Experimental Investigation of Flexural Performance of Prestressed Precast Hollow-Core Slabs with Foam Inserts and Wet Joints

Abstract

This study introduces a novel prestressed precast hollow-core slab system with integrated polystyrene foam inserts and a spliced design utilizing cast-in-place wet joints, aiming to resolve the inherent trade-off between self-weight reduction and load-bearing capacity in precast floor systems. An experimental investigation was conducted on three individual slabs with varied foam geometries and four jointed slabs with different foam configurations and joint concrete types. All specimens were tested under four-point bending. The experimental program meticulously assessed failure modes, load–displacement characteristics, and load–strain relationships of the proposed slab systems. Results reveal that the hollow-core slabs exhibited failure mechanisms similar to conventional cast-in-place slabs, with cracking initiating in the pure bending region and then propagating along the slab edges. Specifically, specimens with square and circular foam inserts achieved weight reduction rates of 24.57% and 19.78%, respectively. Concurrently, their ultimate loads increased by 32.39% and 46.46% compared to the cast-in-place control. The prestressing tendons remained elastic at ultimate load, confirming that failure was governed by concrete crushing in the compression zone rather than tendon rupture, which represents a ductile failure mode providing sufficient warning prior to collapse. For the jointed specimens, while cracks fully penetrated the foam inserts in the pure bending zone, no cracking occurred at the wet joint interfaces, signifying robust composite action. The load capacity of these jointed specimens surpassed that of the equivalent cast-in-place slab by 19.1% to 50.7%. Based on an evaluation of material cost, structural efficiency, and flexural performance among the tested configurations, the combination of square foam inserts and conventional C40 concrete in the wet joint is recommended. This research provides a critical experimental foundation for the development of lightweight, high-performance precast floor systems in prefabricated concrete construction.

1. Introduction

Prefabricated concrete construction has emerged as a predominant global building methodology, owing to its inherent advantages of factory-based component production, standardized design, rapid on-site assembly, reduced labor demand, and enhanced energy and environmental efficiency [1,2,3]. Among various prefabricated structural systems, the precast concrete frame structure—consisting of precast columns, beams, stairs, and floor slabs—stands out as one of the most widely adopted configurations [4,5]. Within this system, the floor slab functions not only as the primary horizontal load-transferring element but also as a critical diaphragm, ensuring the overall stability and integrity of the structure under both gravity and lateral loads. Consequently, achieving an optimal balance between lightweight design, load-bearing capacity, and construction efficiency in precast floor systems is of paramount importance for the widespread adoption of prefabricated concrete buildings.

The precast composite slab, which comprises a prefabricated bottom layer 50–80 mm thick and a cast-in-place topping layer about 70 mm thick, is currently the most widely used floor system in prefabricated concrete buildings [6,7]. However, this system presents a significant drawback: its total thickness considerably exceeds that of a conventional cast-in-place concrete slab under equivalent span and loading conditions. This increased thickness results in greater structural self-weight and higher material consumption. Furthermore, it adversely impacts the seismic performance and cost-efficiency of the overall building by amplifying inertial forces and reducing the available clear story height [8,9,10]. Consequently, there is a pressing need for innovative precast floor systems that achieve substantial self-weight reduction while maintaining or enhancing structural performance.

Recent studies have increasingly focused on the mechanical performance of prestressed hollow-core slabs and precast lightweight floor systems. Researchers have systematically investigated the structural responses of such members from various aspects, including web-shear failure mechanisms, cross-sectional detailing, and boundary conditions [11,12]. Meanwhile, various innovative wet joint or hybrid joint details for precast slab connections have been proposed and validated for their composite load-transfer behavior under flexural action [13]. Nevertheless, systematic experimental investigations that integrate high-strength prestressing tendons, multiple foam geometries, and optimized wet joints into a unified precast hollow-core slab system, while further comparing the effects of different foam shapes and joint concrete types on flexural performance, remain limited.

To mitigate the excessive self-weight of floor slabs, various alternative systems have been developed, such as precast hollow-core slabs and autoclaved aerated concrete panels [14,15,16]. Cui et al. [14] proposed a composite hollow slab with precast edge beams, integrating it with bolted beam–column joints. Tests demonstrated that the hollow slab reduced structural self-weight while ensuring the seismic performance of the joint zone. Han et al. [15] introduced a precast steel truss frame–prestressed concrete composite hollow slab. Static tests indicated that the truss frame significantly enhanced the load-bearing capacity and stiffness of the slab, leading to the establishment of a corresponding theoretical calculation model. Zhou et al. [16] fabricated precast concrete slabs using wood–wool cement board (WWCB), a lightweight filler derived from industrial waste and plant fibers, and demonstrated through flexural tests that these slabs maintained good mechanical performance under weight reduction conditions. Furthermore, deformation compatibility between floor slabs and the primary frame during seismic actions is crucial for maintaining overall structural integrity [17]. However, the benefits of individual hollow slabs can only be fully realized if their inter-panel connections exhibit reliable force-transfer behavior, as connection failure often dictates overall structural performance [18].

