Seismic performance of a novel occlusive fully bolted beam-column joint

Abstract

The fully bolted beam-column joint offers advantages such as good load-bearing performance, convenient construction, and excellent seismic resistance. Traditional fully bolted beam-column joints typically require a greater quantity of bolts or bolts with larger diameters to meet the force transfer requirements, which not only complicates construction and leads to material redundancy but also weakens the net section of the flange and cover plate due to bolt holes. To address the above issues, a new type of occlusive fully bolted beam-column(OFBBC) joint has been proposed. To study the seismic performance of this joint, cyclic loading tests were conducted on two OFBBC joints and one traditional fully bolted joint. Additionally, a static loading test was conducted on one OFBBC joint. The seismic performance of the joint was studied from aspects such as failure mode, ultimate load, ductility, and energy dissipation capacity. Subsequently, finite element models of the OFBBC joint were established using ABAQUS program, and their accuracy was validated based on the test results. Research results indicate that compared to traditional fully bolted joint, the new OFBBC joint can not only enhance the ultimate bearing capacity without increasing the number of bolts but also exhibit better rotational stiffness, ductility, and energy dissipation capacity, demonstrating superior seismic performance. These findings can provide valuable insights for the application of the OFBBC joint.

Keywords

occlusive fully bolted beam-column joint grooves cyclic loading test static loading test seismic performance finite element models

Introduction

Beam-column connections are crucial components in steel structures, responsible for transmitting axial forces, shear forces, and moments within the structure. Their performance directly affects the overall safety and reliability of the structure (Zhan et al., 2021) Based on the types of connections in steel structure beam-column joints, these joints can be categorized into welded joints, bolted connections, and welded-bolted hybrid joints. The quality of on-site welding is influenced by various uncontrollable factors such as weather, welding techniques, and workspace constraints, making it challenging to ensure high-quality welds. Additionally, welding can introduce residual stresses and deformations, reducing the ductility of materials and leading to brittle failure of the structure. Post-earthquake investigations following events like the Northridge earthquake in the United States (Miller, 1998) and the Hanshin earthquake in Japan (Huang and Zhang, 1995) revealed that beam-column welded joints failed due to the formation of brittle cracks in the welds, causing severe damage to buildings during earthquakes.

Compared to welded connections, bolted connections offer simplicity in operation, high construction quality, good load-bearing performance, easy disassembly and high recyclability. They also exhibit strong fatigue resistance and good ductility, providing better seismic performance. Therefore, bolted connections have shown significant potential for development in high-intensity seismic areas, and the seismic performance of fully bolted beam-column joints has become a focal point of research interest for numerous scholars.

(Chu et al., 2022, 2024) proposed a fully bolted beam-column joint with a core tube and conducted hysteretic tests on the joint, demonstrating its seismic performance. They found that this joint exhibited better ductility and energy dissipation capabilities compared to traditional joints (Liu et al., 2017, 2024; Liu et al., 2025). introduced a beam-column joint designed for modular construction, dividing the joint into upper and lower parts pre-welded in the factory to ensure weld quality. Cyclic loading tests were conducted on five joints, which showed significant improvements in hysteresis behavior and energy dissipation. Similarly (Zhong et al., 2024, 2025), proposed a novel configuration for fully bolted connections in modular units. They connected the upper and lower columns using inner sleeves and utilized high-strength bolts and cover plates for beam connections, demonstrating clear load paths and easy installation. Static tests and numerical simulations verified the excellent mechanical performance of the joints, promoting the development of modular construction (Korol et al., 1993; Mourad et al., 1995). suggested using high-strength blind bolts to connect hollow steel columns with end-plate beams to address issues with nut tightening inside hollow columns. The study conducted hysteretic tests and frame tests on the joint, revealing that it exhibited excellent performance in terms of stiffness, bearing capacity, and ductility. Additionally, the performance of blind bolts was found to be superior to that of conventional high-strength bolts. Building on this (Li et al., 2023), tested five joints with different configurations for beams (steel beams and composite steel beams with concrete slabs) and columns (steel square tubes and concrete-filled steel square tubes). They compared the load-carrying capacity, rotational stiffness, and energy dissipation capabilities. (Zhang et al., 2022) proposed a flange splice joint for H-shaped steel beams and concrete-filled steel tube (CFST) columns, where H-shaped steel beams were prefabricated and welded to column bases in the factory. The column bases were flange-connected to upper and lower columns for on-site fully bolted connections. Testing four full-scale specimens provided insights into the seismic performance of the joints, indicating that thickened beam end plates and beam flanges could enhance ductility and energy dissipation. (Fan et al., 2022) addressed the challenge of opening holes in CFST column flanges by introducing a method using plastic tubes for through-bolt holes. Testing six specimens under cyclic loads revealed good energy dissipation and ductility. (Chen et al., 2021) presented a beam-column joint entirely composed of standardized H-shaped steel components, interconnected with bolts for ease of assembly and transportation. Through experiments and numerical analyses, they obtained hysteretic curves and stiffness degradation curves, and analyzed the energy dissipation capabilities of the joints.

