Investigation of Interface Behavior Between Offshore Pipe Pile and Sand Using a Newly Modified Shearing Apparatus

Abstract

With the rapid development of marine engineering, large−diameter steel pipe piles are increasingly used in infrastructure construction, such as bridges, docks, and offshore wind power projects. Therefore, studying the shear behavior of the sand–steel interface is of great importance. In this study, the traditional vane shear apparatus was improved by utilizing its torsional shear actuator, adding an overlying pressure fixing device, and applying lateral pressure through a compressive spring. The original cross plate was replaced with a cylindrical steel rod to simulate the shear behavior of the large−diameter pile–sand interface under different stress states. Experimental results show that this apparatus effectively solves the problem of soil loss due to the shear gap in both the ring shear and direct shear tests under smooth interface conditions. As the shear rate (2°/min, 4°/min, 6°/min) increased, the peak and residual shear stresses decreased, while the shear stress increased with vertical confinement pressure, accompanied by significant residual stress. As the relative density of sand increased from 27.4% to 72.2%, the shear behavior transitioned from contraction to dilation. Regarding surface roughness, the experiment identified a critical threshold: when roughness is below this threshold, it significantly affects the peak shear strength; when above this threshold, the effect is smaller, and failure shifts to the internal sand body. This study provides valuable insights into the mechanics of the sand–steel interface and contributes to optimizing the foundation design for marine infrastructure.

1. Introduction

As marine engineering construction continues to expand, the dimensions and scale of monopiles used in these projects are also increasing. Large−diameter monopiles are becoming increasingly common in offshore wind power, cross−sea bridges, deep−water wharves, and other major marine infrastructures due to their superior bending resistance, high bearing capacity, and relatively straightforward installation process [1,2,3,4]. In sandy soil environments, the behavior of the pile–soil interface is of particular importance, as axial friction plays a crucial role in resisting external loads. Steel pipe piles, which are typically open−ended, are commonly driven into dense strata, such as dense sand, to serve as foundations [5,6,7,8,9,10]. However, in complex marine environments, these piles may experience potential bearing capacity failures due to a reduction in the strength of the interfacial materials, often leading to pile–soil interface sliding [11,12,13,14,15,16]. Therefore, it is essential to determine the design parameters for large−diameter steel pipe pile foundations in marine environments in a more rational and accurate manner in order to improve both the design precision and the construction quality of these foundations.

Recent investigations have focused on the interfacial shear properties between sand and steel surfaces, considering various influencing factors, such as sand particle size, normal stress, and surface roughness [17,18,19,20,21,22,23]. It has been found that the peak strength of coral sand increases with a higher stress level and relative density but decreases as the fine content increases [24]. Conversely, during cyclic shear under constant normal stiffness, the interfacial strength exhibits rapid degradation, predominantly influenced by the initial particle shape [25]. Under conditions of lower steel surface roughness, volumetric contraction tends to dominate cyclic interfacial shear behavior. However, at higher surface roughness, alternating phases of expansion and contraction are observed, which enhance the interfacial shear strength [26]. Additionally, the surface roughness of steel plates significantly affects the shear behavior between calcareous sand and steel. Peak shear strength has been shown to increase with greater surface roughness, with the interfacial friction angle reaching up to 90% of the internal friction angle of calcareous sand for various grain sizes [27]. However, although the interactions between steel pipe piles and sand have been investigated in previous studies, no research has specifically focused on the interaction between large−diameter steel pipe piles and sand. Furthermore, there is still a lack of systematic analysis regarding their behavior, particularly in complex marine environments.

The commonly used methods for studying the sand–steel interface shear behavior include direct shear tests [28], ring shear tests [29], and triaxial tests [30]. The direct shear test is simple to operate, standardized, and cost−effective; however, it is limited by small shear displacements at the interface. When the displacement increases, eccentric loading or even overturning may occur, and the contact area decreases, making it unsuitable for testing interface characteristics at large shear displacements. The ring shear test, on the other hand, can accommodate large shear displacements at the interface, but it is relatively expensive, and the testing procedure is more complex, limiting its widespread use. Like the direct shear test, the ring shear test requires a shear gap to be pre−formed, and for soft soil specimens, significant damage may occur, necessitating the adjustment of the shear gap width, which adds operational difficulty. The triaxial test does not require a pre−formed shear gap and can be used to study interface coupling under complex stress conditions; however, its testing cost is high, similar to that of the ring shear test.

Moreover, studies on large−diameter steel pipe piles have predominantly focused on their stress responses and construction techniques [31,32,33,34,35,36]. For instance, Li et al. [37] employed finite element analysis to investigate the effect of scour−induced unloading on the lateral response of a large−diameter monopile embedded in dense sand. Kong et al. [38] derived an energy transformation equation for pipe piles during sliding by combining working and dynamic principles, which enabled the calculation of the sliding length and spacing of the piles. Zhu et al. [39] conducted axial compression tests on a 550 mm large−diameter, high−strength steel pipe filled with C30 concrete using a 40,000 kN press, determining an ultimate load capacity of approximately 30,000 kN. Heins et al. [40] explored the influence of pile installation methods and sand relative density on both dynamic and static load tests through centrifugal testing.

The construction site of an offshore wind power project is characterized by a thick soft soil layer, with large−diameter steel pipe piles extending more than 60 m into the soil [41]. A typical loading scenario for offshore, large−diameter steel pipe piles is depicted in Figure 1. The primary loads include wind, waves, currents, self−weight, buoyancy, and equipment loads [42,43,44]. Among these, wind, waves, and currents constitute horizontal loads [45,46,47,48], while self−weight, buoyancy, and equipment loads are predominantly vertical loads [49,50,51]. The loads in Figure 1 are applied at the connection point between the pile and the turbine central axis, including a vertical load of 6 MN and a horizontal load of 2 MN. Additionally, the turbine blades generate a substantial moment of 110 MN·m at the connection point, primarily due to wind loads acting at the turbine height. The water depth further amplifies the horizontal load, leading to an increased moment at the mudline. Consequently, understanding the interaction between oversized marine steel pipe piles and the surrounding soil is critical for the subsequent design and long−term safety assessments.

Building upon the limitations of previous studies, this research presents an improved version of the traditional cross−shaped shear test apparatus by replacing the conventional cross−shaped plate with a circular steel bar. This modification allows for a more accurate simulation of the shear behavior at the interface between offshore large−diameter steel pipe piles and sandy soil. The primary aim of this study is to investigate the mechanical properties of the sand–steel interface, with a particular focus on the relationship between shear stress and relative displacement, as well as the shear dilation characteristics. In addition, this study examines the effects of the shear rate, relative density, confining pressure, and surface roughness of the steel on the shear behavior at the sand–steel interface. By utilizing the improved experimental setup, this study addresses the limitations of the traditional cross−shaped shear test, offering a novel approach for examining the mechanical response of the sand–steel interface in offshore structures. Furthermore, the findings provide valuable insights for optimizing pile design and enhancing the stability of engineering structures, particularly regarding the variation in shear strength and behavior under different conditions.