Furthermore, when assembling multiple precast concrete floor slabs, the force-transfer performance of their joints is a critical factor in ensuring the integrity of two-way stressed composite slabs [19,20,21]. This issue is fundamentally a shear-transfer problem, analogous to those studied in steel–concrete composite connections [22]. Zhu et al. [19] proposed a new hollow concrete composite slab with closed joints. This system uses lapped rebar and transverse bars to connect precast bottom slabs without protruding the stressed rebar from the joint. Tests showed that the ultimate deflection and load-bearing capacity of the closed-joint hollow slab were 11% and 3.1% higher than those of the cast-in-place slab, respectively, verifying that a reasonable joint configuration can achieve mechanical performance equivalent to that of a cast-in-place slab. Li et al. [20] conducted full-scale tests on the negative-moment behavior of beam–slab joints that combine prestressed hollow slabs with cast-in-place RC layers. They found that the height of the cast-in-place layer and the degree of prestressing significantly affected the load-bearing capacity and ductility, and recommended a support length of not less than 30 mm and a shear span-to-depth ratio greater than 4 to avoid shear failure. Zhang et al. [21] proposed a new multi-ribbed composite slab system with sleeve-connected wet joints. In this system, sleeve connections embedded in the wet joint join the protruding reinforcement bars from adjacent precast panels, enabling effective force transfer without extensive field welding. It was found that its ultimate load-bearing capacity was only about 7% lower than that of the cast-in-place slab, and there was no strain difference between the sleeve connections and ordinary rebar, indicating that the wet joint has reliable force-transfer performance. While these joint configurations have demonstrated improved performance, the design philosophy of controlled deformation and damage concentration, as explored in recent connection studies [23], offers valuable insights for optimizing slab joints. Beyond structural optimization through geometric design and prestressing techniques, the functional enhancement of concrete materials themselves has garnered increasing research interest in recent years [24]. These advancements indicate that integrating multifunctional concrete materials into precast floor systems presents a promising avenue for developing lightweight, high-performance, and durable building components for future applications.

To address these research gaps, this study proposes a novel precast hollow-core slab system that integrates high-strength prestressing tendons with embedded polystyrene foam inserts, as shown in Figure 1. Two distinct foam geometries—square and circular—are investigated to assess their impact on weight reduction and flexural performance. Furthermore, a cast-in-place wet joint detail is developed for connecting adjacent hollow-core panels. The objective of this detail is to achieve composite action comparable to that of monolithic cast-in-place slabs, as shown in Figure 1b. The experimental program involved fabricating and testing three individual slab specimens—one cast-in-place control and two precast hollow-core slabs with different foam shapes—and four jointed slab specimens—one cast-in-place control and three precast hollow-core slabs with varying foam configurations and joint concrete types. All specimens were subjected to four-point bending. The investigation focuses on analyzing the failure modes, load–displacement responses, and load–strain relationships of the proposed slabs. It also evaluates the effectiveness of the wet joint detail in ensuring force transfer and structural integrity and provides design recommendations regarding the optimal combination of foam geometry and joint concrete type for practical engineering applications. The findings of this study are anticipated to establish a theoretical and experimental foundation for the development of lightweight, high-performance precast floor systems in prefabricated concrete buildings.

2. Specimen Design and Fabrication

2.1. Precast Concrete Hollow Slab

One cast-in-place concrete slab, designated as XJB-1, and two precast concrete hollow-core slabs with prestressing tendons at the bottom and filled with square and circular polystyrene foam boards, designated as PCS-1 and PCS-2, were designed. The dimensions and parameters of all specimens are summarized in

Table 1. Main parameters of specimens PCS.

Specimen	Bottom Rebar Type	D/mm	Shape of Foam Board	Foam Board Dimensions/mm	Thickness/mm	Weight Reduction Rate/%
XJB-1	HRB400	8	-	-	-	-
PCS-1	YL82B	7	square	150	60	24.57
PCS-2	YL82B	7	circular	150	60	19.78

, and the reinforcement details are illustrated in Figure 2. All three specimens had identical external dimensions of 1500 mm in length, 500 mm in width, and 100 mm in thickness, with a concrete strength grade of C40. For specimen XJB-1, 8 mm diameter HRB400 steel bars were used as the longitudinal bottom reinforcement. For specimens PCS-1 and PCS-2, 7 mm diameter YL82B prestressing tendons were applied as the longitudinal bottom reinforcement. All other steel bars, including the top longitudinal and transverse reinforcement in both layers, were 8 mm diameter HRB400 bars with a spacing of 200 mm in both directions. The polystyrene foam boards were positioned at the mid-depth of the slab, embedded within the grids formed by the bottom longitudinal and transverse reinforcement.

Table 1 summarizes the main parameters of the three PCS specimens, including bottom rebar type, diameter, foam shape, foam dimensions, foam thickness, and weight reduction rate relative to the cast-in-place control XJB-1. The weight reduction rates for specimens PCS-1 and PCS-2 were calculated based on the volume of concrete displaced by the embedded foam inserts and were determined to be 24.57% and 19.78%, respectively.