In these fully bolted joints, bolts typically resist shear forces. However, the individual shear capacity of a single bolt is limited. When bolts alone serve as the load-transferring elements, numerous bolts or large diameters are often required to ensure full shear transfer. This can weaken the net section, reduce efficiency, and necessitate larger plates to maintain bolt spacing. As a result, joints may become oversized, increasing material use and cost, and compromising aesthetics. Excessive bolts may also slow construction speed and reduce efficiency. Therefore, improving the bearing capacity of fully bolted joints without increasing bolt number or diameter has become a pressing engineering issue.

To address these problems, (Jiang et al., 2016; Guo et al., 2016, 2020) proposed a new occlusive high-strength bolted (OHSB) connection, which improved contact surfaces via grooves between the connecting components. By tightening bolts to press the components together, the grooves enabled shear transfer. Research showed that under similar surface treatment and bolt arrangement, the anti-slip capacity of OHSB joints was significantly higher than traditional friction-type connections. However, the application of this technique was limited to connections and had not been extended to joints.

Based on the OHSB concept, this paper proposes a novel occlusive fully bolted beam-column (OFBBC) joint. In this joint, shear forces are transferred through grooves, thereby reducing the number of bolts. To evaluate the seismic performance of OFBBC joints, cyclic loading tests were conducted on two OFBBC joints with different groove counts and one traditional high-strength bolted joint. A static loading test was also performed on one OFBBC joint. Finally, finite element analysis (FEA) was carried out to analyze the hysteretic behavior of the joints.

The construction of the OFBBC joint

The construction of the OFBBC joint is shown in Figure 1. The joint consists of a square-tube column, an outer ring plate, a welded web plate, cover plates, and an H-beam, all connected using high-strength bolts. The H-beam flange is connected to the outer ring plate through upper and lower flange-cover plates, while its web is connected to the welded web plate via two web-cover plates, as shown in Figure 1(a). Grooves are milled into the outer ring plates, beam flanges, and flange-cover plates, as shown in Figure 1(b). To accommodate erection tolerances and ensure groove alignment, slotted holes are drilled in the outer ring plate and beam flange, while circular holes are bored in the flange-cover plates. This joint is easy to assemble, requires fewer bolts, and enhances installation efficiency.

Figure 1.

Construction of the OFBBC joint.

Experimental design

Design of specimens

A total of four joint specimens were set up for the experiment, among them, J-0, J-4, and J-6,with 0, 4, and 6 grooves, respectively, were subjected to cyclic loading tests. Specimen D-4, with grooves, was tested under a static loading. The square-tube column had a cross-sectional size of 200 mm × 8 mm and a length of 1300 mm. The H-beam had a cross-sectional size of 300 mm × 150 mm × 6.5 mm × 9 mm and a length of 1150 mm. High-strength M16 bolts of grade 8.8 were used. The square-tube column was connected to the outer ring plate and welded web plate by welding. The dimensions of the other components are shown in Figure 2. It is noteworthy that J-0 was a traditional high-strength bolted joint without grooves, standard circular holes were bored in the outer ring plate and flanges. According to reference (Guo et al., 2020), the size of all grooves were 1.5 mm × 2 mm, as shown in Figure 2(f). The actual appearance of the groove area is showed in Figure 2(g).

Figure 2.

Dimensions of joint components.

Material properties

All components in the experiment were made of Q355B steel. To determine the material properties, static tensile tests were conducted on the Q355B steel used in the joint specimens. Six standard specimens were designed (GB/T 2975-1998, 1998), with three 8 mm thick specimens taken from the column flange and three 9 mm thick specimens taken from the beam flange. The loading procedure followed the specifications of reference (GB/T 228-2010, 2010). The stress-strain curves of specimens with two thicknesses are shown in Figure 3. The curves exhibit a similar trend, featuring a distinct yield plateau. The obtained experimental results are shown in Table 1, and all indicators meet the requirements (GB/T 228-2010, 2010).