2. Interaction Behavior of Pipe Pile–Sand

2.1. A Modified Apparatus for the Experimental Tests of the Interface

Due to the inability of the original indoor cross−plate shear apparatus to apply confining pressure and measure the shear strength parameters of standard sand, modifications were made to the device for this experiment. The modification fully utilized the torsional shear mechanism, allowing for soil–structure interaction tests under vertical normal stress conditions. The modified laboratory vane shear apparatus is shown in Figure 2. The modification was based on the original sample container, but several additional components were incorporated to improve functionality. These components include an extended threaded rod, a spring base circular steel plate, a spring, a top support circular steel plate for the spring, a dial gauge with an extended probe, a nut, and a round steel rod. This improved setup enhanced the accuracy and applicability of the device for the intended tests.

(1) Extended Threaded Rod: The extended threaded rod has three pieces, each with a diameter of 8 mm and a total length of 270 mm. These rods are used to fix the sample container and the top of the spring to support the round steel plate. (2) Spring Base Plate: The plate has a thickness of 5 mm and a diameter of 99 mm, with a 6.5 mm diameter hole at the center to accommodate the vertical rod. Around the center, at a radius of 25 mm, four 4 mm diameter steel rods, each 70 mm high, are welded perpendicularly to the plate’s surface. These rods serve as supports for the spring. The entire plate is used to apply compression to a standard sand specimen, ensuring that the spring pressure is uniformly distributed across the specimen. (3) Springs: Two types of springs are used, with four pieces of each type. The first type has a wire diameter of 2 mm, an outer diameter of 15 mm, and a height of 50 mm. Its strength coefficient is K = 8.7702 N/mm. The second type has a wire diameter of 3 mm, an outer diameter of 15 mm, and a height of 50 mm. Its strength coefficient is K = 61.2980 N/mm. These springs are employed to apply pressure around the dry sand specimen. (4) Spring Top Support Round Steel Plate: The plate has a thickness of 5 mm and a diameter of 160 mm, with a 6.5 mm diameter hole at the center for the vertical rod to pass through. Around the center, at a radius of 25 mm, four 5 mm diameter holes are drilled to accommodate the four steel rods from the spring base plate. These holes facilitate the compression of the spring. Additionally, at a radius of 70 mm, three 10 mm diameter holes are evenly spaced around the plate. These holes are designed to allow passage of the displacement gauge stylus, enabling measurement of the vertical displacement of the standard sand specimen. (5) Extended Stylus of the Meter: The stylus has a length of 100 mm and a diameter of 4 mm. It is used to pass through the spring top support round steel plate and make contact with the spring base round steel plate in order to measure the vertical displacement of the standard sand specimen during the shear process. (6) Nuts: Twelve nuts in total are used, each with an inner diameter of 8 mm. Three nuts are used to secure the extended threaded rods to the base of the instrument. Three more nuts are used to fix the sample container to the instrument’s base, while the remaining six nuts are used to secure the top of the spring to the round steel plate in its original position.

The torque measured by the cruciform vane shear apparatus consists of two components: one primarily generated by the cruciform vane itself and a smaller contribution produced by the penetrating vertical rod that extends into the specimen. While this secondary contribution is relatively minor, it is important to minimize testing errors by conducting an idle rotation test of the penetrating rod under the same conditions after each cruciform vane shear test (i.e., rotation without the cruciform vane attached). The data from this idle test should be recorded, and during the subsequent data analysis, the torque generated by the penetrating rod should be subtracted from the total measured torque. This ensures more accurate results by isolating the torque contribution from the cruciform vane alone.

The torque generated by the cross plate can be divided into two distinct components: (1) the torque generated by the upper and lower circular surfaces resulting from the rotation of the cross plate, denoted as $M_1$, and (2) the torque generated by the sides of the cross plate acting against the center of the circular cross−section, denoted as $M_2$. A schematic diagram illustrating the torque generated by the cross plate is shown in Figure 1b. Assuming that the maximum torque generated by the cross plate during shear failure is denoted as $M_{\max }$ (in N−cm), this can be expressed as follows [52,53]:

$$M_{\max }=M_1+M_2$$

(1)

where $M_1$ and $M_2$ are expressed as follows:

$$M_1=2\times {\displaystyle \frac{\mathsf{\pi }{D}^2}{4}}\times {\displaystyle \frac{D}{3}}{\tau }_{fh}$$

(2)

$$M_2=\mathsf{\pi }DH{\displaystyle \frac{D}{2}}{\tau }_{fv}$$

(3)

where ${\tau }_{fh}$ is shear strength on the horizontal plane, $\mathrm{kPa}$; $D$ is the diameter of the cross plate; $H$ is the height of the cross board; and ${\tau }_{fv}$ is the shear strength on the vertical plane, $\mathrm{kPa}$. Assuming that the soil is an isotropic, ${\tau }_{fh}={\tau }_{fv}={\tau }_f$, where ${\tau }_f$ is the cross−plate shear strength:

$${\tau }_f={\displaystyle \frac{6H}{\mathsf{\pi }{D}^{2}H(3H+D)}}{M}_{\max }$$

(4)

To directly convert the measured torque into shear stress and facilitate the generation of stress–strain curves, a conversion factor k is introduced to relate torque to shear stress, as follows [54]:

$$k={\displaystyle \frac{6H}{\mathsf{\pi }{D}^{2}H(3H+D)}}$$

(5)

$$C=k\cdot \epsilon \cdot R_y$$

(6)

where $C$ is the shear strength, $\epsilon$ is the calibration coefficient of cross plate head sensor, and $R_y$ is the maximum micro−strain value when undisturbed soil loss reduction occurs. The k values for relating shear stress to torque were calculated for the three different sizes of cross plates, and the results are presented in

Table 1. Different cross plate conversion factors.

Types	K (m−3)
10 mm × 20 mm	2.728
15 mm × 30 mm	0.808
20 mm × 40 mm	0.341

.

We conducted shear tests at the sand–steel interface using a modified indoor cross−plate shear apparatus. Vertical pressure was applied on top of the sand by compressing four springs. The total spring stiffness is the sum of the stiffnesses of each individual spring. To calibrate the spring’s strength coefficient, various methods can be employed, such as tensile or compression tests. Regardless of the method used, all of these approaches rely on the same underlying principle, which is based on Hooke’s law [55].

$$F=k\times \Delta x$$

(7)

where F is the tensile force or pressure, $k$ is the spring’s coefficient of strength, and $\Delta x$ is the amount of change in tension or compression.