The tensile stress of the prestressing tendons was determined in accordance with the Chinese design code GB 50010-2010 [25], adopting σ_con = 0.75 f_ptk. For the YL82B tendons with a specified tensile strength of 1569 MPa, the corresponding jacking stress is approximately 1177 MPa. This value maximizes the prestressing efficiency within the code limit while remaining well below the yield strength of 1346 MPa, ensuring that the tendons remain elastic from the initial tensioning stage up to the ultimate failure. This expectation is confirmed by the measured strain data presented in Section 4.2.2 and Section 5.2.2. The prestressing losses, including anchorage set, elastic deformation, stress relaxation, and concrete shrinkage and creep, were estimated in accordance with the Chinese design code GB 50010-2010 and the SP Prestressed Hollow-Core Slab Technical Manual (99ZG408). The total estimated prestress loss was approximately 15% of the initial jacking stress, resulting in an effective prestress of approximately 1000 MPa at the serviceability limit state. It should be noted that the present study focuses on comparing the effects of foam geometry and joint concrete type; therefore, a uniform jacking stress was applied to all prestressed specimens. A systematic parametric optimization of the prestressing level is beyond the scope of this paper and is suggested as a direction for future investigation.

The fabrication process for the precast concrete hollow-core slab specimens comprised the following sequential steps. First, a steel mold with internal dimensions of 1500 mm × 500 mm × 100 mm was assembled. Second, for specimen XJB-1, ordinary HRB400 steel bars were arranged at the bottom layer, whereas for specimens PCS-1 and PCS-2, prestressing tendons were placed at the bottom using the pretensioning method, as shown in Figure 3a. Third, for specimens PCS-1 and PCS-2, square and circular polystyrene foam blocks, respectively, were placed in the designated positions, followed by the arrangement of the top layer of ordinary steel bars, as shown in Figure 3b. Fourth, concrete was cast into the mold and adequately vibrated to ensure proper compaction. Finally, the specimens were demolded after 24 h and subsequently cured under standard conditions at a relative humidity above 90% for 28 days.

2.2. Dimensional Design of Precast Concrete Hollow Slab with Joint Connection

One cast-in-place slab, designated as SJXJB-1, and three precast concrete hollow slabs with spliced joints, designated as SJPCS-1 to SJPCS-3, were designed. The dimensions of all spliced slabs were 1500 × 1100 × 100 mm, as shown in Figure 4, and the reinforcement diagram of the board surface is shown in Figure 5. Specifically, a 50 × 50 mm space was reserved at the upper end of one side of each single precast concrete hollow slab to allow for welding with the longitudinal steel bars of the adjacent slab, followed by concrete casting, thereby forming a precast concrete hollow slab splice joint. Moreover, the thickness of the foam boards for specimens SJPCS-1 to SJPCS-3 was 60 mm, and the steel bar arrangement was identical to that of specimen PCS-1. Furthermore, specimens SJPCS-1, SJPCS-2, and SJPCS-3 used square, square, and circular foam boards, respectively, and the post-cast concretes were SFRC40 (C40 concrete with added steel fibers), C40, and SFRC40, respectively. Detailed parameters and configurations of each specimen are presented in

Table 2. Main parameters of SJPCS.

Specimen	Shape of Filling Material	Thickness	Prestressing Steel	Cast-in-Place Concrete
SJXJB-1	-	-	-	-
SJPCS-1	square	60 mm	YB82B	SFRC40
SJPCS-2	square	60 mm	YB82B	C40
SJPCS-3	circular	60 mm	YB82B	SFRC40

.

The fabrication steps of the precast concrete hollow slabs with spliced joints included the following: (1) Based on the fabrication method described in Section 2.1, a single cast-in-place slab and single precast concrete hollow slabs with different foam fillings were fabricated. A 50 mm × 50 mm space was reserved at the upper end of one side of each single precast concrete hollow slab, and the longitudinal steel bars within this space were left exposed, as shown in Figure 6a. (2) The two fabricated precast concrete hollow slabs were aligned at the exposed upper steel bars, and the concrete at the lower parts of the two slabs were vertically placed together to form a butt joint, as illustrated in Figure 6b. (3) The reserved longitudinal steel bars at the top of the slabs were connected by welding to enhance the integrity of the two slabs. Subsequently, C40 or SFRC40 concrete was post-cast at the slab splice joint. After curing under standard conditions for 28 days, the precast concrete hollow slabs with spliced joints were obtained.

3. Test Program

3.1. Material Properties

The mechanical properties of the steel reinforcement were determined according to GB/T 228.1-2021 [26]. For each rebar type, three coupon specimens were tested, and the reported values in

Table 3. Geometric and mechanical properties of reinforcing steel.

Type	D/mm	fy/MPa	SD	fu/MPa	SD	Elongation/%	E/GPa
HRB400	8	443	12.8	607	8.5	26.7	207
YL82B	7	1346	10.5	1569	13.6	4.56	205

Note: f_y denotes the yield strength; f_u denotes the ultimate tensile strength; E denotes the modulus of elasticity.

represent the average of the three tests, with standard deviations (SDs) provided in parentheses. In addition, the mix proportions of the C40 concrete were cement:sand:gravel:fly ash:mineral powder:water = 1000:600:480:40:50:170 (kg/m³). The water-to-cement ratio was 0.41, and the sand ratio was 56%. For the SFRC40 mixture, steel fibers were added at a volume fraction of 1.5% with a water-to-cement ratio of 0.41, a sand ratio of 56%, and a superplasticizer dosage of 0.6%. The polystyrene foam inserts had a density of 60 kg/m³, a compressive strength of 200 kPa, a thermal conductivity of 0.030 W/(m·K), and a water absorption of 2% by volume [27].

