Experimental and Finite Element Analysis of New Flange Connections for Prefabricated CFST Core Columns

Literature Overview

This study investigates a novel flange connection system designed specifically for prefabricated concrete-filled steel tube (CFST) core columns used in modular building construction. The research combines full-scale cyclic loading experiments with three-dimensional nonlinear finite element analysis to characterize the mechanical behavior, failure modes, and design capacity of the proposed connection. The motivation stems from the growing demand for prefabricated construction methods that can reduce on-site labor, construction time, and quality variability in high-rise building projects.

Connection Configuration and Design Philosophy

The proposed flange connection utilizes a bolted flange plate welded to the top of each CFST column segment, with a matching flange plate welded to the bottom of the segment above. The connection is designed to transfer axial compression, shear, and bending moment between prefabricated column segments while maintaining the integrity of the concrete core. The key design features include:

A reinforced flange plate with increased thickness at the bolt holes to prevent net-section failure.
A stiffener ring welded to the outer surface of the steel tube at the flange location to distribute local bearing stresses.
High-strength bolts (Grade 8.8 or 10.9 per GB/T 1228) arranged in a circular pattern with controlled bolt pretension.
A shear key or dowel pin arrangement to limit lateral slip between segments during seismic loading.

The design philosophy follows the "strong joint, weak member" principle, ensuring that the plastic hinge forms within the column segment rather than at the connection. This is achieved by designing the flange connection to have a nominal moment capacity at least 1.25 times the plastic moment capacity of the column cross-section.

Experimental Program and Results

The experimental program included six full-scale specimens with varying parameters to evaluate the influence of key design variables on connection performance:

Specimen	Steel Tube Diameter	Wall Thickness	Flange Thickness	Bolt Grade	Concrete Strength	Axial Load Ratio
F-1	300 mm	10 mm	20 mm	8.8	40 MPa	0.3
F-2	300 mm	10 mm	20 mm	10.9	40 MPa	0.3
F-3	300 mm	10 mm	25 mm	10.9	40 MPa	0.3
F-4	300 mm	10 mm	20 mm	10.9	60 MPa	0.3
F-5	300 mm	10 mm	20 mm	10.9	40 MPa	0.5
F-6	300 mm	10 mm	20 mm	10.9	40 MPa	0.6

The cyclic loading protocol followed the displacement-controlled method with drift ratios ranging from 0.5% to 6.0%, applying two load cycles at each drift level. The test results revealed that the connection exhibited excellent ductility with drift capacities exceeding 5.0% for most specimens, and the hysteresis loops were full and stable, indicating good energy dissipation capacity.

Finite Element Modeling Approach

The nonlinear finite element model was developed in ABAQUS using the following element and material formulations:

The steel tube was modeled using S4R four-node reduced-integration shell elements with a 5 mm mesh size at the flange region and 15 mm elsewhere.
The concrete core was modeled using C3D8R eight-node reduced-integration solid elements with a 30 mm mesh size.
The flange plates and stiffener rings used S4R shell elements with appropriate thickness assignments.
The bolt connections were simulated using a combination of beam elements for the bolt shanks and penalty contact pairs for the bolt-nut interfaces.
The concrete material followed the Concrete Damaged Plasticity model with calibrated fracture energy and dilation angle parameters.
The steel material used a multilinear kinematic hardening model to capture the Bauschinger effect under cyclic loading.

The model was validated against experimental results by comparing the load-displacement curves, strain distribution patterns, and failure modes. The maximum deviation in peak load was within 8%, confirming the model's accuracy for parametric studies.

Key Findings and Design Recommendations

The parametric study revealed several important design trends:

Bolt grade influence: Upgrading from Grade 8.8 to Grade 10.9 bolts increased the connection's moment capacity by approximately 15%, but the ductility improvement was marginal, suggesting that bolt strength alone is not the governing factor for seismic performance.
Flange thickness: Increasing the flange thickness from 20 mm to 25 mm improved the local buckling resistance of the flange plate and delayed the onset of plastic deformation at the bolt holes.
Axial load ratio: The connection's ductility degraded significantly when the axial load ratio exceeded 0.5, with a 40% reduction in drift capacity at 0.6 compared to 0.3. This is attributed to the increased compressive stress in the flange plate, which reduces the tensile reserve at the bolt holes.
Concrete strength: Higher concrete strength (60 MPa vs. 40 MPa) contributed to improved post-peak load retention due to the enhanced confinement effect of the steel tube on the concrete core near the connection.

Engineering Practice and Quality Control

The prefabrication of CFST column segments with flange connections requires strict quality control at each manufacturing stage:

Welding inspection: The flange-to-tube welds must undergo 100% ultrasonic testing (UT) per GB/T 11345, with acceptance criteria based on GB/T 3323-2005 Level B quality.
Bolt pretension: Each bolt must be tightened to the specified pretension force using calibrated torque wrenches or hydraulic tensioners, with verification by direct tension indicators.
Dimensional tolerance: The flange plate flatness must not exceed 1 mm over the bolt circle diameter, and the bolt hole position tolerance must be within ±0.5 mm per GB/T 3323.
Concrete quality: The concrete used for the CFST core must be high-strength, low-slump concrete (slump ≤ 50 mm) to ensure proper compaction within the confined steel tube.

Study Insights and Implications

This research provides a comprehensive framework for the design and verification of flange connections in prefabricated CFST columns. The combined experimental and numerical approach offers confidence in the proposed design method, which can be directly applied to engineering projects involving modular construction. The finding that axial load ratio is the most critical parameter for connection ductility has important implications for structural system design, suggesting that the seismic design of prefabricated CFST buildings should limit the axial load ratio to below 0.5 in the plastic hinge region.

The study also highlights the importance of proper detailing at the connection interface, where the transition between the prefabricated segments occurs. Future research should address the long-term durability of the bolted flange connections under environmental exposure, particularly in corrosive environments where the bolt threads and flange edges are vulnerable to rust and pitting.