The spring stiffness coefficient in this study was calibrated using a compaction device with a ring cutter sample from a triaxial test. The pressure was applied using several 5.1 kg weights, and the compression was measured using a vernier caliper. The experiment followed these steps: ① Label the springs (numbered 1 to 4) for easy data recording. Measure and record the original length of each spring using a vernier caliper, with an accuracy of 0.01 mm. ② Weigh the mass of the spring support (a cylindrical steel rod), which serves to evenly distribute the pressure from the weights to the spring. Use two models: Model 1, weighing 0.3141 kg, and Model 2, weighing 0.3073 kg. ③ Insert the spring onto the central vertical rod of the compaction device and place the support on top to begin the experiment. ④ Gently place the 5.1 kg weights onto the upper support, increasing the number of weights in multiples. After each weight is added, record the total mass of the weights and support, and measure the compression of the spring using the vernier caliper, accurate to 0.01 mm. ⑤ Compile and analyze all experimental data to assess the reliability of the results. Exclude large errors, and for springs with unstable data, conduct additional sets of parallel experiments for verification. Finally, take the average value to determine the spring’s stiffness coefficient. Each spring underwent three sets of parallel tests. The experimental data used to calibrate the spring stiffness coefficient are presented in

Table 2. Spring with a wire diameter of 2 mm.

Number	Weight Quality (kg)	F (N)	△x (mm)	K (N/mm)	Average K (N/mm)
1	10.2	105.07	12.56	8.3657	8.39
	15.3	156.07	18.70	8.3461
	20.4	207.07	24.52	8.4551
2	10.2	105.07	11.52	9.1209	9.04
	15.3	156.07	17.16	9.0952
	20.4	207.07	23.22	8.9179
3	10.2	105.07	11.88	8.8445	8.84
	15.3	156.07	17.90	8.7192
	20.4	207.07	23.10	8.9642
4	10.2	105.07	11.72	8.9653	8.81
	15.3	156.07	17.82	8.7583
	20.4	207.07	23.80	8.7005
K	K = (K1 + K2 + K3 + K4)/4 = 8.7702 N/mm

and

Table 3. Spring with a wire diameter of 3 mm.

Number	Weight Quality (kg)	F (N)		△x (mm)		K (N/mm)		Average K (N/mm)
1	10.2	105.14		1.68		62.5839		62.86
	20.4	207.14		3.18		65.14
	30.6	309.14		5.08		60.85
2	10.2	105.14		1.62		64.90		63.69
	20.4	207.14		3.22		64.33
	30.6	309.14		5.00		61.83
3	10.2	105.07		2.04		51.51		59.04	59.22
	20.4	207.14	156.07	3.48	3.58	59.52	57.84
	30.6	309.14	207.07	5.28	5.10	58.55	60.60
4	10.2	105.14	105.07	2.00	1.62	52.57	64.86	56.14	62.88
	20.4	207.14	156.07	3.56	3.34	58.19	61.99
	30.6	309.14	207.07	5.36	5.00	57.68	61.81
K		K = (K1 + K2 + K3 + K4)/4 = 61.2980 N/mm

Note: Spring 3 and Spring 4 were tested in parallel with the upper support of No. 1 and No. 2, respectively.

.

2.2. Test Material

The standard sand used in the experiments was dry, with particle sizes ranging between 0.25 mm and 0.5 mm, a shape factor of 0.6, and a mineralogical composition in which the silica (SiO₂) content exceeds 96%. In compliance with the Geotechnical Test Method Standard (GB/T 50123−1999), the selected sand met the criterion of a maximum particle size of 5 mm, with particles ranging from 2 mm to 5 mm constituting no more than 15% of the total sample mass. As a result, the testing could be conducted using the methods recommended by the standard. The minimum dry density of the standard sand was determined to be 1.50 g/cm³, while the maximum dry density was found to be 1.76 g/cm³.

To determine the relative density of the sand, both the maximum and minimum dry densities were measured. The minimum dry density is typically measured using either the funnel method or the cylinder method, whereas the maximum dry density is determined using the vibrating hammer method. It is essential that both the minimum and maximum dry density tests are performed simultaneously, with the difference in density between the two measurements not exceeding 0.03 g/cm³. The average of the two measurements was taken as the final dry density. The results from the minimum dry density test for sand are provided in

Table 4. Minimum dry density of standard sand.

Test	Mass (g)	Volume A (mL)	Volume B (mL)	Volume C (mL)	Dry Density (g/cm3)
Test 1	700	455	468	468	1.496
Test 2	700	457	465	465	1.505
Minimum dry density	ρdmin = (ρ1 + ρ2)/2 = 1.50 g/cm3

and

Table 5. Maximum dry density of standard sand.

Test	Mass (g)	Volume (mL)	Dry Density (g/cm3)
1	439.03	250	1.756
2	438.52	250	1.754
Maximum dry density	ρdmax = (ρ1 + ρ2)/2 = 1.76 g/cm3

. Specifically, the minimum dry density of the standard sand is 1.50 g/cm³, the maximum dry density is 1.76 g/cm³, and the internal friction angle is 30°.

In this study, the sand–steel interface test was conducted by replacing the conventional cross plate with a cylindrical steel rod to perform torsional shear tests. Four cylindrical steel rods with different surface roughness values (20 mm in diameter and 40 mm in height) were fabricated for the experiment (Figure 3). The surface roughness was modified by machining uniform rectangular grooves on the rods’ surfaces, resulting in varying levels of roughness as follows: type 1, a smooth surface, hereafter referred to as the smooth steel structure interface; type 2, rectangular grooves with dimensions of 1 mm by 1 mm, spaced at 30° intervals along the circumference, creating a total of 12 vertical grooves; type 3, rectangular grooves with dimensions of 2 mm by 2 mm, spaced at 60° intervals along the circumference, resulting in a total of 6 vertical grooves; and type 4, rectangular grooves with dimensions of 2 mm by 2 mm, spaced at 30° intervals along the circumference, yielding a total of 12 vertical grooves.

2.3. Test Program

A series of interface shear tests was conducted to investigate the effects of the shear rate on various factors, including peak strength, residual strength, the shear stress–relative displacement relationship, and shear−induced dilation or contraction at the sand–steel interface. The tests involved three different shear rates: 2 degrees per minute, 4 degrees per minute, and 6 degrees per minute. These rates were applied between the standard sand and smooth steel surfaces, with a constant normal stress of 50 kPa maintained by a spring mechanism. Throughout the testing, the relative density of the standard sand specimens was kept constant at 72.2%.

