Experimental and Numerical Research on the Mechanical Properties of a Novel Prefabricated Diaphragm Wall–Beam Joint

Abstract

Based on the engineering context of prefabricated underground station structures, this paper proposed a novel diaphragm wall–beam joint based on post-poured ultra-high-performance concrete (UHPC) and non-contact lap-spliced steel bars. This research study designed and conducted a full-scale experiment on the diaphragm wall–beam joints. The failure modes, bearing capacity, overall stiffness, crack resistance performance, and force transmission mechanism of the new diaphragm wall–beam joint were investigated. Additionally, a three-dimensional finite element model (FEM) of the wall–beam joint was developed using the software ABAQUS 2020. The model was validated against experimental results and used for further analysis. The results showed that the tensile through-cracks at the UHPC-diaphragm wall interface characterize the final failure process. The proposed UHPC joint could satisfy the structural design requirements in terms of crack resistance and bearing capacity. No rebar pulled-out damage was observed, and the non-contact lap-spliced length of 10d in the UHPC joint was sufficient. Compared with the traditional cast-in-place concrete joint, the cracking moment and yield moment of the proposed UHPC joint increased by 8.7% and 5.4%, respectively.

1. Introduction

Prefabricated structures offer significant advantages over traditional cast-in situ concrete structures in terms of standardized design, factory-based production, assembly-based construction, and contributions to low-carbon development [1,2]. Driven by strong national policies, China’s above-ground prefabricated buildings have experienced rapid development. With the increasing diversity of prefabricated structural systems and prefabricated components, as well as the corresponding improvement of construction methods, prefabricated structures are very popular. Compared to the rapid development of prefabricated buildings on the ground, China’s application of prefabricated technology in underground structures is still in the early stages. The slower adoption can be attributed to several technical and economic challenges unique to underground construction. The prefabricated elements for underground structures are often heavy, requiring specialized equipment for transportation and installation. In addition, underground construction sites often have restricted space for assembly. Underground structures must withstand loads from surrounding soil and water pressure, and the joint connection is more complex than above-ground prefabricated structures. Additionally, underground projects are often one-off or limited in scope, and the investment in manufacturing plants, molds and equipment can be expensive. Despite these challenges, prefabrication holds promise for underground structures. Notably, prefabricated construction technology holds significant potential for application in subway station structures, offering an effective solution to the problem of traffic congestion in large urban areas [3,4].

In the 1980s, the first prefabricated station structures emerged in the former Soviet Union to overcome the shortcomings of cast-in situ concrete construction in cold weather. Prefabricated technology for subway stations has been adopted in Russia, Japan, and France. The early prefabricated station structures typically utilized a complex rectangular system with a precast roof, floor, side wall, column, and beam components. However, the connection joints at the base were formed using cast-in situ concrete. In 2012, a fully prefabricated subway station structure system under open-cut conditions was first proposed in the Changchun Metro Line 2 project [5]. Currently, China has initiated the construction of prefabricated stations in Changchun, Qingdao, Shenzhen, Shanghai, Guangzhou, Harbin, Jinan, Wuxi, and other cities.

The joint connection is the most critical problem to be solved, which affects engineering cost, construction speed, waterproof performance, and overall mechanical properties in prefabricated underground structures [6,7,8]. The research team carried out experimental tests and theoretical calculations to study the mechanical properties of the joint. Yang [9] proposed a new grouted mortise-tenon joint. After the assembly of the precast components is completed, grout is injected into the gap between the mortise and the tenon through the reserved grouting hole, so as to effectively transmit the internal force of the section and limit the interface dislocation. The principle of bending resistance, and the bending and shearing characteristics of the joint were deeply explored [5]. Du [10,11,12] proposed a new prefabricated construction technology for Jinanqiao Station. The station floor was designed as an integral cast-in-place slab, while the columns and sidewalls were precast. The precast column–column joints were connected using embedded grout sleeves. Additionally, Yanmazhuang West Station employed an internally braced support system, integrating permanent structures with temporary supports [13]. The station utilized a combination of precast slabs and post-cast slabs, as well as precast pillars and post-cast temporary pillars [14]. A hybrid construction method combining precast and in situ-cast techniques for constructing large-span underground vaults was adopted in the Wuzhong Road Metro Station in Shanghai [15]. The structural analysis of such an underground vault was conducted. Furthermore, a new U-shaped bar connection technology for sidewall-slab joints was adopted at Nanmen Station in Wuxi. The static and seismic behaviors of the joint were systematically studied through experimental research and numerical simulations [16,17,18].

