Study on Sand-Steel Interface Shear Test Method and Mechanism

Abstract

Soil-structure interface properties play an essential role in geotechnical engineering. The interface shear test is widely used to measure the interface properties. However, in the traditional interface shear test (TIST), distribution of shear stresses along contact surface is not uniform due to boundary effects. Thus, average mechanical response of the whole interface measured by TIST cannot be used to evaluate interface friction properties. In this paper, a novel interface shear apparatus (MIDST) is presented to investigate the shear behaviours of the soil-structure interface. A series of shear tests were conducted on Fujian standard sand-steel interface. Two shear force sensors simultaneously monitor the shear force along the interface: a pre-embedded sensor inside the interface/steel plate (responding to MIDST), while the other outside the interface (responding to TIST). Laboratory test results show that the pre-embedded internal sensor successfully detects the weakening characteristics of the interface, while the external sensor monitors the hardening law. The interface shear strength measured by internal sensor is significantly higher than that monitored by external sensor. A commercial DEM software Version 5.0, Particle Flow Code in Two Dimensions (PFC^2D), is employed to study the soil-structure interaction mechanism, and numerical test results show that the main reasons for the internal and external differences are the uneven shear stress distribution at the soil-structure interface and the boundary effect. In addition, numerical test results agree with the laboratory test results, indicating that the shear behaviours monitored by MIDST are relatively accurate and can provide a reference for engineering design.

1. Introduction

With the significant development of marine resources, various offshore engineering structures (e.g., offshore platforms, large-span bridges, manufactured islands, etc.) widely employ pile and mooring foundations as the underlying supporting systems. They also transfer their self-weight and external loads to deeper, stiff soils or bedrock to guarantee stability. Under self-weight and external loads, the foundations are prone to bearing capacity failures, usually caused by relative sliding between the soil and foundation structure due to interface strength reduction [1]. Thus, the soil-structure interface characteristics are vital for the safety and stability of offshore engineering structures. However, the present parametric investigations are mainly based on the laboratory test considering the interface stress conditions, foundation materials, interface roughness, etc. [2,3,4,5,6,7,8]. However, the test results do not accurately reflect the natural characteristics of the soil-structure interaction [9,10,11]. Therefore, it is necessary to investigate the actual behaviours of the soil-structure interface to better understand the associated interface strength variation law and illustrate the foundation failure mechanism.

Many scholars conduct variation test methods to investigate the soil-structure interface behaviour and use it to assess the stability of the engineering structure. Two main test methods have been widely adopted to survey the interaction between the soil and solid materials: full-scale in situ and laboratory tests. The full-scale in situ tests can reflect the realistic natural state of the soil, whereas the scaling effect or the sample disturbance of the laboratory tests can affect the investigation results [12,13,14,15,16,17]. However, the field trial cost of crews and devices is quite expensive, and the test period is quite long. For this reason, many scholars consider the effectiveness and convenience of the laboratory tests, as various test conditions can be employed to explore the soil-structure interaction mechanism [18,19,20,21,22,23,24,25,26,27]. With the development of science and technology, understanding the soil-structure interface behaviours is becoming more comprehensive and reasonable regarding their stress state and geotechnical factors. However, the realistic characteristics of the soil-structure interface and the gap between the laboratory and full-scale situ tests have not been thoroughly understood [23,24,25]. Most researchers develop and improve the unit interface test methods based on the direct shear apparatus, simple shear apparatus, torsion shear apparatus, cylinder shear apparatus, and triaxial apparatus. The interface direct shear test has been widely adopted in the research of the soil-structure interface [19,28,29,30,31]. Despite most researchers conducting a necessary correction on the contact surface and the system friction (actually, it was not completely revised), they usually assume that the boundary effect of the sample box has no influence on the test results, and that the stress and strain are uniform on the interface. Hu and Pu [9] conducted a series of numerical tests on the soil-concrete interface and found that the distribution of shear stress and normal stress on the contact surface was uneven. Yin et al. [10] performed microscopic deformation laboratory tests and obtained the uneven distribution phenomenon of the shear stress and strain by observing the local soil deformation. The interface simple shear test can weaken the boundary effect on the soil-structure interface test and distinguish the relative shear displacement of the soil-structure and the shear deformation of the soil [32,33,34,35,36]. However, the overlaying ring near the soil-structure contact surface will constrain the soil particles during the shearing, for it has a certain thickness. So, there may still be a certain stress concentration phenomenon with non-homogeneous strain, and there is also system friction due to its complex overlaying-ring structure. The torsion shear test can carry out the continuous shear without the boundary effect [37,38,39,40,41]. However, sandy soil is not suitable for preparing a thin-walled cylindrical sample. The principle of the cylinder shear test [42,43,44] and the triaxial interface shear test [5,7,8] is consistent with that of the direct shear test. In addition to the uneven shear stress distribution phenomenon, it is worth noting that there must be strict control on the ratio of pile diameter to soil sample diameter, which aims to reduce the influence of the boundary effect. Even though the large-scale laboratory test can decrease the impact of the boundary effect on the interface behaviours, the sample heterogeneity will also cause inaccuracy in the test results.