3.2. Test Setup

The experimental study on precast concrete hollow slabs and precast concrete hollow slabs with spliced joints was conducted in the Structural Laboratory of Nanchang University. The test setup was a structural mechanics combined loading system, with a maximum test space of 3500 × 750 × 1700 mm and a maximum vertical load capacity of 500 kN, as shown in Figure 7. The test setup mainly consisted of a hydraulic actuator, a loading column, a loading beam, two distribution beams, and two supports. All slab specimens were placed on two supports located 75 mm from each end, and the two distribution beams were positioned 450 mm from the nearest support. The distribution beams were connected to the actuator through the loading beam and loading column, thereby enabling the flexural performance testing of the specimens.

3.3. Loading Setup and Measurement Scheme

3.3.1. Loading Setup

The loading process for all precast concrete hollow slab and precast concrete hollow slab with spliced joint specimens consisted of two stages: preloading and formal loading. During preloading, displacement-controlled loading was applied until the mid-span displacement of the specimen reached 0.5 mm, in order to verify that all equipment functioned properly, after which the readings were reset to their initial values. During formal loading, a displacement-controlled loading method was adopted, and each specimen was continuously loaded until the bearing capacity dropped to 85% of the ultimate load or the specimen could no longer sustain further loading, at which point the test was terminated. The loading rate throughout the entire process was 0.2 mm/min [28].

3.3.2. Measurement Scheme

The measurement items for all precast concrete hollow slab and precast concrete hollow slab with spliced joint specimens included the load and rebar strain at different displacements. Among them, the load during the entire loading process was recorded by the hydraulic jack, while the displacement at key locations and rebar strain were measured as follows:

(1) Displacement measurements were recorded using HAD-YHD-50 linear variable differential transformers (LVDTs) with a range of ±25 mm and a calibration coefficient of 0.005 mm/με. The LVDT signals were acquired by a DH3818Y static strain indicator and data acquisition system, which has a strain range of ±60,000 με, a resolution of 0.1 με, and a strain measurement error of ±(0.5%red ± 3 με). Strains in the steel reinforcement were monitored using BFH120-3AA electrical resistance strain gauges (resistance: 120 Ω; gauge factor: 2.0 ± 1%; grid dimensions: 3 mm × 2 mm). The arrangement of displacement transducers for each specimen is shown in Figure 7. They were installed at the mid-span bottom of the slab and at the top of the two end supports. During installation, the concrete surface at the transducer locations was ground, and horizontal glass plates were placed to ensure the accuracy of the displacement transducer readings. Since the primary objective was to capture mid-span deflection as the key indicator of flexural performance, no transducers were placed beneath the loading points. The support transducers monitored any settlement, allowing accurate calculation of the net mid-span deflection. Under the symmetric loading and support conditions, the maximum deflection occurred at mid-span, making additional measurements at the loading points unnecessary.

(2) Strain measurement. The strain measurement points on the bottom rebar of the precast concrete hollow slab are shown in Figure 8a. Strain measurement points S1 to S5 were arranged on the longitudinal rebar in the pure bending zone, and points S6 to S7 were arranged on the transverse rebar. The strain measurement points on the bottom rebar of the precast concrete hollow slab with spliced joints are shown in Figure 8b,c. Strain measurement points S1 to S4 were arranged on the longitudinal rebar at the mid-span bottom, and points S5 to S8 were arranged on the transverse rebar.

Owing to laboratory constraints, only one specimen was tested for each configuration. To minimize measurement uncertainty, all specimens were cured under standard conditions and tested with the same loading system using calibrated transducers and strain gauges. The loading rate was controlled by a closed-loop computer system to eliminate human operational variability. The observed failure modes, load–displacement trends, and key characteristic values, including cracking and ultimate loads, exhibited good physical regularity and consistency with theoretical expectations, which supports the reliability and representativeness of the reported experimental data.

4. Test Results and Discussion of Precast Concrete Hollow Slab

4.1. Failure Mode and Crack Development

4.1.1. Specimen XJB-1

The failure process of the cast-in-place control specimen XJB-1 is illustrated in Figure 9a. At the initial loading stage, no cracks were observed on the bottom surface, and the load–displacement relationship remained essentially linear, indicating elastic behavior. The first crack appeared at the mid-span bottom within the pure bending zone when the displacement reached 1.81 mm, and the corresponding load was 17.12 kN. This crack propagated along the short-side direction. Subsequently, as the displacement increased to 2.79 mm and the load to 20.40 kN, two additional cracks emerged on the left and right sides of the mid-span region and gradually extended along both short-side directions. With continued loading, multiple cracks developed progressively in the pure bending zone, with an average spacing of approximately 74 mm, and these cracks were generally symmetrically distributed along the longitudinal axis. When the displacement reached 27.60 mm, the specimen exhibited severe damage and could no longer sustain further loading, and the corresponding ultimate load was 36.18 kN. The ratio of mid-span deflection to span length at the ultimate state reached 1/50 [28], confirming that the specimen possessed adequate deformation capacity.