Diaphragm wall-internal support is a reliable enclosure system for prefabricated underground station structures, especially suitable in water-rich and soft soil geological conditions [19,20]. However, the overlapped enclosure structure has significant negative drawbacks. The prefabricated stations require a wider excavation width than cast-in-place stations due to the dual-wall overlapping design. In addition, the two-wall structure consumed significantly more building materials and construction time. Compared with cast-in situ stations, the construction cost of prefabricated stations is approximately 10–20% higher [21,22,23]. These challenges have hindered the widespread application of prefabricated technology in underground structures.

The single diaphragm scheme was implemented in Shangchong Station in Guangzhou. In this design scheme, the cast-in-place diaphragm wall served dual functions as both the external wall for the enclosure and the sidewall for the station structure. The cast-in-place concrete transverse beam also serves as a temporary support. A novel diaphragm wall-waler beam-strut joint was proposed, incorporating welded steel plates to enhance steel bar connection tolerance capability [24,25,26]. Furthermore, a novel prefabricated diaphragm wall was introduced at Xichong Station. The single-diaphragm wall design enabled the structure to function simultaneously as both the enclosure wall and the sidewall of the underground station. Compared to the dual-wall design, the prefabricated single-wall enclosure structure significantly reduced excavation requirements, minimizing labor and material waste [21,23,27]. The post-cast beam was also utilized as a temporary horizontal internal support and a permanent beam.

Post-cast concrete has been widely adopted in wall–beam joint connection technology. However, the steel bar connection in diaphragm beam–wall joints is labor-intensive and constrained by limited operation space. Additionally, the maintenance process of cast-in-place concrete significantly hinders the construction speed of prefabricated stations. Therefore, it is of great significance to solve the joint connection problem. Ultra-high-performance concrete (UHPC) is a promising material for prefabricated structures owing to its high strength, toughness, and durability [28,29,30,31,32,33]. It is particularly popular in bridge engineering structures due to its excellent mechanical properties. UHPC could provide superior bond strength with reinforcement, narrowing the width of the connection joint [34]. In the joint where the steel bars are densely arranged, UHPC provides an effective solution due to its high fluidity and self-compacting properties without vibration.

A novel diaphragm wall–beam joint utilizing UHPC was proposed in this study. The joint design employs a non-contact lap-splice method to replace traditional reinforcement binding and welding techniques. To reduce the anchorage length, the pre-embedded bars in the precast beam and diaphragm wall were designed in U-shape and L-shape configurations, respectively. The construction process of the precast single diaphragm wall–beam joint is illustrated in Figure 1. The method involves the following steps: Step 1: The slurry trench was constructed, and the precast diaphragm wall was positioned and sunk into place. The foundation pit excavation was performed. Step 2: L-shaped steel bars were post-installed on the diaphragm wall surface using connectors. The precast beam was then positioned and lowered into place. Step 3: The L-shaped rebars and U-shaped rebars were fixed by stirrups. Step 4: UHPC was post-poured into the wall–beam joint to complete the connection.

In this study, two full-scale diaphragm wall-slab specimens were prepared to investigate the mechanical behaviors of the proposed joint. In addition, a three-dimensional finite element model of the diaphragm wall–beam joint was developed using ABAQUS 2020. The model was validated through experimental results and subsequently used to investigate the influence of UHPC–concrete interface adhesion performance. This study aims to solve the diaphragm wall–beam joint connection problem. The research findings can offer valuable theoretical reference and engineering guidance for the structural design of prefabricated underground structures.