Figure 3.

Stress-Strain curves of standard specimens.

Table 1.

Results of material tests.

Specimen	Elastic modulus/Gpa	Yield stress/Mpa	Ultimate stress/Mpa	Elongation at fracture/%
8-1	205.19	383.81	566.19	27.5
8-2	202.20	389.31	564.63	29.0
8-3	207.17	387.38	564.75	33.0
Average	204.85	386.83	565.19	29.8
9-1	202.80	409.94	534.94	27.0
9-2	206.85	417.83	546.78	33.5
9-3	197.07	411.17	544.67	32.9
Average	202.24	412.98	542.13	31.1

Instrumentation

Layout of instrumentation is shown in Figure 4. Layout of linear variable displacement transducers (LVDTs) is shown in Figure 4(a). DT-1 to DT-2 and DT-4 to DT-5 were used to measure the rotation of the column within the plane. DT-3, DT-8 to DT-10 were used to measure the deflection of the outer ring plate and the beam, with DT-8 to DT-10 also measuring the rotation angle of the connection region. DT-6 and DT-7 measured the column base support rotation. DT-11 and DT-12 measured the displacement at the beam ends as well as the degree of torsion at the beam ends.

Figure 4.

Layout of instrumentation.

Layout of strain gauges is shown in Figure 4(b).On the column surface, S1 to S8 and S9 to S16 were arranged in eight directions above and below to check for eccentricity and asymmetry. Four strain gauges (S17 to S20) were placed on the beam flange near the joint area to monitor flange yielding and check for potential torsion.

Experimental setup and loading system

The experimental setup is shown in Figure 5. Approximate Design Method was used to preliminarily estimate the yield load of the joint. It was assumed that the beam-end moment was entirely carried by the flanges, while the shear force was entirely carried by the web. The load corresponding to the yielding of the weakened beam flange was taken as the yield load P_y, which was calculated using equation (1).

P_{y} = \frac{f_{y} (b_{f} - 2 d) (t_{f} - h_{g}) h_{b}}{l}

(1)

In the equation, l is the distance from the loading end to the weakened beam flange, b_f is the beam width, d is the bolt hole diameter, t_f is the beam flange thickness, h_g is the groove depth, and h_b is the center-to-center distance between the upper and lower beam flanges.

Figure 5.

The experimental setup.

Before loading, all high-strength bolts of the specimen were tightened to the required preload specified (JGJ 82-2011, 2011). Before formally applying the load at the beam ends, preload was performed. Then, an axial force of 560 kN was applied to the top of each column. Before applying the axial force, the screws at both ends of the lateral restraint device were loosened to allow the horizontal restraint device to move with the column’s deformation through its elongated holes, thus preventing excessive deformation and ensuring the subsequent usability of the specimen. After the jack reached the predetermined load, the bolts of the horizontal restraint device were tightened.

During static loading, displacement control was adopted. Before yielding, the displacement increment for each step was 10 mm, with a loading rate of 1 mm∙min⁻¹. After each step, the load was held for 5 minutes. After yielding, the displacement increment for each step was reduced to 5 mm, with a loading rate of 1 mm∙min⁻¹. The load was held for a sufficient amount of time after each step. Loading was stopped, and the test concluded when significant damage occurred in the joint region or the joint could no longer sustain the load.

The loading protocol for cyclic loading was primarily based on reference (JGJ 101-96, 2005) and adjusted to suit the characteristics of this experiment. A force-displacement hybrid control method was adopted. Before reaching the yield load P_y, force control was applied, with each loading step set as a multiple of P_y, and each step was repeated twice. The actual yield load P_y′was determined as the beam-end load at which the stiffness of the strain gauges S17-S20 significantly decreased, and the corresponding yield displacement δ_y was recorded. After yielding, displacement control was used, with each deformation step set as a multiple of δ_y, and each step was repeated twice, as shown in Figure 6.

Figure 6.

The loading protocol for cyclic loading.

Experimental results and analysis

Experimental phenomena and failure modes

The failure mode of J-0 was the outer ring plate fracture, as shown in Figure 7(a). Under cyclic loading, J-4 exhibited misalignment of the grooves, and its failure mode was the slip-out of groove disengagement, as shown in Figure 7(b). The J-6 showed no obvious groove misalignment, and its failure mode was outer ring plate fracture (Figure 7(c)). The failure mode of D-4 was the grooves pulled out from the slot, as shown in Figure 7(d).