To ensure the accuracy of the experiment, it was crucial to determine an appropriate shear rate. According to the British testing standard BS 1377−7, torque was applied to the vane by rotating the torsional head at a speed ranging from 6°/min to 12°/min until the soil was fully sheared. However, the shear rate range of the instrument used in this study was limited to 0.6°/min to 6°/min. Therefore, this experiment was divided into two parts: first, a comparative analysis of different shear rates was conducted through interface shear tests to investigate how the shear rate affects the mechanical properties of the interface; second, aside from the shear rate comparison tests, all other interface shear tests were performed at a shear rate within the standard−recommended range, specifically, at 6°/min. The shear rate of the indoor vane shear apparatus was adjusted by rotating the speed-setting dial on the front panel of the main unit. Figure 4 illustrates the corresponding relationship between the vertical rod rotation speed and the numerical values on the speed-setting dial.

In order to study the effect of the relative density on the peak strength, residual strength, shear stress–relative displacement relationship curve at the sand–steel interface, and shear expansion (shrinkage) of sand and soil, the relative density of the standard sand was taken as 27.45%, 50.98%, and 72.2%, representing three compactness states of loose (15–35%), medium (35–65%), and compact (65–85%), respectively. Then, the sand and smooth steel structures were subjected to interfacial shear tests with a spring−applied vertical pressure of 50 kPa and a shear rate of 6°/min.

To investigate the effect of roughness on the peak shear stress, residual shear stress, shear stress–relative displacement relationship curve of the sand–steel interface, and shear dilation (or contraction) of sand, four types of round steel rods with varying roughness values were fabricated. Shear tests were conducted at the interface of the standard sand and steel structures under different peripheral pressures. This study also aimed to establish the variation patterns in normal stress and shear strength in smooth sand–steel interfaces and to confirm the existence of critical roughness. The vertical pressure applied by the spring was 0 kPa, 25 kPa, 50 kPa, 75 kPa, and 100 kPa, and the shear rate was 6°/min. The specific test program arrangement is shown in

Table 6. Test design for standard sand–steel interface shear experiments.

Experiment Scheme			Contact Surface Type	Φ (°)	Vertical Pressure (kPa)
Shear rate	2°/min	0.349 mm/min	Type 1 round steel bars	30	50
	4°/min	0.698 mm/min
	6°/min	1.047 mm/min
Relative density	27.4%		Type 1 round steel bars	30	50
	51.0%
	72.2%
Roughness	Type 1		Round steel bars	30	0/25/50/75/100
	Type 2
	Type 3				25
	Type 4				25
					25

below.

The modified interface testing device differs significantly from the original apparatus. Taking the shear test of the interface between a smooth steel rod and a 72.2% standard sand specimen as an example, with a vertical pressure of 50 kPa and a shear rate of 6°/min, the operational steps were as follows: ① Loosen the vertical rod fixing clamp and the penetrating hole fixing screws, raise the vertical rod slightly to avoid interfering with the sample loading, and then retighten the vertical rod fixing clamp.② Place the sample container on the base and fix it with screws to begin the sample loading. ③ Load the standard sand specimen in three layers. After each layer is added, use a steel ruler to roughly level the sand surface, tap the container’s sidewalls to loosen the sand, and then use the spring−loaded circular steel plate to vibrate and compact the sand. Finally, measure the height of each layer with a steel ruler. The height of each sand layer is 53.3 mm, giving a total sample height of 160 mm. Pre−calculate the mass of the standard sand specimen. The specimen has a diameter of 99.24 mm, a height of 160 mm, and a volume of 1,237,608 mm³. With a relative density of 72.2%, the dry density is 1.684 g/cm³. Thus, the total mass of the specimen is 2084.28 g. Load the first layer with 694.76 g, the second layer with 673.28 g (after subtracting the standard sand mass corresponding to the volume occupied by the second layer’s steel rod and 6.6 mm penetrating vertical rod, 21.48 g), and the third layer with 692.22 g (after subtracting the volume occupied by the 53.4 mm penetrating vertical rod, 2.54 g). ④ After the first layer of standard sand is loaded, loosen the vertical rod fixing clamp, gently lower the vertical rod, and pass it through the top spring support steel plate, the spring assembly, and the bottom spring support steel plate. Finally, attach the smooth steel rod, tighten it, and insert it into the container, ensuring the smooth steel rod’s center is 80 mm from the container’s bottom (measured directly, the top of the vertical rod is 107.7 mm above the torque transmission arm’s upper surface, which can be precisely located using a vernier caliper). ⑤ Loosen the torque transmission rod fixing screw, allow the torque transmission arm and the torque receiver to gently touch, and then retighten the torque transmission rod fixing screw. ⑥ Continue loading the second and third layers of the standard sand specimen, following the procedure in step ③. Due to limited space at the top of the sample container, load the sand by raising the spring base circular steel plate with small iron blocks or screws, leaving a gap, and using paper trays to pour the sand in. ⑦ After the specimen is loaded, apply confining pressure by compressing the springs. The springs used have a stiffness coefficient of 61.2980 N/mm. From the calculations, the compression of each spring is 1.577 mm. Tighten the top three nuts on the springs to achieve the specified compression and ensure the spring base circular steel plate and the top spring support steel plate are level, ensuring uniform load transmission. Finally, tighten the three nuts on the second layer to fix the position of the top spring support steel plate. ⑧ Fix a dial indicator with an extended contact probe on the universal magnetic lever base and place it on the base of the indoor cross−shear apparatus. Ensure the contact probe of the dial indicator passes through the hole around the top spring support steel plate and contacts the bottom spring support steel plate. Activate the magnetic switch to fix the universal magnetic lever base to the base and record the reading on the dial indicator. ⑨ Set the shear rate to 6°/min. ⑩ The maximum torque displayed on the instrument’s screen is 98.07 N, and the arm length is 5 cm, so the maximum torque value displayed on the screen is 490.4 N·cm. As the goal of this test is to plot the shear stress–relative displacement relationship curve, use only the instantaneous torque, not the maximum torque. ⑪ During the test, record data every 10 s for the first 20 min, every 15 s for the next 20 min, and every 30 s after 40 min. ⑫ Stop the test once the data no longer change or the rotation angle reaches 360°.

3. Results

3.1. Effect of the Shearing Rate

In this study, interfacial shear tests were performed between standard sand and a smooth steel surface under three different shear rates: 2° per minute, 4° per minute, and 6° per minute. A constant vertical pressure of 50 kPa was uniformly applied throughout the tests. The experiments were conducted using a modified indoor cross−plate shear apparatus to investigate the interfacial behavior. The resulting curves, which depict the relationship between interfacial shear stress and relative displacement, were obtained for smooth interfaces at the three shear rates. These shear rates correspond to displacement rates of 0.35 mm per minute, 0.7 mm per minute, and 1.05 mm per minute, respectively. The results are presented in Figure 5.