4.1.2. Specimen PCS-1

The failure process of specimen PCS-1, which incorporated bottom prestressing tendons and square foam inserts, is depicted in Figure 9b. During the initial loading stage, the specimen remained in an elastic state without visible cracking. The first crack appeared at the bottom surface within the pure bending zone when the displacement reached 2.58 mm and the load reached 20.41 kN. This crack developed along the short-side direction. Subsequently, additional cracks initiated from the mid-span region and propagated toward the longitudinal edges, with most cracks concentrated in the vicinity of the short-side boundaries of the foam inserts. The ultimate displacement and ultimate load of specimen PCS-1 were 30.13 mm and 47.90 kN, respectively, both of which were considerably higher than those of the cast-in-place control specimen XJB-1. Notably, at the ultimate state, no transverse through-cracks were observed at the foam insert locations within the pure bending zone. Furthermore, both the number and width of cracks were significantly smaller than those observed in specimen XJB-1, indicating that the application of prestressing effectively enhanced the cracking resistance and ductility of the precast hollow-core slab.

4.1.3. Specimen PCS-2

The failure process of specimen PCS-2, which incorporated bottom prestressing tendons and circular foam inserts, is presented in Figure 9c. The first crack emerged at the mid-span bottom surface and propagated along the short-side direction, extending across the foam insert region, when the displacement reached 2.51 mm and the load reached 21.24 kN. With continued loading, cracks progressively developed from the mid-span of the pure bending zone toward the longitudinal edges, exhibiting a similar propagation pattern to that observed in specimens XJB-1 and PCS-1. The ultimate displacement and ultimate load of specimen PCS-2 were 30.56 mm and 52.99 kN, respectively, which were higher than those of specimen PCS-1 and significantly higher than those of specimen XJB-1. Moreover, the number and width of cracks in specimen PCS-2 were smaller than those in specimen XJB-1, confirming that the prestressing tendons effectively controlled crack development. The specimen maintained satisfactory load-bearing capacity throughout the loading process while ensuring adequate deformation capacity.

4.2. Test Result Analysis

4.2.1. Load–Displacement Curve

Figure 10 presents the load–displacement curves of specimens XJB-1, PCS-1, and PCS-2. At the initial stage of loading, all three specimens exhibited linear elastic behavior, and the cast-in-place control specimen XJB-1 displayed a slightly higher initial stiffness than the two precast hollow-core specimens PCS-1 and PCS-2. This difference can be attributed to the replacement of concrete with polystyrene foam inserts in the precast specimens, which reduced the effective cross-sectional area and consequently lowered the flexural stiffness. Additionally, the prestressing tendons in specimens PCS-1 and PCS-2 were subjected to relatively low tensile stress at this stage, and their contribution to the overall stiffness was therefore limited.

Table 4. Bearing force of specimens PCS.

Specimen	Fc/kN	Dc/mm	Fu/kN	Du/mm
XJB-1	17.12	1.81	36.18	27.60
PCS-1	20.41	2.58	47.90	30.13
PCS-2	21.24	2.51	52.99	30.56

Note: F_c denotes the cracking load; D_c denotes the displacement at cracking; F_u denotes the ultimate load; D_u denotes the displacement at ultimate load.

summarizes the cracking load F_c, cracking displacement D_c, ultimate load F_u, and ultimate displacement D_u for all three specimens. The cracking loads of specimens XJB-1, PCS-1, and PCS-2 were 17.12 kN, 20.41 kN, and 21.24 kN, respectively, while their corresponding ultimate loads were 36.18 kN, 47.90 kN, and 52.99 kN. Compared with the cast-in-place control specimen XJB-1, specimens PCS-1 and PCS-2 exhibited increases in ultimate load of 32.39% and 46.46%, respectively, despite achieving weight reductions of 24.57% and 19.78%. These substantial improvements can be primarily attributed to the combined effect of two factors. First, the high-strength prestressing tendons, designated as YL82B with a yield strength of 1346 MPa as listed in Table 3, provided a significantly higher tensile capacity than the ordinary HRB400 steel bars with a yield strength of 443 MPa used in the control specimen. Second, the prestressing force introduced a beneficial compressive stress state at the bottom fiber, which delayed crack initiation and propagation. Notably, although specimen PCS-1 achieved a greater weight reduction than PCS-2, its ultimate load was approximately 9.6% lower. This difference suggests that circular foam inserts, by virtue of their geometry, induce less stress concentration and maintain a more favorable effective concrete area in the tension zone compared to square inserts, thereby contributing to higher load-carrying capacity. Nevertheless, from the perspective of weight efficiency, defined as the load-to-weight ratio, specimen PCS-1 may remain advantageous for applications where self-weight reduction is the primary design criterion.