Figure 7.

Experimental phenomena.

After removing the cover plate to observe the deformed shape of the grooves, it was found that the grooves of J-4 were worn to a very smooth surface (Figure 7(e)), while those of J-6 had only slight wear (Figure 7(f)). The grooves of D-4 were sheared off and all tilted in the same direction (Figure 7(g)).

Moment–rotation hysteretic curves

To obtain the moment-rotation curve of the specimens, the joint’s rotation angle φ needs to be calculated. Since the rotation area of the joint is mainly concentrated at the connection between the beam and column, the column, outer flange, and welded web plate area are considered as the rigid region without rotation. Therefore, the rotation angle is calculated using the deformation schematic diagram shown in Figure 8 (Garifullin et al., 2017). From the figure, it can be seen that the beam end displacement Δ_mi consists of two parts: (1) the displacement δ_i caused by bending deformation at the joint; and (2) the displacement Δ_i caused by bending deformation of the structural column. Thus, the rotation angle due to bending at the joint can be calculated using equation (2).

φ \approx \frac{δ_{i}}{l_{w}} = \frac{Δ_{mi} - Δ_{i}}{l_{w}}

(2)

Figure 8.

Schematic of joint rotation calculation.

The displacement Δ_mi can be obtained from the LVDTs DT-12 and DT-13 at the beam end. Δ_i is the displacement generated at both ends of the joint region, assuming the joint domain is rigid, and it is obtained through numerical calculation. l_w is the distance from the loading point to the beam-column connection face.

By applying the method, the beam-end moment M and joint’s rotation angle φ were calculated, and the hysteretic curves of the three joints were plotted, as shown in Figure 9. It was observed that in the early stage of loading, the hysteretic curves of all three joints were close to straight lines, with loading and unloading paths roughly overlapping. As the load increased, a significant horizontal slip segment appeared in the hysteretic curve, as shown in Figure 9(a), indicating that the high-strength bolts in joint J-0 began to slip and entered the bearing stage. In the later loading stages, the hysteresis loop exhibited a distinct pinching effect with an inverse S-shape, and the curve was not full.

Figure 9.

Hysteretic curves of the three joints.

Figure 9(b) showed the hysteretic curve of joint J-4. The hysteresis loop remained spindle-shaped and relatively full. In the later stages of loading, as the grooves began to fail, the sliding of the cover plate caused a change in joint stiffness. Groove slippage was evident near a rotation angle φ of 0, leading to fluctuations in the curve. At the end of loading, all grooves on the lower cover plate of the joint failed and slid out of the slots, resulting in a large horizontal slip segment in the curve, which indicated joint failure. The failure occurred during unloading, suggesting that the grooves were subjected to complex forces during cover plate slippage and gradually wore down, rather than failing at the peak shear force.

Figure 9(c) showed the hysteretic curve for joint J-6. It was evident that, until failure, the hysteretic loop remained spindle-shaped and very full, demonstrating that the joint had strong energy dissipation capacity.

Figure 9(d) compared the hysteretic curves of the three joints. It was seen that joints J-4 and J-6, due to the presence of engaging grooves, significantly reduced member slippage under cyclic loading and maintained higher stiffness throughout the loading process. Their hysteretic loops were fuller and showed no pinching effect, indicating stronger energy dissipation capacity. Joint J-6, with more engaging grooves than J-4, exhibited higher rotational stiffness, greater ultimate bearing capacity, and superior energy dissipation performance.

Backbone curves and ductility factor

Based on the hysteretic curves, the backbone curves can be constructed by connecting the peak load points of each displacement loading cycle. Figure 10(a) showed the backbone curves for the three joints. Figure 10(b) showed a comparison between the static monotonic loading curve of D-4 and the backbone curve of the hysteretic loading for J-4, both directions were “Negative”.

Figure 10.

Backbone curves of the three joints.

Ductility refers to the deformation capacity of a structure beyond its elastic limit, without significant degradation in strength and stiffness. It is an important factor influencing the seismic performance of structures. The ductility factor μ is calculated using equation (3).