The shear rate has a significant influence on the relationship between shear stress (τ) and relative displacement (δ) at the interface between standard sand and a smooth steel structure. However, this effect differs from the intrinsic shear behavior of the soil itself. Typically, when the shear rate increases, there is a corresponding rise in peak shear stress, which is accompanied by an increase in the associated relative displacement. For example, values of 28.53 kPa, 25.341 kPa, and 22.119 kPa are observed at different shear rates. In contrast, during the interface shear tests between standard sand and a smooth steel structure, the peak shear stress at the sand–steel interface decreases as the shear rate increases. Furthermore, the relative displacement at which the peak shear stress occurs is very close to the residual shear stress, and it exhibits a similar trend in response to variations in the shear rate.

At the interface between standard sand and a smooth steel structure, the shear stress–relative displacement (τ−δ) curves under the three applied shear rates show distinct peak values, indicating strain−softening behavior. Notably, the magnitude of these peaks diminishes slightly as the shear rate increases. The τ−δ relationship for this interface under varying shear rates demonstrates an elastic–plastic characteristic that can generally be divided into three distinct phases: an elastic–plastic growth phase, a plastic softening phase, and a residual friction phase. As an example, Figure 4 presents the τ−δ curve at a shear rate of 6° per minute.

Figure 6 illustrates the τ−δ relationship for the interface between sand and a smooth steel surface. This curve can be characterized by three distinct phases: the elastic–plastic growth stage (OP), the plastic softening stage (PM), and the residual friction stage (MN).

During the elastic–plastic growth stage (OP), the interfacial shear stress increases with the relative displacement but at a progressively decreasing rate. This behavior is indicative of pronounced elastic–plastic deformation. In this phase, the primary contribution to the interfacial shear stress comes from the friction between the sand particles and the smooth steel surface, with an additional contribution from the adhesion force between the two materials. As the relative displacement continues to increase, the sand particles experience both translation and rolling, which enhances the friction between the sand and the smooth steel structure. However, despite this increased friction, the rate of increase in shear stress gradually slows over time.

In the plastic softening stage (PM), the interfacial shear stress begins to decrease as the relative displacement increases. Initially, the rate of decrease accelerates, but it eventually slows down as the process progresses. Beyond point P, as the sand particles continue to slide and roll, the friction between the sand and the steel surface reaches a maximum value and then starts to diminish. This occurs as the original interfacial structure, formed by the interaction between the sand and the steel, begins to degrade. As a result, the frictional resistance decreases, and the interfacial shear stress declines, marking the onset of stress softening.

In the residual friction stage (MN), the interfacial shear stress gradually stabilizes and approaches a horizontal line as the relative displacement increases. At this stage, the original structure of the sand and the integrity of the steel surface have been entirely compromised. Consequently, the interfacial shear stress is primarily governed by the sliding friction between the sand and the smooth steel surface. Additionally, as the sand particles progressively align along the shear direction, the interfacial shear stress continues to decrease.

3.2. Effect of Vertical Confining Pressure

The relationship between shear stress and relative displacement at the smooth sand–steel interface under varying vertical confining pressures (0 kPa, 25 kPa, 50 kPa, 75 kPa, and 100 kPa) was investigated through torsional shear tests using a modified indoor triaxial shear apparatus. The results are shown in Figure 7. As evident from the figure, the vertical confining pressure significantly affects the shear stress–relative displacement relationship at the smooth sand–steel interface. As the vertical pressure increases, both the peak shear stress and the residual shear stress increase significantly, with the peak shear stress reaching 3.213 kPa, 11.37 kPa, 18.758 kPa, 24.361 kPa, and 32.906 kPa, respectively. The relative displacement corresponding to the peak shear stress also increases with rising pressure. Notably, when the vertical pressure is 0 kPa, the initial shear stress is not zero, indicating the presence of not only sliding friction but also noticeable cohesion, or adhesive force, between the sand and steel structure. As the vertical pressure increases, the initial shear stress during the relative displacement at the smooth sand–steel interface also increases. This phenomenon follows a similar principle: in order to generate sliding friction, the maximum static friction force must first be overcome, and as the normal stress on the interface increases, the maximum static friction force increases correspondingly.

For the interface shear strength, there are two parameters: peak shear stress and residual shear stress. The peak shear stress is taken as the smooth sand–steel interface shear strength, while the variation in the residual shear stress is analyzed. In addition, the shear stress values obtained in this experiment must be corrected by the instrument to subtract the shear stress generated by the hollow rotation of the vertical rod in the sand sample, i.e., the shear stress solely generated by the round steel rod. The peak shear stress and residual shear stress of the smooth sand–steel interface under different vertical pressures are shown in

Table 7. Percentage change in the peak shear stress and residual shear stress at different vertical pressures.

Experiment Scheme		Vertical Pressures (kPa)
Type	Stress (kPa)	25 to 0	25 to 50	75 to 50	100 to 75
Instrument display	peak shear stress	73%	41%	23%	25%
	residual shear stress	72%	31%	23%	41%
Vertical rod idling	peak shear stress	39%	28%	11%	18%
	residual shear stress	37%	25%	41%	48%
Round steel rod	peak shear stress	78%	42%	24%	26%
	residual shear stress	76%	32%	20%	39%

.

The vertical pressure applied in this experiment acts in the vertical direction of the sand sample rather than directly on the normal direction of the sand–steel interface. Therefore, the pressure actually acting on the sand–steel interface is equivalent to the at−rest earth pressure. The formula for calculating the at−rest earth pressure is:

$${\sigma }_0=K_0 {\sigma }_z$$

(8)

where $K_0$ is the coefficient of earth pressure at rest and ${\sigma }_z$ is the vertical confining pressure, in kPa.

The static earth pressure coefficient is calculated using Jaky’s empirical formula.

$$K_0=1-\sin \phi$$

(9)

where φ represents the effective internal friction angle of standard sand, in °.

The peak shear stress and residual shear stress at the sand–steel interface are clearly shown in Table 7 and Figure 8 to increase with the normal stress. Both exhibit an approximately linear relationship with normal stress. Therefore, a representation similar to the Mohr–Coulomb criterion can be used.

$$\tau =c+\sigma tan\phi$$

(10)

where $c$ is the cohesion force, in kPa, and $\sigma$ is the interface normal stress, in kPa.