It is worth further explanation that the circular foam specimen (PCS-2) exhibited a higher ultimate load than the square foam specimen (PCS-1), primarily due to geometric effects on local stress concentration. The sharp corners of square inserts induce localized stress concentration under flexural tension, promoting earlier micro-cracking, whereas circular inserts provide a smoother transition and more uniform stress distribution, while retaining a larger effective concrete area in the tensile zone. However, square inserts achieve a higher concrete replacement rate with a weight reduction of 24.57% compared to 19.78% for circular inserts, and are also more convenient and cost-effective in fabrication. Considering the balance among weight reduction efficiency, mechanical performance, and economy, square inserts still offer superior overall value for practical engineering applications.

4.2.2. Load–Displacement Curve

Figure 11 presents the load–strain relationships of the bottom longitudinal reinforcement at measurement point S2 for specimens XJB-1, PCS-1, and PCS-2. This measurement point, located in the pure bending zone, exhibited the largest strain values among all gauges throughout the loading process and was therefore selected for analysis. The load–strain curves of all three specimens comprised three distinct stages: an initial elastic stage, followed by an elastoplastic transition, and finally a plastic development stage. The maximum strain recorded for the HRB400 steel bar in specimen XJB-1 was 3645 με, which exceeded its yield strain of 2215 με, confirming that the longitudinal reinforcement had yielded at the ultimate state. In contrast, the maximum strains of the prestressing tendons in specimens PCS-1 and PCS-2 were 4390 με and 4601 με, respectively, both of which were substantially lower than the yield strain of 6568 με for the YL82B steel. This observation indicates that the prestressing tendons remained in the elastic range throughout the entire loading process. Consequently, the failure of the prestressed hollow-core slabs was governed by crushing of the concrete in the compression zone rather than by rupture of the tendons, which represents a desirable failure mode providing sufficient warning prior to collapse.

A comparison between specimens PCS-1 and PCS-2 reveals that the strain values of the prestressing tendons at corresponding load levels were remarkably similar, with a maximum difference of less than 5% throughout the entire loading range. This observation indicates that the foam insert shape had a negligible influence on the strain distribution of the prestressing tendons. The primary function of the foam inserts is to displace concrete for weight reduction rather than to participate in load transfer. Consequently, the selection of foam geometry should be based primarily on weight efficiency and load-bearing capacity, as the strain state of the prestressing reinforcement remains largely unaffected by the foam shape.

5. Precast Concrete Hollow Slab with Joint Connection

5.1. Failure Mode and Crack Development

5.1.1. Specimen SJXJB-1

The failure process of the cast-in-place control specimen SJXJB-1 is illustrated in Figure 12. At the initial loading stage, no cracks were observed on the slab bottom, and the specimen exhibited linear elastic behavior. The first crack appeared at the mid-span bottom within the pure bending zone when the displacement reached 1.98 mm and the load reached 28.79 kN, and it propagated along the short-side direction. As the displacement increased, additional cracks developed progressively in the pure bending zone. When the displacement reached 30.19 mm and the load reached 64.12 kN, the specimen reached its ultimate state with severe damage and could no longer sustain further loading. At this stage, the cracks had penetrated through the entire slab thickness, and the deflection-to-span ratio exceeded 1/50, satisfying the deformation requirement specified in GB50010 [25]. These observations confirm that the cast-in-place control specimen possessed adequate load-bearing capacity and deformation capacity, serving as a reliable benchmark for evaluating the performance of the jointed precast hollow-core slabs.

5.1.2. Specimen SJPCS-1

The failure process of specimen SJPCS-1, which incorporated bottom prestressing tendons, square foam inserts, and steel-fiber-reinforced concrete with a strength grade of C40 in the cast-in-place wet joint, is depicted in Figure 13. At the initial loading stage, no cracks were visible on the slab bottom. The first crack appeared in the pure bending zone at the mid-span bottom when the displacement reached 3.09 mm and the load reached 32.63 kN. As the displacement increased, multiple cracks developed progressively in the pure bending zone and propagated along the short-side direction. When these cracks reached the polystyrene foam inserts, no detachment or debonding occurred at the foam–concrete interface, indicating adequate bond between the foam and the surrounding concrete. When the displacement reached 30.05 mm and the load reached 84.53 kN, the specimen attained its ultimate state and could no longer sustain further loading. At this stage, cracks were primarily concentrated in the pure bending zone, and no cracking or separation was observed at the wet joint locations on the slab bottom. This failure mode was essentially identical to that of the cast-in-place control specimen SJXJB-1, confirming that the proposed wet joint detail successfully achieved composite action equivalent to that of a monolithic slab.