μ = \frac{δ_{u}}{δ_{y}} = \frac{φ_{u}}{φ_{y}}

(3)

In this equation, δ_u is typically selected as the ultimate displacement of the specimen, δ_y is the yield displacement of the specimen, φ_u is the ultimate rotation of the joint, and φ_y is the yield rotation of the joint.

The methods for determining the yield displacement mainly include the General Moment Yield Method, the Equivalent Elastic-Plastic Yield Method, and the Tangent Stiffness Method. In Figure 11(a), the General Moment Yield Method takes point B as the approximate yield point, where the displacement and load are the yield displacement δ_y and the yield load P_y. The Equivalent Elastic-Plastic Yield Method takes point A as the approximate yield point. In Figure 11(b), the Tangent Stiffness Method takes point B as the approximate yield point.

Figure 11.

The methods for determining δ_y.

In this study, the average values from three calculation methods were adopted as the yield displacement. The ultimate displacement was defined as the displacement corresponding to a 15% drop in peak load. If no load drop occurred, the displacement corresponding to the peak load is used as the ultimate displacement. The final results are shown in Table 2. For the J-0 joint, two failure modes were considered: the friction-type (FT) mode, where the ultimate capacity was defined as the load at which bolts began to slip, and the bearing-type (BT) mode, where bolts continued to bear load after entering the bearing stage.

Table 2.

Calculation results of the test specimens.

Specimen	Direction	M_y/kN·m	φ_y/rad	M_u/kN·m	φ_u/rad	μ	K₀/(kN·m·rad⁻¹)
FT J-0	Positive	181.27	0.07800	66.619	0.02548	1.21	4743.94
FT J-0	Negative	190.15	0.07928	73.506	0.02552	1.19	4415.97
BT J-0	Positive	181.27	0.07800	193.787	0.09418	1.21	4743.94
BT J-0	Negative	190.15	0.07928	200.616	0.09469	1.19	4415.97
J-4	Positive	148.80	0.04233	169.97	0.05429	1.28	4926.28
J-4	Negative	133.12	0.04317	150.89	0.05464	1.27	4536.25
J-6	Positive	160.51	0.03114	185.56	0.05365	1.72	8660.15
J-6	Negative	177.33	0.03169	199.62	0.05296	1.67	8168.00
D-4	Negative	166.82	0.05688	185.614	0.1009	1.74	5099.68

In the table, K₀ is the initial rotational stiffness, defined as the slope from the first point of the backbone curve to the origin. M_y is the yield load, φ_y is the yield angle, M_u is the ultimate load, φ_u is the ultimate angle, and μ is the ductility factor. Based on the results in Table 2, the following comparisons were made:

By comparing the OFBBC joints with the traditional fully bolted joint, it was found that:

(1) The ultimate bearing capacity of J-4 was 120% higher than that of FT J-0. The bolts in J-6 did not slip at failure, indicating higher shear resistance. The ultimate load of FT J-0 was lower than its yield load, leading to premature failure, limited ductility, and poor energy dissipation.

(2) BT J-0 and J-6 shared the same failure mode—fracture at the outer ring plate—and had similar ultimate capacities. However, J-6 had a smaller ultimate displacement, indicating a smaller drift angle and better seismic performance. Its ductility factor was about 40% higher than that of BT J-0.

(3) J-6 had the highest initial rotational stiffness, and J-4 was slightly stiffer than J-0.

Through the above comparison, it can be concluded that the addition of grooves enhances the ultimate bearing capacity of the joint, increases its stiffness, improves its ductility, and provides better plastic deformation capacity, allowing the joint to dissipate more energy during seismic events.

By comparing two OFBBC joints, it was found:

(1) J-6 had a higher bearing capacity, with an ultimate bearing capacity approximately 20% higher than that of J-4.

(2) J-6 yielded earlier with a yield rotation angle smaller than J-4, and its ductility was about 33% higher.

(3) The initial rotational stiffness of J-6 was approximately 75% higher than that of J-4, and the stiffness of J-6 remained consistently greater than that of J-4.

This indicates that the number of grooves affects the load-bearing performance of the joint, with the groove quantity positively correlated with the ultimate bearing capacity, ductility, and rotational stiffness of the joint.

By comparing J-4 with D-4, it was found:

(1) Both had similar initial stiffness, but J-4 degraded faster in later loading.

(2) Comparing with D-4, although both joints exhibited the same failure mode, J-4 had approximately a 23% reduction in ultimate load capacity, an 84% reduction in ultimate displacement, and a 37% reduction in ductility.