To sum up, the current research methods of the soil-structure interface shear test are mainly limited in three aspects. First, the shear stress is unevenly distributed along the shear plane; second, how can the resistance entirely be eliminated in the complex test system to achieve accurate measurement; and third, the homogeneity of the sample. For the third factor, the unit laboratory interface test can ensure sample homogeneity and is convenient for sampling. However, the soil-structure interface shear test method depends on the soil properties. For the kit with unique shape sample requirements, the preparation and loading process of the sandy soil sample is more complex than that of the clayey soil sample. For the first and second factors, a new set of shear apparatuses should be developed and evaluated to assess the soil-structure interface’s actual shear behaviours. Thus, to eliminate the boundary effect and the complex system friction, this paper proposes a modified interface test method. The new soil-structure interface test method sets one tangential force sensor inside the soil-structure interface, as shown in Figure 1. This method was validated and supported by Liu et al. [45].

For this purpose, this paper aims to reduce the influence of the boundary effects and the system friction, designing and producing an interface shear device called the “Modified Interface Direct Shear Test” (MIDST), with several verification tests conducted to investigate the soil-structure interface behaviour. Furthermore, the discrete element method (DEM) is combined to explain the softening mechanism of the soil-structure interface. The laboratory test and numerical test results show that the experimental method and the apparatus can effectively obtain the actual characteristics of the soil-structure interface.

2. MIDST Apparatus

The direct shear instrument with a simple internal structure, a significant transformation space, and the manufacturing, saturation, and installation process of clayey and sandy soil samples are relatively convenient and easy. And many interface shear test studies are based on displacement-controlled direct shear tests to evaluate the interfacial strength (including peak and residual strength). The test results are very representative and can be compared with the test results of this study, which will be presented in the following. Thus, this paper upgraded and modified the traditional direct shear instrument ZFY-1 (Figure 2) using a steel block with an internal sensor to replace the lower shear box to accurately obtain the soil-structure interface strength (see Section 3 for details). We refer to the improved testing system as the MIDST apparatus, as shown in Figure 2. It consists of four primary systems: normal loading system, shear loading system, interface-measurement block, and data acquisition and control system. The normal loading system mainly comprises a normal load actuator to apply the normal stress on the top of the sample. A normal displacement sensor monitors the normal shear deformation during shearing. The sample diameter is 61.8mm, and the height is 20 mm. The upper surface of the friction block is 20 mm by 20 mm. The shear loading system permits shear rates ranging from 0.002 mm/min to 4.5 mm/min.

The interface-measurement block consists of a friction block with an internal shear load sensor, an internal slide rail, and a frame (Figure 3). The whole structure consists of the top plate, middle frame, and bottom plate. The interface-measurement block is 130 mm in length by 130 mm in width by 25 mm in height, which is placed inside and fits precisely with the lower shear box. The upper surface roughness of the friction block and the top plate is identical. The tension-compression sensor connects the friction block to the sidewall of the middle frame structure, adopting the methods of loading and unloading in steps to calibrate the sensor’s sensitivity assembled in the interface-measurement block. The weights are loaded in 3 steps and unloaded in 3 steps. The sensitivity assigned to the whole interface-measurement block is 65.27 N/V.

