Double Nonlinear Finite Element Analysis of Concrete-Filled Circular Steel Tube Arch Structures

Literature Overview

The paper by Chen Baoshun and Lin Jiayang (Journal of the Railway Society, 2005, Vol. 27, No. 6, pp. 77–84) presents a dual-nonlinear finite element methodology for analyzing the full-range spatial behavior of concrete-filled steel tube (CFST) arches. The authors developed a dedicated finite element program employing spatial beam elements with a fiber model for normal stress–strain and a unified shear stress–strain curve. This work was supported by the Fujian Provincial Key Science and Technology Program (2003F007) and the Ministry of Transport Western Transportation Science and Technology Program (2003-318-798-20-1). The study is particularly significant for engineers working on railway viaduct arches, where the interaction between the steel tube and infill concrete under spatial loading demands rigorous nonlinear characterization.

Core Technical Approach

The dual nonlinearity addressed in this work encompasses both geometric nonlinearity (large displacements, large rotations, and P-Δ effects) and material nonlinearity (plastic yielding, concrete crushing, and steel tube buckling). The fiber element model discretizes the cross-section into multiple material fibers, each independently following its own stress–strain law, which is far superior to smeared-section models for capturing the progressive confinement effects in CFST members. The shear behavior is modeled using a unified theory curve that accounts for the coupled effect of normal and shear stresses on concrete degradation.

From a pipe manufacturing perspective, the fiber model's accuracy is directly dependent on the material properties of the steel tube itself. The yield strength, ultimate tensile strength, and strain-hardening behavior of the steel tube (typically Q235 or Q345 per GB/T 1591 or GB/T 700) must be precisely characterized through tensile testing in accordance with GB/T 228.1. The longitudinal and transverse properties may differ slightly in ERW or HFW welded pipe due to the heat-affected zone, and this anisotropy should ideally be incorporated into the fiber model for high-fidelity analysis.

Key Findings and Their Engineering Implications

The authors conducted comparative analyses on single-rib and double-rib model arches, examining the influence of spatial lateral forces. The principal conclusions are summarized below.

The finding that geometric nonlinearity dominates over material nonlinearity in spatial loading has direct implications for pipe selection. Thicker-walled tubes (e.g., t ≥ 12 mm for large-diameter arch ribs) provide greater stiffness against P-Δ effects, thereby reducing the geometric nonlinearity contribution. However, thicker tubes require more energy-intensive forming processes such as cold-bending or hot-bending, and the residual stress distribution after bending must be carefully controlled. For arch ribs with radius-to-thickness ratios exceeding 30, the residual stress from cold forming can reduce the local buckling capacity by 5–10%, a factor that must be incorporated into the fiber model's initial stress field.

The observation that single-rib arches exhibit more pronounced spatial behavior than double-rib arches is also relevant to welding practice. In single-rib configurations, the entire load path passes through fewer weld joints, concentrating the demand on each joint. The longitudinal weld seams in LSAW or UOE-manufactured arch ribs become critical load-bearing elements, and their quality directly governs the arch's overall spatial stability. According to API 5L and SY/T 5047, the weld seam must achieve full radiographic inspection with acceptance criteria no looser than Level II per GB/T 3323.

Connection with Pipe Manufacturing and Welding Quality

From my experience in pipe manufacturing, the success of a CFST arch structure depends fundamentally on the quality of the steel tube and its welds. The following table summarizes the critical manufacturing parameters that influence the structural performance predicted by the dual-nonlinear analysis.

The dual-nonlinear analysis assumes idealized material behavior, but in reality, the steel tube may exhibit local imperfections such as weld undercut, overlap, or incomplete fusion. These defects act as stress concentrators that trigger localized yielding at loads well below the theoretical yield point. A single-rib arch with a longitudinal weld defect can experience a 15–25% reduction in peak load compared to the analytical prediction, a discrepancy that must be accounted for in safety factors.

Study Insights and Reflections

This paper represents a significant advance in the computational analysis of CFST arches, bridging the gap between simplified linear elastic models and the complex reality of spatial nonlinear behavior. However, from a manufacturing and quality control standpoint, I note that the fiber model's accuracy is only as good as the input material data. In practice, batch-to-batch variations in steel tube properties—stemming from mill heat treatment differences, welding procedure variations, or forming process inconsistencies—can lead to significant scatter in actual structural performance. Engineers should therefore recommend representative material testing on each production batch of arch tube, rather than relying solely on mill test certificates.

The work also highlights the importance of considering initial imperfections in the finite element model. In my experience with UOE pipe production, the typical initial ovality is 0.5–1.0% of the nominal diameter, and this imperfection can reduce the local buckling load by 10–15% for slender tubes. Incorporating measured initial imperfection profiles into the fiber element model would significantly improve the predictive accuracy of the dual-nonlinear analysis. This remains an area for future research that I believe deserves more attention from both the structural engineering and pipe manufacturing communities.