5.1.3. Specimen SJPCS-2

The failure process of specimen SJPCS-2, which incorporated bottom prestressing tendons, square foam inserts, and conventional C40 concrete in the cast-in-place wet joint, is presented in Figure 14. At the initial stage, the specimen remained elastic without visible cracks. The first crack appeared at the mid-span bottom when the displacement reached 2.17 mm and the load reached 31.22 kN. With increasing displacement, cracks in the pure bending zone gradually multiplied and propagated along the short-side directions. When the cracks extended to the foam inserts, no debonding occurred at the foam–concrete interface, indicating satisfactory bond performance. When the displacement reached 30.36 mm, the specimen reached its ultimate state with an ultimate load of 76.38 kN. At this stage, the maximum crack width in the pure bending zone reached 1.5 mm, and the cracks had penetrated through the entire slab thickness. Notably, no cracks developed along the longitudinal joint connections on the slab bottom, and the failure mode was essentially consistent with that of specimen SJPCS-1, indicating that the use of conventional C40 concrete in the wet joint, rather than steel-fiber-reinforced concrete, did not alter the failure mechanism or compromise the joint integrity.

5.1.4. Specimen SJPCS-3

The failure process of specimen SJPCS-3, which incorporated bottom prestressing tendons, circular foam inserts, and steel-fiber-reinforced concrete in the cast-in-place wet joint, is illustrated in Figure 15. Cracks began to appear at the mid-span bottom when the displacement reached 2.33 mm and the load reached 36.14 kN. With increasing displacement, cracks continued to develop in the pure bending zone, each propagating from the center toward the two short-side edges, and no debonding occurred when the cracks extended to the foam inserts. When the displacement reached 29.61 mm and the load reached 96.63 kN, the specimen suffered severe damage and could no longer sustain further loading. At this stage, the cracks had penetrated through the entire slab thickness. The failure mode was essentially the same as that of specimens SJXJB-1 and SJPCS-1, further confirming that neither the foam shape nor the joint concrete type significantly influenced the fundamental failure mechanism of the jointed precast hollow-core slabs. The specimen exhibited the highest ultimate load among all tested jointed specimens, demonstrating the superior load-bearing capacity of the configuration combining circular foam inserts with steel-fiber-reinforced concrete in the wet joint.

5.2. Test Result Analysis

5.2.1. Load–Displacement Curve

Figure 16 presents the load–displacement curves of specimens SJXJB-1, SJPCS-1, SJPCS-2, and SJPCS-3, and

Table 5. Ultimate bearing capacity of specimen SJPCS.

Specimen	Fc/kN	Dc/mm	Fu/kN	Du/mm
SJXJB-1	28.79	1.98	64.12	30.19
SJPCS-1	32.63	3.09	84.53	30.05
SJPCS-2	31.22	2.17	76.38	30.36
SJPCS-3	36.14	2.33	96.63	29.61

Note: Fc denotes the cracking load; Dc denotes the displacement at cracking; Fu denotes the ultimate load; Du denotes the displacement at ultimate load.

summarizes the corresponding cracking load F_c, cracking displacement D_c, ultimate load F_u, and ultimate displacement D_u for each specimen. The cast-in-place control specimen SJXJB-1 exhibited a slightly higher initial stiffness than the three jointed precast specimens. This difference can be attributed to the replacement of concrete with polystyrene foam inserts in the precast specimens, which reduced the effective cross-sectional area, and to the relatively low tensile stress in the prestressing tendons at the initial loading stage. The ultimate loads of specimens SJXJB-1, SJPCS-1, SJPCS-2, and SJPCS-3 were 64.12 kN, 84.53 kN, 76.38 kN, and 96.63 kN, respectively. Compared with the cast-in-place control, the jointed precast hollow-core slabs exhibited increases in ultimate load ranging from 19.1% to 50.7%, demonstrating the substantial enhancement in load-bearing capacity achieved by the combination of prestressing tendons and foam inserts.

A comparison between specimens SJPCS-1 and SJPCS-2 reveals the influence of joint concrete type on load-bearing capacity. Specimen SJPCS-1, which incorporated steel-fiber-reinforced concrete in the wet joint, exhibited an ultimate load of 84.53 kN, representing a 10.7% increase over specimen SJPCS-2 with conventional C40 concrete in the joint, which achieved 76.38 kN. This improvement can be attributed to the enhanced tensile strength and crack-bridging effect provided by the steel fibers. However, the wet joint is located in the compression zone of the slab under four-point bending, where tensile stresses are minimal. Consequently, the contribution of steel fibers to the overall flexural capacity is limited, and the observed 10.7% enhancement may not justify the additional material cost of steel-fiber-reinforced concrete in practical applications. This finding suggests that conventional C40 concrete is sufficient for the cast-in-place wet joint in terms of structural performance and cost-effectiveness. From a broader practical implementation perspective, the recommended square foam inserts with C40 concrete in the wet joint achieve a weight reduction of 24.57% while delivering a 19.1% increase in ultimate load for jointed slabs compared with the cast-in-place control. Although circular foam inserts offer higher load capacity, their lower weight reduction efficiency and greater fabrication complexity result in a higher cost per unit weight saved. Thus, the proposed combination of square foam inserts and C40 joint concrete offers the best balance among structural performance, self-weight reduction, construction convenience, and material cost.