2. Experimental Program

2.1. Materials

2.1.1. Steel Bars

The longitudinal steel bars at the prefabricated diaphragm wall–beam joint were selected with Φ28 and Φ32. All the steel reinforcements were of HRB400 strength grade. The steel bar tensile tests were conducted on the universal testing machine with a maximum range of 1000 kN according to GB/T 228.1-2010, as shown in Figure 2a. The mechanical parameters of the steel bars, including elastic modulus, yield strength, and ultimate strength are determined and listed in

Table 1. Mechanical parameters of steel bars.

d	Strength Grade	fy (MPa)	Cov	fu (MPa)	Cov	Es (GPa)	Cov	εy (10−6)	Cov
28	HRB400	436	0.05	602	0.04	201	0.06	2160	0.05
32	HRB400	432	0.06	596	0.05	199	0.07	2140	0.05

Note: d denotes the steel bar diameter; f_y denotes yield tensile strength; f_u denotes ultimate tensile strength; E_s denotes elastic modulus; ε_y denotes yield tensile strain; Cov denotes Coefficient of Variations.

.

2.1.2. Concrete

The concrete grade of both the precast diaphragm wall and the precast beam components is C50 [35]. The concrete adopted in this paper was self-compacting concrete. The compressive strengths of concrete were determined using 100 × 100 × 100 mm³ cube specimens based on the Chinese standard GB/T 50081-2019 [36]. The cubic compressive tests of concrete were conducted on the hydraulic testing machine with a maximum range of 3000 kN, as shown in Figure 2b. The loading rate of concrete was 0.5 MPa/s. The measured compressive strength of concrete was 50.1 MPa after 28 days of maintenance.

2.1.3. UHPC

The UHPC adopted in the joint was UC120. The UHPC mixture consisted of water, powder, and steel fibers [37]. The water-to-cement ratio of UHPC was 0.27. The steel fibers were in a straight shape and the volume content in UHPC was 2.5%. The tensile strength of steel fibers was 2500 MPa. The length and diameter of steel fibers were 11 mm and 0.2 mm, respectively.

Due to the low water-to-cement ratio and high content of steel fibers, a special mixing process was required. First, the UHPC powder was mixed in the mixer for two minutes. Then, water was gradually added in and mixed for three minutes. Finally, dispersed fibers were carefully incorporated and mixed for three minutes until the UHPC appeared to have suitable liquidity. The UHPC cubes were cured for 28 days before testing on the hydraulic testing machine, and the loading rate of UHPC was 1.2 MPa/s [38]. The measured compressive strength of UHPC was 123.1 MPa.

2.2. Experimental Design and Specimen Fabrication

The detailed geometric dimensions and cross-sectional reinforcement of the joint are illustrated in Figure 3. The test specimen was designed in a π-shaped configuration, and the precast diagram wall component was assembled with two precast beam components. Two parallel joint specimens were designed as “QC-1A-1” and “QC-1A-2”. During the loading process, the movement of the diaphragm wall was constrained while the beam was subjected to horizontal load. After testing the joint specimen “QC-1A-1”, the π-shaped specimen was lifted and rotated 180 degrees to enable the testing of the joint specimen “QC-1A-2”.

The geometric dimensions of the precast diaphragm wall and precast beam were 4460 mm × 1500 mm × 600 mm and 1730 mm × 700 mm × 600 mm, respectively. The width of the post-poured UHPC joint was 470 mm. The lap-splice length of reinforcements was 330 mm, approximately 10d of the longitudinal reinforcements. The L-shaped bars were assembled with pre-embedded anchoring bars of the diaphragm wall using connectors. The tensile and compressive anchoring reinforcements in the diaphragm wall were 6Φ32 and 6Φ28, respectively. In the precast beam, the embedded tensile and compressive reinforcements were 4Φ28 + 4Φ32 and 4Φ32 + 4Φ28, respectively. The diameter and spacing of stirrups in the joint were 16 mm and 100 mm, respectively. The stirrups in the precast beam had a diameter of 12 mm and were arranged at 200 mm intervals.