As shown in Figure 8a, the curve fitting of peak shear stress has an R² value of 0.99, indicating a good linear fit and adherence to the Mohr–Coulomb failure criterion. From the fitted curve equation in Figure 8a, it can be observed that the apparent cohesion and internal friction angle at the interface between standard sand and round steel bars are c = 2.638 kPa and φ = 28.8°, respectively. The internal friction angle of the standard sand used in this study is 30°. Therefore, for the sand–steel interface, the interface internal friction angle is smaller than the internal friction angle of the sand itself, implying that shear failure occurs at the sand–steel interface. As shown in Figure 8b, the curve fitting of the residual shear stress has an R² value of 0.9601, also indicating a good linear fit and adherence to the Mohr–Coulomb failure criterion. From the fitted curve equation in Figure 8b, the apparent cohesion and internal friction angle at the sand–steel interface are c = 0.98 kPa and φ = 12.6°, respectively. The interface internal friction angle is much smaller than the internal friction angle of the sand. When a significant relative deformation of the interface occurs (greater than 60 mm), the residual shear stress should be selected as the shear failure strength of the interface.

3.3. Shearing Strength of the Interface

For the interface shear failure strength, both the peak shear stress and the residual shear stress were considered. The peak shear stress was defined as the interface failure strength, while the evolution of the residual shear stress was also examined. Furthermore, all shear stress measurements obtained from this experiment were corrected by subtracting the shear stress value generated by the vertical rod in the sand sample during the idling phase. This value represents the contribution solely from the round steel rod. The corrected peak and residual shear stress values at the smooth sand–steel interface, measured under different shear rates, are presented in

Table 8. Percentage change in the peak shear stress and residual shear stress at different shear rates.

Experiment Scheme		Shear Rate (°/min)
Type	Stress (kPa)	2 to 4	4 to 6
Instrument display	peak shear stress	11%	13%
	residual shear stress	9%	8%
Vertical rod idling	peak shear stress	4%	8%
	residual shear stress	11%	−2%
Round steel rod	peak shear stress	12%	14%
	residual shear stress	9%	9%

.

The peak shear stress and residual shear stress both exhibit a significant decreasing trend as the shear rate increases. Figure 9 illustrates the relationship between these two shear stresses and the shear rate for the round steel rod. Through regression analysis, the correlation between both shear stresses and the shear rate is well expressed, with an R² value of 1 for the peak shear stress and 0.99 for the residual shear stress, indicating a high degree of correlation and a strong goodness of fit for the regression models. This relationship suggests that at lower shear rates, stronger bonding and internal friction between the material contact surfaces result in higher peak and residual shear stresses. In contrast, at higher shear rates, the reduced slip time between materials weakens the viscous effects, leading to a marked reduction in the shear stress. In this study, the fragmentation of sand particles during shear was observed, a phenomenon previously reported in earlier studies [56,57]. Parathiras [58] conducted residual strength tests on various soil types at different shear rates using the ring shear apparatus developed by Imperial College London. The results indicated that, for non−cohesive soils, the strength decreased as the shear rate increased. Particle breakage may cause changes in particle shape and alter the contact patterns between particles, which in turn affect the macroscopic mechanical behavior of the soil. As the shear rate increases, the degree of particle breakage becomes more pronounced, leading to structural changes in the soil that subsequently influence the peak shear stress behavior. Specifically, the breakage of sand particles alters the surface friction and adhesion between particles, resulting in a decrease in shear stress at higher shear rates. Furthermore, the material type plays a crucial role in determining the interfacial shear strength. For example, differences in frictional properties and surface roughness between smooth sand steel and round steel bars result in notable variations in peak and residual shear stresses. Overall, the observed negative correlation and linear trend between the interfacial shear strength and shear rate highlight the combined effects of friction, adhesion properties, and the material’s kinetic response.

3.4. Shear Dilation Analysis

In the interfacial shear test, the contact area between the standard sand specimen and the smooth steel structure remained constant, allowing the vertical displacement of the sand specimen to serve as an indicator of changes in its volume. Based on the test data, a curve illustrating the relationship between vertical displacement and relative displacement at the interface between the standard sand and the smooth steel structure under varying shear rates was plotted, as shown in Figure 10.

For clarity, an increase in the volume of the sand specimen is defined as a negative change, while a decrease in volume is considered a positive change. No shear shrinkage was observed at the interface between the standard sand and the smooth steel structure under the three different shear rates. Instead, all test conditions exhibited shear dilation. This behavior can be attributed to the interlocking between sand particles, as well as the adhesive forces and frictional resistance at the sand–steel interface. During shearing, relative rolling between the sand particles and the particles at the sand–steel interface leads to an increase in volume, which aligns with the typical shear characteristics of dense sand. Across all three shear rates, before the relative displacement reached 30 mm, the highest vertical displacement was observed at a shear rate of 6°/min, followed by 4°/min, and the lowest at 2°/min. The lower shear rate results in stronger friction between the sand particles, which enhances particle rearrangement and the interlocking effect. This may suppress the increase in vertical displacement. In contrast, a higher shear rate likely facilitates faster sliding between particles, leading to a more pronounced shear dilation effect. Consequently, during the shear process, although the vertical displacement initially increases rapidly, this increase may eventually stabilize over time due to particle rearrangement or changes in the frictional forces. This results in a trend where the vertical displacement rises to a certain value before stabilizing. The vertical displacements observed at the three shear rates were minimal, with the maximum value not exceeding 0.02 mm [59]. It is generally recognized that shear dilation at the sand–steel pipe pile interface is approximately 0.02 mm. Furthermore, shear dilation also occurs within the influence zone of the interface. In this test, vertical pressure was applied using a spring, resulting in negligible changes in the volume of the sand specimen during shearing at the interface between the standard sand and the smooth steel structure. Given the compression of the spring, the variation in vertical displacement is extremely small and does not significantly affect the experimental results. Therefore, the sand–steel interface in this test can be treated as a normal stress boundary condition.

Figure 11 illustrates the relationship between the vertical displacement and relative displacement at the interface of standard sand and smooth steel under varying relative densities. The shear contraction and shear expansion behaviors at this interface are influenced by factors such as the arrangement of sand particles, the density state, and the inter−particle interaction mechanisms. In the loose density state (27.45%), the sand particles have fewer contact points with each other, resulting in a more loosely arranged structure. During shear, particle rearrangement occurs to fill the voids, leading to shear contraction. In the medium−density state (50.98%), the sand particles are initially more compactly arranged. At the onset of shearing, particle sliding increases the pore volume, resulting in shear expansion. However, as the particles undergo compression, the behavior transitions from shear expansion to shear contraction, reflecting a shift in the internal structural response. In the dense state (72.2%), the particles are tightly packed, and particle sliding during shearing is more restricted. This restriction causes particle rotation or redistribution, leading to an increase in volume, which manifests as shear expansion. This behavior underscores the dynamic adjustment of the internal particle structure and its response to external stresses in sandy soils under different density conditions. Shear dilation plays a significant role in the long−term performance and stability of offshore pipe piles in sandy soils. On one hand, shear dilation alters the particle arrangement and pore structure of sandy soils, which can reduce shear strength and cause fluctuations in pore water pressure. This weakens the surrounding soil’s restraining capacity on the pile, thereby decreasing its lateral stability. Furthermore, the softening effect of the soil leads to a decrease in skin friction, ultimately reducing the pile’s load−bearing capacity, especially under horizontal loads. On the other hand, the soil deformations caused by shear dilation are transmitted to the pile, inducing local deformations and potential fatigue damage, which further compromises the overall stability of the system. In marine environments, the long−term effects of external forces, such as waves and tides, exacerbate the loss of stability, thereby increasing the risk of local instability.