This indicates that under cyclic loading, the grooves of the OFBBC joint are more prone to losing shear resistance due to cyclic friction, leading to lower bearing capacity. Therefore, in the seismic design of grooves, it is crucial to incorporate adequate safety redundancy to minimize premature sliding and misalignment caused by friction, thereby reducing the influence of the friction-induced reduction effect.

Stiffness degradation

To evaluate the stiffness degradation of the joints, the secant stiffness K_i under each loading level is calculated using equation (4):

K_{i} = \frac{| + M_{i} | + | - M_{i} |}{| + φ_{i} | + | - φ_{i} |}

(4)

where K_i is the secant stiffness at the i-th loading level; +M_i and −M_i are the maximum positive and negative moments under cyclic loading, and +φ_i and −φ_i are the corresponding rotation angles.

The variation of secant stiffness with the number of loading cycles n for specimens J-0-C, J-4-C, and J-6-C was shown in Figure 12. It was observed that the joint stiffness decreased rapidly in the early loading cycles and more gradually in the later stages. Compared with the OFBBC joints, the J-0 joint exhibited a more significant stiffness degradation due to bolt slippage. The final stiffness reductions were 65% for J-0, 43% for J-4, and 45% for J-6.

Figure 12.

Stiffness degradation curve.

Energy dissipation of the joints

Energy dissipation capacity refers to the ability of a joint to dissipate energy through its own deformation under cyclic loading. The dissipated energy U of the joint is measured by the area enclosed by the hysteretic curves—the larger the area, the more energy is dissipated. The energy dissipation capacity is evaluated using the equivalent viscous damping coefficient ξ where a larger ξ indicates stronger energy dissipation capacity. ξ is calculated using equation (5):

ξ = \frac{1}{2 π} \cdot \frac{U}{S}

(5)

In this equation, S represents the area of the triangle formed by peak point of the hysteresis loop, the corresponding rotation angle on the horizontal axis, and the origin.

Since the FT J-0 joint exhibited insufficient energy dissipation, only the energy dissipation performance of the BT J-0 joint was analyzed. Figure 13 compared U and ξ of the three joints as functions of rotation. J-6, having the highest stiffness and the fullest hysteretic loops, dissipated the most energy at the same rotation. J-4, with slightly lower stiffness, exhibited lower energy dissipation and ξ than J-6. In contrast, J-0 experienced bolt slippage, which led to pinched hysteretic loops and the smallest hysteretic area at the same rotation, resulting in the least energy dissipation. Due to significant slippage, J-0 also had the smallest ξ, indicating the weakest energy dissipation capacity. These results demonstrated that incorporating interlocking grooves not only enhanced the ultimate load-bearing capacity of the joint but also significantly improved its energy dissipation ability, making it more efficient in dissipating seismic energy.

Figure 13.

Energy dissipation of the joints.

Finite-element analysis

Modeling assumptions

The finite element models (FEMs) of the joints were established using the finite element analysis (FEA) software ABAQUS. The geometric dimensions and construction methods were consistent with the experimental specimens described in Section “Design of specimens”. Three FEMs with different groove numbers were developed, as shown in Figure 14(b). Structured meshing was performed using C3D8R elements, with finer mesh near the joint domain and coarser mesh farther from it, as illustrated in Figure 14(c). All bolts in the models were constructed using 3D solid elements C3D8R. Based on experimental results, bolts did not to fail due to tensile forces. Therefore, a simplified dumbbell-shaped solid model was adopted to represent bolts, providing better computational efficiency (Figure 14(d)).

Figure 14.

Finite element models.

For the high-strength bolts, a bilinear elastic–plastic model was used, with an elastic modulus of 200 GPa, a yield strength of 640 MPa, an ultimate tensile strength of 800 MPa, and an ultimate plastic strain of 0.2. Similarly, the Q355B steel was modeled using a bilinear constitutive model based on results from material property tests. The elastic modulus was taken as 200 GPa, the yield strength as 400 MPa, the ultimate tensile strength as 540 MPa, and the corresponding plastic strain as 0.3. The plastic hardening behavior of Q355B steel was defined using the “Combined” hardening option in ABAQUS, with “Cyclic Hardening” enabled in the sub-options. The parameters used in the combined hardening model were adopted with reference to the values reported in Reference (Zhang, 2020).