The fabrication progress, curing and transportation of the diaphragm wall–beam joint specimen were presented in Figure 4. The precast diaphragm wall and beam were constructed using C50 self-compacting concrete, and the components were cast in steel molds. Strain gauges with a length of 5 mm were attached to the reinforcements adjacent to the connectors before concrete pouring. These strain gages were protected by epoxy resin before concrete casting. Rebar connectors were embedded inside the diaphragm wall to link the L-shaped bars and embedded anchorage bars. All the concrete components were de-molded 24 h after casting and cured outdoors for 28 days. Meanwhile, concrete cubes were cast in the same batch with the concrete components. To improve the interfacial bonding performance between the UHPC and precast concrete, the interfaces of concrete were roughened. The transverse reinforcement in the non-contact splice region was strengthened to provide sufficient constraints on longitudinal bars [39,40]. The diameter and spacing of stirrups in the joint were 16 mm and 100 mm, respectively. The UHPC joint was de-molded 24 h after casting and cured outdoors for 28 days. The prefabricated diaphragm wall–beam joint specimen was transported to the laboratory for testing after sufficient maintenance.

2.3. Loading and Measurement Scheme

Figure 5 depicts the experiment setup at the China Railway Design Corporation. The loading equipment consists of a three-dimensional loading apparatus and a data acquisition system. The precast beam was subjected to horizontal load to apply deformation and induce joint bending moments. The loading process was managed by a digital operating system.

The 3D loading system provided a loading capacity of 12,000 kN, 10,000 kN, and 5000 kN in the vertical, longitudinal, and horizontal directions, respectively. The operational space of the loading frame measured 9.0 m × 5.0 m × 3.5 m. The loading system met the requirements for full-scale experiments and was capable of capturing the complete failure process.

The reaction wall structure provided the necessary reaction force and stable boundary conditions for the joint specimens. The specimens were placed on the supports and anchored to the ground. The experiment measurement system monitored the force, deflection, strain, crack width, and interface opening of the joint specimen. First, strain gauges with a length of 50 mm were pasted on the surfaces of precast concrete and post-cast UHPC to monitor the strain. Additionally, laser displacement sensors were installed to measure the specimen deformation. All experimental data were integrated into the DH-3823 system. The monotonic loading test was carried out according to the Chinese Standard GB/T 50152-2012 [41]. The load was incrementally increased until the specimen failed. Loading at each level was maintained for 15 min, during which the experimental phenomenon was observed and the cracks were marked. The initial values of cracking moment Mcr and ultimate moment Mu were determined by theoretical calculation according to the Chinese Standard GB/T 50010-2010 [42].

3. Experimental Results

3.1. Experiment Phenomenon and Failure Modes

Figure 6 shows the crack distribution and typical failure modes of the diaphragm wall–beam joints at the final loading stage. At the beginning of the loading process, the first tensile crack occurred on the cross-section interface between the UHPC and diaphragm wall. The bending moment caused by the loading was the largest in the joint specimen. As the load increased, some flexural cracks appeared on the side surface of the joint and extended from the bottom to the top of the joint specimen and gradually formed flexural-shear cracks. Meanwhile, some tensile cracks appeared on the top surface of the diaphragm wall. The tensile crack also appeared in the UHPC, and the width is much smaller than it is in concrete. This can be attributed to the steel fiber bridging effect of UHPC.

Although tensile cracks developed in the beam–wall joints, the crack propagation was slower than the opening of the UHPC–concrete interface. According to durability design in GB/T 50010-2010, the crack width should not exceed 0.2 mm to prevent water ingress. The crack width first appeared at the UHPC–concrete interface. Finally, the joint specimen forms a typical failure mode characterized by the UHPC–concrete interface opening exceeding 1.5 mm. According to Chinese standard GB/T 50152-2012, the joint specimen is failed [41]. Meanwhile, neither the steel bar strain nor specimen deflection reached the threshold for test termination.