3.5. Effect of Interface Roughness

Four sets of torsional shear tests were conducted to examine the interface between standard sand and steel structures with varying surface roughness levels. The roughness levels were characterized by the average sand depth values: y = 0 mm, y = 0.0365 mm, y = 0.0730 mm, and y = 0.2918 mm. The tests were carried out under a uniform vertical pressure of 25 kPa, a constant shear rate of 6 degrees per minute, and a relative density of 72.2%. The resulting experimental data were analyzed to investigate the influence of surface roughness on the mechanical properties of the interface between standard sand and steel structures.

In this study, following the methodology proposed by Li [60], a sand−filling technique was used to assess the surface roughness of concrete structures. While this method is effective for evaluating the roughness of concrete surfaces exhibiting regular variations, it has two significant limitations. To address these limitations, two correction coefficients were introduced: one for the depth of influence, denoted as K1, and another for the degree of dispersion, denoted as K2. These correction factors helped refine the assessment of surface roughness, making the evaluation more accurate and reliable.

The impact depth correction factor K₁ is:

$$K_1={\displaystyle \frac{V_{\mathrm{d}}}{V}}$$

(11)

where $V_{\mathrm{d}}$ represents the volume of sand filling within the range of the interface shear depth of influence, which is related to the grain size and type of soil. For sandy soils, the maximum depth of influence is typically estimated to be approximately 7 to 8 times the average grain size. In contrast, for clay soils, the depth of influence is generally considered to be around 10 mm. The term $V$ refers to the volume of standard sand filling. When the depth of sand filling is less than or equal to the maximum depth of influence, the corresponding coefficient is set to 1.

The dispersion correction factor K₂ is:

$$K_2={\displaystyle \frac{S_{\mathrm{c}}}{S}}$$

(12)

where $S_{\mathrm{c}}$ represents the area of the rough portion of the surface of the concrete structure and $S$ denotes the total surface area of the concrete structure.

Ultimately, the equation employed to evaluate the surface roughness of concrete structures is:

$$y={\displaystyle \frac{V_d}{V}}{\displaystyle \frac{S_c}{S}}{\displaystyle \frac{V}{S}}$$

(13)

The average depth of sand filling, denoted as $y$, is utilized as a measure to represent the extent of surface roughness.

In this study, a regular distribution of surface grooves was observed on the steel structure, with the width, depth, and spacing of the grooves showing consistent changes. However, the centerline average roughness was found to be inadequate in effectively distinguishing the surface roughness between type 3 and type 4 round steel bars. As a result, the sand replacement method was employed to correct Formula 13 for a more accurate assessment the surface roughness of the round steel bars. The average particle size of the standard sand used in this study was set at 0.375 mm, and the depth of the interfacial shear influence was considered to be seven times the average particle size. Based on this correction, modified Formula 13 was derived using the sand replacement method. The parameters related to the surface roughness of Nos. 1–4 round steel bars are presented in

Table 9. Surface roughness of round steel bars.

Round Steel Bar Type	Square Groove (Length × Width × Depth)	Distance Along the Circumference	Roughness Y (mm)	Standard Deviation
1	0 mm × 0 mm × 0 mm	0	0	0.015
2	40 mm × 1 mm × 1 mm	30°	0.0356	0.023
3	40 mm × 2 mm × 2 mm	60°	0.0730	0.011
4	40 mm × 2 mm × 2 mm	30°	0.2918	0.018

.

To further investigate the mechanical behavior at the sand–steel interface, a modified indoor cross−plate shear apparatus was used to conduct torsional shear tests. The shear stress–relative displacement relationship curves at the sand–steel interface were obtained under varying surface roughness conditions. The results, shown in Figure 12, indicate that surface roughness significantly influences the shear stress–relative displacement relationship. As the surface roughness of the steel structure increases, both the peak shear stress and the residual shear stress at the sand–steel interface increase substantially, while the corresponding relative displacement gradually decreases. For the interface between standard sand and steel structures, the shear stress–relative displacement curves under different roughness conditions exhibit clear peak characteristics, typical of peak−type responses.

These shear stress–relative displacement curves can generally be categorized into three distinct stages: the growth stage, the softening stage, and the residual friction stage. With an increase in the surface roughness of the steel structure, notable changes occur in the characteristics of these stages. In the growth stage, the behavior transitions from elastic–plastic to purely elastic. The softening stage evolves from plastic softening to brittle softening, and in the residual friction stage, the curve no longer stabilizes as a horizontal line but fluctuates around a certain residual value. As the surface roughness increases, the magnitude of these fluctuations becomes more pronounced. This suggests that greater surface roughness induces more significant perturbations to the surrounding sand particles, creating a shear zone at the interface. This shear zone not only causes slippage but also results in rolling and vertical movements of the sand particles, leading to dynamic fluctuations in the interfacial shear stress. Additionally, the residual shear stresses at the interface between standard sand and steel structures with surface roughness levels corresponding to type 2 and type 3 round steel rods were found to be nearly identical. This similarity can be attributed to the minimal difference in surface roughness between the two. However, the residual shear stress at the interface of the type 2 round steel rod showed a slight tendency to decrease further, which highlights the complex and nuanced mechanisms through which surface roughness influences interfacial shear behavior.

4. Discussion

The peak shear stress is defined as the interfacial shear failure strength between the sand and steel. Based on the data obtained from interfacial shear tests conducted on standard sand and round steel bars with varying surface roughness, a relationship curve between the peak shear stress and surface roughness was constructed, as shown in Figure 13. To further analyze the data, an exponential fitting curve was superimposed on the same figure, and the corresponding functional relationship of the exponential curve is provided. This analysis highlights the correlation between surface roughness and the interfacial shear failure strength, offering valuable insights into the behavior of the sand–steel interface under different surface conditions.

$$\tau =64.011-53.748\times (1-e^{-{\displaystyle \frac{\delta }{0.037}}})$$

(14)

where τ is the peak shear stress at the interface of sand and steel with different roughness and δ is the surface roughness of the round steel bar. The curve fitting similarity R² = 0.988 for this exponential fitting curve is good.