5.2.2. Load–Rebar Strain Curve

Figure 17 presents the load–strain relationships of the bottom longitudinal reinforcement at measurement point S1 for specimens SJXJB-1, SJPCS-1, SJPCS-2, and SJPCS-3. This measurement point, located at the mid-span bottom, exhibited the largest strain values among all gauges throughout the loading process and was therefore selected for analysis. The maximum strain recorded for the ordinary steel bar in specimen SJXJB-1 was 4186 με, which exceeded its yield strain of 2215 με, confirming that the longitudinal reinforcement had yielded at the ultimate state. In contrast, the maximum strains of the prestressing tendons in specimens SJPCS-1, SJPCS-2, and SJPCS-3 were 4356 με, 2209 με, and 4196 με, respectively, all of which were substantially lower than the yield strain of 6568 με for the YL82B steel. This observation indicates that the prestressing tendons in all jointed precast specimens remained in the elastic range throughout the entire loading process. Consequently, similar to the individual hollow-core slabs described in Section 4.2.2, the failure of the jointed precast specimens was governed by crushing of the concrete in the compression zone, representing a ductile failure mode that provides sufficient warning prior to collapse.

A notable observation from Figure 17 is the strain variation among the three jointed precast specimens. Specimen SJPCS-2 exhibited a maximum prestressing tendon strain of only 2209 με, which is substantially lower than the values of 4356 με and 4196 με recorded for specimens SJPCS-1 and SJPCS-3, respectively. This difference can be explained by the fact that specimen SJPCS-2 had the lowest ultimate load among the three jointed specimens, as reported in Table 5, and consequently its prestressing tendons were subjected to lower tensile stress at the ultimate state. In contrast, specimens SJPCS-1 and SJPCS-3, which exhibited higher ultimate loads, experienced more pronounced strain development in the prestressing tendons. Nevertheless, the strain values for all three specimens remained below the yield strain of 6568 με, confirming that the prestressing tendons were not fully utilized in terms of their tensile capacity. This observation implies that the load-bearing capacity of the jointed precast hollow-core slabs was governed by the concrete compressive strength rather than by the tensile capacity of the prestressing tendons, suggesting that further increases in load capacity could potentially be achieved by using higher-strength concrete in the slab section.

To clarify the force transfer mechanism across the wet joint, it should be explained that the tensile forces in the bottom zone are transferred through the welded lap splices of protruding longitudinal reinforcement from adjacent panels, while the compressive forces in the top zone are transmitted through the roughened concrete interfaces and the shear key action between the precast panels and the cast-in-place joint concrete. Although the cracks completely penetrated the foam inserts in the pure bending zone, no cracking was observed at the wet joint locations in any of the jointed specimens. This confirms that the stresses developed at the joint remained below its capacity throughout the loading process. The observed behavior validates that the proposed joint detail achieves composite action equivalent to that of a monolithic slab, consistent with the design philosophy of “Stronger connections, weaker slab panels”.

6. Conclusions

To investigate the flexural behavior of precast concrete hollow slabs with joint connections, three precast concrete hollow slab specimens considering the influence of different foam filler shapes were initially designed, followed by four precast concrete hollow slab specimens with joint connections. Experimental studies were subsequently carried out on these specimens. The main conclusions are as follows:

(1) The failure mode of the precast concrete hollow slab specimens equipped with prestressed steel bars and filled with polystyrene foam boards was essentially the same as that of ordinary concrete slabs. Cracks primarily occurred in the pure bending zone between the two loading points, extended across the entire slab width, and were largely symmetric along the longitudinal direction.

(2) Compared to the cast-in-place concrete slab, the weight reduction rates for slabs with square and circular foam boards were 24.57% and 19.78%, respectively. Concurrently, the load-bearing capacity increased by 32.39% and 46.46%, respectively. The prestressed steel bars at the slab bottom remained within the elastic stage throughout the tests. It is recommended to use precast concrete hollow slabs with square polystyrene foam boards in design.

(3) Cracks in the precast concrete hollow slabs with joint connections mainly occurred in the pure bending zone and propagated along the short-side direction and through the slab thickness. The post-cast C40 and SFRC40 concrete had no significant influence on the failure mode. Cracks penetrated through the foam boards, and no cracking occurred at the joint connections, which was essentially consistent with the failure mode of the cast-in-place slab.

(4) The load-bearing capacity and flexural performance of the precast concrete hollow slabs with joint connections were superior to those of the cast-in-place concrete slab. The prestressed steel bars remained in the elastic stage, while the longitudinal rebar in the cast-in-place slab yielded. Considering the cost of steel fiber-reinforced concrete, mechanical performance, and load-bearing capacity, it is recommended to use square foam boards and post-cast ordinary concrete.

(5) The recommendation of C40 concrete for the wet joint is based on the observation that the joint is located in the compression zone under four-point bending, where steel fibers provide limited improvement in ultimate load relative to their additional cost.

This study generates valuable experimental data and summarizes the mechanical behavior rules that can benefit future related research on precast hollow-core slab systems. The conclusions are expected to provide references for the improvement of design specifications for precast concrete floor systems and promote the field application of the proposed slab system in prefabricated concrete buildings. This paper focuses on the experimental investigation of the novel slab system. Moving forward, we will conduct finite element simulations using the validated test data to gain a deeper understanding of its internal load transfer, local stress distribution, crack propagation mechanisms, and parametric effects, ultimately facilitating the development of a comprehensive design methodology.