3.2. Crack Development

As the applied load increased, cracks developed at the diaphragm wall–beam joint specimen. The joint crack was primarily governed by the opening of the UHPC–concrete interface opening. Figure 7a shows the maximum crack width at the wall–beam joint. The development of the wall–beam joint interface opening was divided into three stages: the initial stage, the development stage, and the accelerated stage. The interface opening initiated at M = 200 kN·m. The beam–wall joints entered the accelerated stage after M > 1050 kN·m. The crack width of the wall–beam joint was calculated according to the Chinese standard GB/T 50010-2010. It showed that the experimental results align well with the theoretical calculations. As the crack width reached 0.2 mm, the experimental joint moment exceeded the serviceability limit state moment Ms = 410 kN·m. The results indicate that the UHPC joint satisfies the structural design requirements. Furthermore, as the crack width reached 1.5 mm, the experimental joint moment exceeded the ultimate limit state moment Mu = 851 kN·m.

3.3. Deflection Curves

The deflection at the loading point was measured using a displacement sensor. Figure 7b shows the joint moment-deflection curves of the wall–beam joint specimens. The moment-deflection curve can be divided into three different stages: the elastic stage, the plastic stage, and the failure stage. The first turning point of the experimental curve corresponds to the joint interface opening and concrete cracking. The second turning point marks the yielding of the tensile reinforcement. It can be observed that the stiffness of the experimental curve is lower than the theoretical results. This discrepancy may be attributed to insufficient boundary constraints of the joint specimen during the test.

3.4. Strain Development

Figure 8 illustrates the strain development of the wall–beam joint specimens with increasing load. Figure 8a shows the tensile strain of the concrete in the joint specimen. A total of five valid strain data were collected. The concrete tensile strain showed a sudden increase due to the crack propagation in the joint after the applied moment exceeded 200 kN·m. It showed that the cracking strain of concrete was approximately 100 με, which was consistent with Chinese standard GB/T 50010-2010. The tensile strain development trend aligns closely with that of crack propagation. Figure 8b shows the compressive strain of concrete in the joint. The compressive strain development increased quickly after the occurrence of joint cracking. The measured maximum compressive strain of concrete was 2500 με. The value is smaller than the ultimate compressive strain 3300 με adopted in GB/T 50010-2010. No concrete crushing phenomenon was observed during the joint test.

Figure 8c depicts the tensile strain of the reinforcements near the rebar connector. The moment in this area was the largest among all the joints. The strain can reflect the actual stress state of the reinforcements before yield and indicate where the plastic area develops in the joint specimens. The beam–wall joint’s interface opening accelerated the rebar tensile strain increase. The measured strain values showed significant dispersion after joint interface opening. In addition, the rebar tensile strain experienced a sudden drop during the final loading stage. Figure 8d shows the rebar strain development in the compression region. The reinforcement compressive strain showed the same tendency of concrete. The strain values were close to that of concrete.

4. Finite Element Analysis

To reveal the failure process and mechanical behaviors of the novel prefabricated diaphragm wall–beam joint, a three-dimensional finite element model (FEM) was established using ABAQUS.

4.1. Material Model

The stress–strain behaviors of concrete and UHPC were modeled using the concrete damage plasticity (CDP) model in ABAQUS. The CDP model has been widely adopted to simulate the nonlinear behaviors of concrete structures under monotonic and cyclic loading conditions [43]. The constitutive model parameters of concrete were calculated using equations recommended by Chinese standard GB/T 50010-2010 [42]. The detailed parameters of the CDP model were determined and summarized in

Table 2. Parameters of the CDP model for concrete and UHPC.

Dilation Angle	Eccentricity	fb0/fc0 (MPa)	K	Viscosity Parameter
35	0.1	1.16	0.667	0.0005

.