As shown in Figure 13, the two curves exhibit a high degree of similarity and a strong correlation, indicating that the exponential model is well−suited to describe the relationship between peak shear stress and surface roughness at the sand–steel interface. As the surface roughness increases, the peak shear stress at the interface also rises, although the rate of increase gradually diminishes. This suggests that the effect of surface roughness on peak shear stress becomes less pronounced as roughness continues to increase. Once the roughness reaches a certain threshold, the interlocking effect of the surface texture on the sand particles becomes nearly saturated, and further increases in the roughness result in only marginal improvements in shear stress. This observation supports the existence of a critical roughness value, referred to as the critical roughness threshold. Specifically, when the roughness value lies between 0 and the critical roughness threshold, variations in the surface roughness of the steel structure significantly affect the peak shear stress at the sand–steel interface, with shear failure occurring at the interface. In contrast, when the roughness value falls between the critical roughness threshold and the maximum roughness, changes in the surface roughness have negligible effects on the peak shear stress, and shear failure occurs within the soil itself, reflecting the intrinsic shear characteristics of the sand. Beyond a certain threshold, the effect of the surface roughness on the shear resistance plateaus, indicating a saturation point. This finding provides valuable guidance for the rational selection of material surface treatments in engineering applications. Additionally, this study highlights the importance of regular inspections and targeted reinforcement measures, particularly in marine environments, to ensure the long−term stability of structures. These results can also be used to optimize pile foundation installation techniques, reducing unnecessary material usage while ensuring structural stability and improving construction efficiency.

Figure 14 illustrates the motion characteristics of standard sand under various surface roughness conditions. At lower surface roughness levels (e.g., y = 0 mm and y = 0.0356 mm), the vertical displacement of the standard sand exhibits notable fluctuations. Specifically, at y = 0 mm, the vertical displacement decreases sharply at the initial stage and then stabilizes, indicating deeper embedding of the standard sand and more pronounced free sliding at lower roughness interfaces. At y = 0.0356 mm, the vertical displacement shows less fluctuation but follows a similar trend. Under higher surface roughness conditions (e.g., y = 0.073 mm and y = 0.2918 mm), the vertical displacement remains almost constant in the early stage and rapidly approaches saturation. This behavior can be attributed to the increased number of anchoring points provided by the rough interface, which generates greater vertical embedding forces, thereby restricting vertical movement. In contrast, the horizontal displacement (relative displacement) of the standard sand continues to increase. As shown in the figure, the growth trend of horizontal displacement remains relatively consistent across different roughness conditions, suggesting that surface roughness has a weaker impact on horizontal displacement.

Figure 15 illustrates that within the range of surface roughness less than 0.036 mm, the absolute value of the vertical displacement increases sharply. This phenomenon reflects the rapid embedding and adjustment process of particles occurring at the interface. At this stage, due to the relatively low surface roughness, the contact between the sand particles and the steel surface becomes smoother. As a result, the frictional resistance and embedding forces between the particles are reduced, allowing the particles to slide and rearrange more freely, leading to a denser contact configuration. As the surface roughness increases to the range between 0.036 mm and 0.15 mm, the sand particles gradually embed into the concave–convex structures of the steel surface. This results in an increased number of contact points, enhanced interfacial friction and embedding forces, and a reduction in the rate of change in vertical displacement. This phase reflects the progressive stabilization of the contact network among the particles. When the surface roughness exceeds approximately 0.036 mm, the variation in the vertical displacement begins to stabilize, indicating that the particle embedding depth has reached a limiting value. Beyond this point, further increases in the surface roughness have a diminished effect on particle movement and interfacial embedding forces, with the overall mechanical behavior being predominantly controlled by the intrinsic properties of the soil matrix.

This study primarily investigates the shear behavior at the interface between standard sand and steel. However, various marine sediments, including coral sand and silty sand, are commonly found in ocean environments. These sediments exhibit significant variations in particle morphology, gradation, and mineral composition, which may result in different shear responses at the interface. Additionally, the experiments in this study were conducted under a single−factor loading condition, whereas real marine environments are subject to the combined effects of wave loading, cyclic loading, and complex stress paths, which have not been fully addressed. Consequently, future research should explore a broader range of marine sediments and incorporate more realistic oceanic conditions, such as wave–current interactions, cyclic loading, and varying burial depths, in order to systematically elucidate the fundamental mechanisms underlying the shear behavior at the sand–steel interface.

5. Conclusions

This study systematically analyzes the effects of shear rate, relative density, stress state, and surface roughness on the shear behavior of large−diameter sand–steel interfaces using an improved indoor cross−plate shear apparatus. The main conclusions are as follows:

At different shear rates, the shear stress–relative displacement curves at the smooth sand–steel interface exhibit distinct peak characteristics. As the shear rate increases, both the peak shear strength and residual shear strength show a decreasing trend. The regression analysis results reveal a strong linear correlation between the shear stress and shear rate, with R² values of 1 and 0.99, respectively, indicating a high degree of linear relationship. As vertical pressure increases, both the peak shear stress and residual shear stress significantly increase, with the peak shear stress reaching 3.213 kPa, 11.37 kPa, 18.758 kPa, 24.361 kPa, and 32.906 kPa, respectively. This demonstrates the crucial role of confining pressure in the interface shear behavior and highlights the key influence of the shear rate and confining pressure on the mechanical properties of the sand–steel interface.

At varying relative densities, the shear deformation behavior of the sand–steel interface exhibits distinct differences. In the loose state (relative density of 27.45%), the sample shows shear contraction; under medium−density conditions (relative density of 50.98%), the sample initially exhibits shear dilation, followed by shear contraction; and in the dense state (relative density of 72.2%), the sample exhibits shear dilation. As density increases, the shear deformation behavior of the sand shifts from contraction to dilation, consistent with the classical theory of the densification–dilatancy relationship in soil mechanics.

An increase in surface roughness leads to a rise in peak strength and a decrease in relative displacement. The impact of roughness can be divided into three stages: the growth stage, the softening stage, and the residual friction stage. The experimental results indicate the presence of a critical roughness value in the relationship between peak strength and surface roughness. When the roughness is below 0.036 mm, changes in roughness significantly affect the peak strength, with shear failure primarily occurring at the sand–steel interface. However, when the roughness exceeds the critical value, its effect on peak strength diminishes, and shear failure shifts to the soil, reflecting the inherent shear behavior of the sand. This finding is significant for understanding the shear failure mechanisms at the sand–steel interface, particularly in the context of improving the shear strength of structural materials in civil engineering design.