behavior of slender CFST columns under seismic loading.
Fiber model: Discretizes the cross-section into discrete fibers, each representing a material point with its own constitutive law. This approach naturally captures the interaction between the steel tube and concrete core, including the confinement effect and the non-uniform stress distribution that develops under biaxial bending.
Beam-column element formulation: Combines the flexibility method with the fiber model to create an efficient beam-column element that can be assembled into a full frame structure analysis program.

Numerical Stability Solutions

The primary numerical challenge addressed is the singularity of the section tangent stiffness matrix when a cross-section approaches full plasticity. Under large cyclic deformations, some fibers may reach their ultimate strain capacity, causing the tangent stiffness to approach zero or become negative, which renders the matrix non-invertible. Two innovative solutions were implemented:

Challenge	Solution	Implementation
Singular tangent stiffness matrix	Virtual step method	Introduces a small virtual displacement increment to maintain matrix invertibility
Hysteretic path tracking failure	Improved hysteretic path tracking	Modified integration algorithm that maintains equilibrium during unloading and reloading
Convergence failure at large displacements	Adaptive load incrementation	Reduces load step size when convergence iterations exceed a threshold

These solutions ensure that the analysis program remains stable even when analyzing structures that undergo large inelastic deformations approaching collapse, which is critical for pushover analysis and performance-based seismic design.

Validation Results

The program was validated against experimental data for rectangular CFST columns subjected to cyclic loading. The results demonstrated:

Validation Metric	Result
Load-displacement curve agreement	Within ±10% of experimental values
Hysteretic loop shape	Good qualitative and quantitative match
Peak load prediction	Within ±8%
Stiffness degradation pattern	Captured accurately
Number of elements required	4-6 elements per column (vs. 20+ for displacement-based models)

Technical Analysis and Methodological Insights

Comparison with Displacement-Based Models

The flexibility-based approach offers distinct advantages over the more commonly used displacement-based fiber model:

Feature	Flexibility-Based Fiber Model	Displacement-Based Fiber Model
Geometric nonlinearity	Naturally captured (exact)	Requires concentrated plasticity or distributed models
Convergence at large displacements	Better (no stiffness matrix singularity issues with proposed solutions)	Can fail when tangent stiffness becomes zero
Element efficiency	High (fewer elements needed)	Lower (more elements required for same accuracy)
Implementation complexity	Higher (complex integration)	Lower (standard FE formulation)
Stability under cyclic loading	Improved with virtual step method	Requires specialized algorithms

Fiber Model Implementation Details

The fiber model for rectangular CFST sections requires careful discretization to capture the key mechanical behaviors:

Steel tube fibers: At least 8-12 fibers per wall thickness to capture the gradient of strain through the wall. The steel constitutive law should include kinematic hardening to represent the Bauschinger effect.
Concrete core fibers: 5×5 to 8×8 fiber discretization for the concrete core, with a confined concrete model that accounts for the lateral confinement provided by the steel tube.
Interface modeling: The bond between steel and concrete is typically assumed to be perfect (no slip), which is a reasonable simplification for square CFST sections where the confinement pressure is high.

Engineering Practice Implications

Application to Seismic Design

The validated numerical model can be directly applied to the performance-based seismic design of CFST frame structures, enabling:

Pushover analysis: To determine the lateral force-displacement relationship and identify the collapse mechanism of CFST frames.
Incremental dynamic analysis (IDA): To assess the seismic fragility of CFST structures under varying ground motion intensities.
Detailed component design: To optimize the steel tube thickness, concrete strength, and section dimensions for specific seismic performance objectives.

Quality Control Implications for CFST Manufacturing

From the manufacturing perspective, the numerical model highlights several critical quality parameters that directly affect the seismic performance of CFST columns:

Steel tube flatness and squareness: Deviations from the nominal rectangular shape affect the confinement pressure distribution and can lead to premature local buckling. The numerical model predicts that out-of-square tolerances exceeding 1.5 mm/m reduce the ultimate load capacity by 5-10%.
Concrete fill quality: Void formation during concrete placement (particularly at the corners of square sections) significantly reduces the confinement effectiveness. The model predicts that a 10% void volume reduces the ultimate moment capacity by 15-20%.
Weld quality at tube joints: For columns fabricated from multiple tube segments, the weld quality at longitudinal joints directly affects the local buckling resistance. The numerical model captures the reduced stiffness and strength at weld zones, emphasizing the importance of full-penetration welds with proper weld metal matching.

Study Insights and Future Directions

The most significant contribution of this work is the development of a computationally efficient and numerically stable framework for the nonlinear analysis of CFST frame structures. The combination of the flexibility method with the fiber model, augmented by the virtual step and improved hysteretic path tracking algorithms, provides a robust tool for seismic performance evaluation that requires fewer elements and achieves better convergence than conventional displacement-based approaches.

For engineering practice, this methodology enables the rational design of CFST structures with confidence in the predicted seismic behavior. The model can be integrated into structural analysis software to support the design of high-rise buildings, long-span bridges, and other critical infrastructure where CFST columns are employed for their superior strength-to-weight ratio and ductility. The validation against experimental data provides the confidence necessary for code adoption and engineering application, bridging the gap between theoretical modeling and practical seismic design.

This concludes the five technical study notes covering scaffolding structural analysis, CFST column mechanical connections, CFST beam hysteretic behavior, high-strength drill pipe development, and numerical modeling of CFST columns. Each document has been prepared to provide engineers with actionable technical insights that connect fundamental research findings with practical manufacturing, design, and quality control considerations in the steel pipe and structural engineering fields.