A bilinear model was employed to describe the uniaxial mechanical behavior of the reinforcements. The elastic stage was characterized by elastic modulus E and yield strength f_y. After reaching the yield strength, the rebar stress was increased to the ultimate tensile strength fu with a hardening modulus Eh, which was considered as 1% of elastic modulus E. The perfect elastic–plastic model was adopted for the loading platen and supports. The elastic modulus values for the rebars and steel plates were 200 GPa and 210 GPa, respectively.

4.2. Finite Element Model

The boundary condition and details of the finite element model of the diaphragm wall–beam joint are illustrated in Figure 9 and Figure 10. The diaphragm wall, precast beam, UHPC, loading platen, support, and high-strength bolt were modeled using reduced integration elements (C3D8R). The reinforcements were modeled using truss elements (T3D2).

In the finite element model, the boundary and loading conditions of the finite element model were aligned with that of the experimental tests. During the loading tests, the movement of the diaphragm wall was constrained by steel supports, high-strength bolts, and reaction walls. The displacements of the supports, bolts, and the right side of the diaphragm wall were restricted in all directions. The concentrated horizontal load was monotonically applied to the loading platen.

The tie constraint method was used to define the interfacial behavior between the loading platen and precast beam. The loading platen was selected as the master surface while the beam was assigned as the slave surface. Surface-to-surface contact was utilized to describe the interaction behavior between the diaphragm wall and the support. The hard contact was adopted in the normal direction while penalty function contact was adopted in the tangential direction. The interfacial friction coefficient was adopted as 0.5 to characterize the friction between the support and the beam. The steel support was designed as the master surface while the concrete was assigned as the slave surface. The bond slip between the reinforcements and concrete was neglected in this paper. Perfect bond was assumed using the embedded element method.

4.3. Model Validation

Figure 11a presents the tensile damage distribution of concrete at M = 450 kN·m. The cracking model was relatively close to the experimental phenomenon. The cracks initially developed at the UHPC–concrete interface and propagated upward from the joint. Figure 11b depicts the rebar strain distribution at M = 1200 kN·m. The maximum stress occurred in the tensile rebar in the precast diaphragm wall and precast beam. Due to the excellent tensile performance of UHPC, the stress in the steel bars at the joint remained relatively low. The results demonstrated that the proposed joint exhibits excellent load transfer capability.

A mesh sensibility analysis of the finite element model was conducted. Three different mesh sizes (150 mm, 100 mm, and 50 mm) were employed in the numerical calculation to evaluate the influence of mesh size. Figure 12a compares the joint moment-deflection curves obtained according to the numerical simulations with different grid sizes. The UHPC–concrete interfaces were modeled using the tie method. It showed that the simulation results were in good agreement with the overall trend of the experimental data. The numerical specimen exhibited higher stiffness and bearing capacity as the mesh size increased. Balancing accuracy and computing efficiency, a global mesh size of 100 mm was selected for the concrete, UHPC, supports, loading platen, and steel bars.

Figure 12b compared the rebar tensile and concrete compressive strains from the test and finite element model results. Solid geometries depict the test results, and the lines show the numerical simulation. It demonstrated that the finite element model results showed the same trend as that of the experimental results. It was noticed that the rebar strain appeared to be larger than the test result. In the finite model, no interfacial slippage was considered between concrete and reinforcement.

4.4. Interface Behavior Analysis

To investigate the influence of interface behavior, the UHPC–concrete interfaces were modeled using both the tie method and the hard contact method [10]. Figure 13 illustrates the tensile damage distribution and interface opening of the joint. It can be observed that the tensile damage at the interface was not significant. However, the interface opening phenomenon was effectively simulated using the hard contact method.

Figure 14a compares the simulated deflection curves for different friction coefficients (1.0, 0.5, 0.4, 0.3, and 0.2). The results indicate that the simulated stiffness and bearing capacity using the tie method exceeded those from the experimental tests. In other words, the UHPC–concrete interface was over-constrained with the tie method. The numerical analysis using the hard contact method provided more conservation calculation results, and the optimal interface friction coefficient value was 0.5. Figure 14b compares the simulated maximum crack width with experimental results. The results demonstrate that the numerical predictions were over-conservative in estimating the maximum crack width.

