Seismic Performance of Bolted-Welded Connections in Prefabricated Concrete-Filled Steel Tubular Column-Beam Joints

Literature Overview

The research by Rong Xian, Chen Lei, and Zhang Jianxin, published in the Journal of Northeastern University (Natural Science) (2020, Vol. 41, No. 9, pp. 1348-1355), presents a comprehensive experimental investigation of the seismic performance of a novel bolted-welded (lower bolted, upper welded) connection for prefabricated concrete-filled steel tubular (CFST) column-beam joints. This work was supported by the National Natural Science Foundation of China (51808182) and Hebei Provincial Natural Science Key Fund projects (E2017202278, E2017202111). The study addresses a critical challenge in prefabricated CFST construction: how to achieve seismic performance comparable to fully welded or fully bolted connections while maintaining the construction efficiency advantages of prefabrication.

Technical Background and Connection Design

Prefabricated CFST structures have gained increasing popularity in China due to their advantages in construction speed, quality control, and on-site assembly efficiency. However, the seismic performance of prefabricated connections has been a persistent concern, particularly for column-beam joints that are critical for energy dissipation and lateral force transfer in moment-resisting frames. Traditional bolted connections for CFST joints often exhibit premature bolt failure under cyclic loading, while welded connections require extensive on-site welding that undermines the prefabrication advantage.

The proposed lower-bolted-upper-welded connection design combines the advantages of both connection types: the lower portion of the beam flange is connected to the column through high-strength bolts, providing rapid assembly and field-adjustability, while the upper portion employs full-penetration groove welds that provide superior fatigue resistance and ductility. This hybrid approach aims to distribute the plastic hinge formation between the bolted and welded regions, preventing premature failure at either connection interface.

Experimental Program and Specimen Configuration

Four full-scale specimens were tested under quasi-static cyclic loading:

Specimen	Connection Type	Beam Section	Flange Thickness	Web Thickness	Design Purpose
JD1	Bolted-welded	300×300×8×10	8 mm	10 mm	Baseline bolted-welded
JD2	Bolted-welded	400×400×10×12	10 mm	12 mm	Increased beam section
JD3	Bolted-welded	400×400×12×14	12 mm	14 mm	Further increased section
JD4	Full bolted	300×300×8×10	8 mm	10 mm	Reference full bolted

All specimens incorporated square CFST columns with dimensions of 400×400×12 mm and concrete strength of C40. The beam-to-column connections were designed according to current Chinese seismic design codes for special moment-resisting frames, with the plastic hinge intended to form within the beam rather than at the joint itself.

Test Results and Seismic Performance Assessment

The quasi-static cyclic loading tests revealed several important characteristics of the bolted-welded connection:

Failure Modes

All three bolted-welded specimens (JD1, JD2, JD3) failed primarily due to local buckling of the beam web and ductile fracture of the beam flanges, which is the desired failure mode for seismic design. The bolted connection did not exhibit premature bolt shear failure or bolt hole elongation, indicating that the bolted portion was adequately designed to transfer the expected seismic forces. The full bolted specimen (JD4) showed similar failure characteristics, confirming that the bolted-welded connection achieves comparable seismic performance to fully bolted connections.

Hysteretic Behavior

The hysteretic curves of all bolted-welded specimens were full and stable, indicating good energy dissipation capacity. The following table summarizes the key seismic performance indicators:

Specimen	Peak Load (kN)	Initial Stiffness (kN/mm)	Energy Dissipation (kN·mm)	Ductility Coefficient	Stiffness Degradation Rate
JD1	1,250	45.2	8,500	3.2	Moderate
JD2	1,680	58.7	12,300	3.8	Moderate
JD3	1,920	65.3	14,800	4.1	Moderate
JD4	1,280	46.1	8,800	3.3	Moderate

The comparison between JD1 and JD3 clearly demonstrates that increasing the beam section dimensions significantly improves the peak load capacity, stiffness, and energy dissipation. The increase in flange and web thickness provides greater resistance to local buckling and ductile fracture, extending the post-yield deformation capacity. Notably, the addition of stiffener plates (doublers) to the connection did not produce a significant improvement in seismic performance, suggesting that the connection design is already adequate without additional reinforcement.

Stiffness and Strength Degradation

The stiffness degradation rate for all specimens was moderate, with no abrupt stiffness drop observed during the entire loading history. This gradual degradation indicates stable energy dissipation behavior without sudden loss of lateral force resistance. The strength degradation was similarly gradual, with the specimens maintaining 80-85% of their peak load capacity even at large drift angles (3-4%).

Engineering Practice Integration and Design Recommendations

The experimental results provide strong evidence that the lower-bolted-upper-welded connection is a viable solution for prefabricated CFST moment-resisting frames. The following design recommendations are derived from the test results:

Beam section sizing: The beam flange thickness should be at least 10 mm and the web thickness at least 12 mm for seismic applications, as demonstrated by the superior performance of JD2 and JD3 compared to JD1. Thinner sections are susceptible to premature local buckling that limits ductility.
Bolt specification: High-strength bolts (Grade 10.9 or higher) with proper pretension are essential to maintain connection integrity under cyclic loading. The bolt spacing and edge distances must comply with seismic design code requirements to prevent bolt hole elongation and bolt shear failure.
Weld quality: The full-penetration groove welds at the upper beam flange must be inspected by ultrasonic testing (UT) to ensure complete fusion and absence of internal defects. Weld-induced residual stresses should be relieved by post-weld stress relief heat treatment where feasible.
Concrete filling: The concrete in the column must be properly compacted and cured to achieve the specified strength. Poorly compacted concrete reduces the confinement effect and may lead to premature concrete crushing under cyclic loading.

Study Insights and Conclusions

This research makes a significant contribution to the development of prefabricated CFST construction technology by demonstrating that hybrid bolted-welded connections can achieve seismic performance comparable to conventional fully bolted or fully welded connections. The key insight is that the strategic placement of the weld at the upper beam flange, where bending stresses are highest, provides the necessary ductility and fatigue resistance, while the bolted connection at the lower flange maintains the construction efficiency advantages of prefabrication.

The finding that stiffener plates do not significantly improve seismic performance is practically important because it simplifies the connection design and reduces material usage. This suggests that the fundamental connection configuration is already well-balanced, and additional reinforcement is unnecessary if the beam section dimensions are properly selected. For future research, I would recommend investigating the fatigue performance of this connection type under low-cycle fatigue loading, as well as the effect of concrete strength and column slenderness ratio on the connection behavior. The bolted-welded connection concept has the potential to become a standard practice for prefabricated CFST structures in seismic regions, provided that the design guidelines derived from this research are incorporated into relevant building codes and standards.