The contact relationships between steel plates were defined with consistent parameters. The normal behavior was set as “hard” contact, allowing separation after contact. The tangential behavior employed a “penalty” friction formulation with a coefficient of 0.35 while the friction coefficient between high-strength bolt nuts and steel plates was set to 0.05(Liu et al., 2019).

Since the square tube column, outer ring plates, and welded web plate of the specimens were welded connections, the tie constraint of “tie” was applied, as shown in Figure 14(e). In the FEMs, coupling constraints were set at the column top, column bottom, and beam ends. A hinged support was applied at the column top, while a fixed support was applied at the column bottom (Figure 14(f)).

To maintain consistency with the experiment, the loading protocol in the FEMs was identical to that used in the test. Before the loading steps, all M16 bolts were tensioned by the “bolt load”, provided in ABAQUS, of 80 kN, and the bolt length was fixed during the whole loading step. Then an axial force of 560 kN was imposed at the top of the column. Subsequently, cyclic loading was applied at the beam ends, following the same loading protocol as the experiment.

Finite element model validation

Figure 15 showed the comparison of hysteretic curves and backbone curves between experimental results and FEA calculations for the three joints. It was observed that the experimental and FEM curves for all three joints were consistent in trend, shape, and turning points. Table 3 presented a comparison of the test and FEA calculation results. The test and FEA results for the yield moment M_y and ultimate moment M_u of the joints showed errors of less than 15%, indicating that the FEMs had sufficient accuracy and could reliably simulate the experimental results under cyclic loading.

Figure 15.

Comparison of test results and FEA calculations.

Table 3.

Comparison of test results and FEA calculations.

Specimen	Direction	M_y/kN·m			M_u/kN·m
Specimen	Direction	Test	FEA	Error/%	Test	FEA	Error/%
J-0	Positive	181.27	173.92	4.1	193.787	193.44	0.2
J-0	Negative	190.15	163.30	14.1	200.616	183.00	8.8
J-4	Positive	148.80	146.35	1.6	169.97	163.57	3.8
J-4	Negative	133.12	140.90	−5.8	150.89	163.00	−8.0
J-6	Positive	160.51	165.99	−3.4	185.56	185.87	−0.2
J-6	Negative	177.33	159.15	10.3	199.62	185.00	7.3

Stress analysis of the FEMs

Figure 16 illustrated the stress distribution in detailed parts of the models, where gray areas indicated stress exceeding the ultimate stress of 540 MPa. Figure 16(a) showed the stress distribution at the outer ring plate of J-0 and J-6 joints at the point of failure. Through internal force analysis in ABAQUS, it was found that the stress was significantly high at the intersection of the three welds between the column and the outer ring plate, exhibiting a clear stress concentration effect. This caused the internal force of the outer ring plate to far exceed the initial design value. Compared to the beam-column connection, this region experienced a larger bending moment, indicating the need for structural reinforcement.

Figure 16.

Stress distribution in detailed parts of the FEMs.

Figure 16(b) showed the stress distribution of the flange cover plate grooves for the J-4 and J-6 models at failure. In the J-4 model, all four grooves reached the ultimate stress, which could be equivalently considered as fully flattened. In the J-6 model, only the first groove was partially damaged at failure. This indicated that the shear resistance of six grooves surpassed that of four grooves, and also demonstrated that the J-6 joint maintained better stiffness and energy dissipation capacity, with significant shear capacity remaining in the grooves.

Conclusions

This study investigates the seismic performance of a novel occlusive fully bolted beam-column joint. The main conclusions of the study are as follows:

(1) Under cyclic loading, the failure mode of the OFBBC joint was groove wear, resulting in the complete disengagement of the grooves between the flange-cover plate and the beam flange. Compared with traditional high-strength bolted joints, the OFBBC joints with grooves exhibited higher ultimate bearing capacity, greater rotational stiffness, improved ductility, and enhanced energy dissipation capacity. Furthermore,s as the number of grooves increased, these performance parameters also improved.

(2) By comparing the OFBBC joint under static and cyclic loading, it was found that the grooves were more prone to wear under cyclic loading, leading to a reduction in bearing capacity. This phenomenon, referred to as the “reduction effect caused by friction,” resulted in significant reductions in ultimate bearing capacity and ductility.

(3) FEMs of the three types of joints were established and validated using experimental results. Additionally, the stress distribution of the FEMs was analyzed.