4.5. Continuous Concrete Joint

To compare the mechanical properties of prefabricated structures and cast-in-place structures, a continuous concrete joint model was developed and analyzed. The longitudinal reinforcements were continuously arranged. Figure 15 presents the concrete tensile damage and rebar stress of the wall–beam joint with continuous reinforcements. The concrete tensile damage distribution and rebar stress were relatively uniform in the joint region, which was different from the behavior observed in the prefabricated joint specimen. The first tensile crack appeared in the wall–beam interface.

Figure 16 compares the joint moment-deflection curves and strain curves of the joint. The results showed that the UHPC joint exhibits greater stiffness and higher load-bearing capacity. The tensile strain in the reinforcement bars of the continuous joint was higher due to the lower tensile strength of the concrete. In other words, the UHPC joint could effectively bear and transmit joint stress. In the later stages of loading, the tensile strain in the reinforcement bars became very close for both joint types. Additionally, the compressive strain of the concrete in the continuous joint was greater than that in the prefabricated UHPC joint. The results demonstrated that the proposed UHPC joint could meet the requirements of engineering design. Compared with the traditional cast-in-place concrete joint, the cracking moment and yield moment of the proposed UHPC joint increased by 8.7% and 5.4%, respectively.

5. Future Study

In future experimental and numerical simulation, parametric studies will be carried out to study the influence of post-poured UHPC strength, steel bar lap splice length, interface characteristics and stirrup ratio on the joint behaviors.

The seismic behaviors of the UHPC joint will be studied in the future. Full-scale joint specimens will be tested under cyclic loading. The influence of bond degradation and fatigue effects on its long-term performance will be revealed.

It should be noted that this paper studied the mechanical properties of the proposed UHPC joint. The long-term durability of the proposed joint is not considered in this study. In future studies, the authors will conduct the water absorption test to assess the long-term durability of the UHPC joint exposed to sulfate attack, chloride penetration, and groundwater movement [44,45].

6. Conclusions

In this study, a novel UHPC-based diaphragm wall–beam joint was proposed to improve the existing connection method in prefabricated underground construction. To investigate the mechanical behaviors of the proposed joint, two full-scale joint specimens were fabricated and tested. In addition, a three-dimensional finite element model was developed to validate the experimental results. According to the experimental results and numerical simulation, the following conclusions were drawn:

(1) In the full-scale experimental test, the failure modes of the joint specimen QC-1A-1 and QC-1A-2 were similar. As the load increased, the maximum crack of the joint was governed by the opening of the UHPC-diaphragm wall interface. The interface opening exceeded 1.5 mm before the rebar strain and specimen deflection reached their respective limitations.

(2) The cracking patterns and strain distribution in the reinforcement bars indicate that the proposed UHPC joint effectively transfers the load from the precast diaphragm wall to the precast beam. The theoretically calculated deflection and maximum crack width closely match the experimental tests. During the loading tests, no rebar pulled-out damage was observed in the overlapping anchor region, indicating that the lap-spliced length of 10d for steel bars in the UHPC joint is sufficient.

(3) The developed three-dimensional finite element model can reasonably describe the mechanical behaviors of the proposed joint. The numerical simulation results are in good agreement with the experimental data in terms of overall stiffness and bearing capacity. However, the UHPC–concrete interface behavior was overestimated when modeled using the tie method. The cracking moment and yield moment increased by 20.1% and 4.7%, respectively. In contrast, the hard contact method provides more reasonable calculation results with an optimal interface friction coefficient of 0.5.

(4) The proposed UHPC joint exhibits superior stiffness, bearing capacity, and crack resistance compared to concrete joints. The results demonstrated that the proposed UHPC joint could meet the requirements of engineering design. Compared with the traditional cast-in-place concrete joint, the cracking moment and yield moment of the proposed UHPC joint increased by 8.7% and 5.4%, respectively. These research results offer valuable theoretical insights and practical references for the structural design of prefabricated underground structures.