Footnotes

ORCID iDs

Xiaonong GUO

Jinhui LUO

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Chen

Niu

Liu

, et al. (2021) Seismic study on an innovative fully-bolted beam-column joint in prefabricated modular steel buildings. Engineering Structures 234: 111875.

Chu

Zhong

Luo

, et al. (2022) Experimental study on the seismic performance of the new fully-bolted beam-column joint. Journal of Constructional Steel Research 199: 107619.

Chu

Chen

Zhong

, et al. (2024) Seismic fragility analysis of steel frames with fully-bolted core tube joints. Advanced Steel Construction 3: 208–221.

Fan

Zhao

Zhu

, et al. (2022) Seismic behavior and analytical model for a fully bolted joint between CFDST columns and steel beams. Structures 42: 515–530.

Garifullin

Pajunen

Mela

, et al. (2017) Initial in-plane rotational stiffness of welded RHS T joints with axial force in main member. Journal of Constructional Steel Research 139: 353–362.

GB/T 228-2010 . (2010) Metallic materials – tensile testing – part 1: method of test at room temperature. (Standard).

GB/T 2975-1998 . (1998) Steel and steel products – location and preparation of samples and test pieces for mechanical testing. (Standard).

Guo

Zhang

Xiong

, et al. (2016) Load-bearing capacity of occlusive high-strength bolt connections. Journal of Constructional Steel Research 127: 1–14.

Guo

Zong

Gao

, et al. (2020) Ductile failure of occlusive high strength bolt connections under shear force. Journal of Constructional Steel Research 168: 105982.

10.

Huang

Zhang

(1995) On disaster of structural steel in great Hanshin quake in Japan. Steel Construction 2: 118–127, (in Chinese).

11.

JGJ 101-96 . (2005) Specification of testing methods for earthquake resistant building. (Standard).

12.

JGJ 82-2011 . (2011) Technical specification for high strength bolt connections of steel structures. (Standard).

13.

Jiang

Guo

, et al. (2016) Experimental research on high strength bolted zigzag-plate connections of aluminum alloy members. Journal of Tongji University 44(9): 1333–1339, (in Chinese).

14.

Korol

Ghobarah

Mourad

(1993) Blind bolting W-shape beams to HSS columns. Journal of Structural Engineering 119(12): 3463–3481.

15.

Jiang

Wang

, et al. (2023) Experimental study on cyclic behavior of blind-bolt joint connecting steel beam and rectangular tube column. Advanced Steel Construction 19(2): 91–101.

16.

Liu

Yang

Wang

, et al. (2017) Seismic performance of H-section beam to HSS column connection in prefabricated structures. Journal of Constructional Steel Research 138: 1–16.

17.

Liu

Cui

Zhan

, et al. (2019) Seismic performance of bolted connection of H-beam to HSS-column with web end-plate. Journal of Constructional Steel Research 156: 167–181.

18.

Liu

Chen

, et al. (2024) Seismic performance of bolted middle joint for modular steel structure. Journal of Constructional Steel Research 220: 108861.

19.

Liu

Lin

Chen

, et al. (2025) Experimental research and numerical simulation of a modular composite steel frame structure. Structures 71: 107914.

20.

Miller

(1998) Lessons learned from the Northridge earthquake. Engineering Structures 20(4–6): 249–260.

21.

Mourad

Ghobarah

Korol

(1995) Dynamic response of hollow section frames with bolted moment connections. Engineering Structures 17(10): 737–748.

22.

Zhan

Liu

Feng

, et al. (2021) Seismic performance of a square HSS column to H-section beam bolted connection with double cover plate. Engineering Structures 231: 111729.

23.

Zhang

(2020) Numerical Simulation of Q345 Steel Ultra-low Cycle Fatigue Fracture Process Based on Cyclic Void Growth Model. Tianjin University. PhD thesis.

24.

Zhang

Liu

, et al. (2022) Seismic performance of joint for H-beam to CFST column with field-bolted flange-splicing. Journal of Constructional Steel Research 196: 107375.

25.

Zhong

Aslani

Feng

, et al. (2024) Mechanical performance study of a novel modular gymnasium inner sleeve all-bolt cross connection joint – part Ⅰ: Experiments and finite element modeling. Thin-Walled Structures 195: 111422.

26.

Zhong

Hui

Zhang

, et al. (2025) Mechanical performance study of a novel modular gymnasium inner sleeve all-bolt cross connection joint – part II: parametric study and design recommendations. Journal of Constructional Steel Research 226: 109211.