An MR micro linear slide rail, CPC MRU3MN SU from Taiwan, is arranged between the friction block and the bottom plate. According to the manufacturer’s laboratory test results, for a single slide rail without any structural design, the friction between the seal plate at the bottom of the slider and the track is 0.1 N, and the sliding friction coefficient between the slider ball and the trail is 0.002~0.003. The friction of the CPC linear slide rail is minimal and negligible, as demonstrated by a simple test. We use a handheld dynamometer to push the side centre of the friction block through a reserved hole and measure the linear slide rail’s resistance after assembly (Figure 4). The dynamometer reading f₁ and the inner tension-pressure sensor reading f₂ are recorded under normal forces of 0, 25, and 50 N, respectively. Table 1 shows the test results, the interface-measurement block friction is the linear slide rail friction, responding to the difference f₁–f₂ between the two sensors. From the experimental results, the friction of the slide rail after assembly is larger than the calibration results of the manufacturer’s laboratory. The friction of a single-section seal after assembly is 0.13 N, and the friction coefficient is about 0.0026. And this minor friction increment can be ignored relative to the shear force on the soil-structure interface.

3. Experimental Procedure

3.1. Test Materials

In this paper, the sand used in the interface shear tests is Fujian Standard sand, which has highly rounded grains and relatively uniform grain size distribution. The soil passes through the 2 mm sieve. The maximum dry density is obtained according to the American Society for Testing and Materials (ASTM) guidance D4253-16e1 [46], and three groups of tests get ρ_dmax = 1.928 g/cm³.

The S316 stainless steel casts the interface-measurement block, the original roughness of its top surface R_a is 1.6, and the wire-cutting process treats the top surface after assembly. The purpose is to make the frame’s top plate and the friction block’s top surface in the same plane and eliminate the influence of structural errors on the shear test results.

3.2. Specimen Preparation

The samples use the dry sample preparation method to achieve the target sample density. A given mass of dry sand is filled into three layers in the cutting ring and compacted by the light compaction method to achieve a uniform density in the model. The height of the specimen is 20 mm, with a density of 1.83 g/cm³. The samples are finally placed in the vacuum saturator for saturation, as shown in Figure 5.

3.3. Testing Procedure

The specimen is inside the upper shear box and laid upon the interface-measurement block. The sample is immersed in the water during shearing to keep saturation. And before shearing, the specimen needs preloading. Constant normal loading (CNL) test condition is employed in all tests, and the normal stress has three levels, including 50, 100, and 150 kPa, commonly found in geotechnical engineering applications, which are undertaken to study the normal stress influence on the interface shear behaviours. The shear load followed the strain-controlled mode at a shearing rate of 0.8 mm/min until the horizontal displacement reached 13mm, corresponding to 20% strain of the samples. According to ASTM D3080/D3080M-11 “Standard Test Method for Direct Shear Test of Soils Under Consolidated Drained Conditions” [47] and the research on the scale effect on direct shear tests [48,49,50], to avoid the influence of the scale effect, the specimen size should be not less than 10 times the maximum particle diameter d_max or 50 times the median particle diameter d₅₀ in length, and not less than 6 d_max in height. The maximum particle diameter, d_max, of the sand employed in this study is 2 mm, the median particle diameter, d₅₀, is 0.24 mm, the sample length is about 20 d_max and 257 d₅₀, and the height is 10 d_max. Therefore, it can be assumed that the scale effect does not affect the test results.

4. Experimental Results and Discussion

4.1. Verification Test

Figure 6 exhibits the shear behaviour of the soil-structure interface shear tests under different levels of normal stress. The typical shear load–time curves are shown in Figure 6a. It can be observed that the soil-structure interface shear load increases with the normal stress increase. The main reason is that the normal stress applied to the interface increases the friction resistance between the soil particles and the steel or the soil particles themselves, increasing the shear zone’s peak strength. It is worth noting that when analyzing the interface shear strength, we found that the interface strength measured by the internal sensor in the interface-measurement block is more significant than that measured by the external sensor. The results measured by the external sensor are hardening curves, while those measured by the internal sensor are softening curves, showing apparent inconsistency. The following will analyze and discuss the above conjecture and test results in combination with the characteristics of the direct shear test and relevant references to demonstrate the reliability of the test method proposed in this paper.

First, we conducted a stress analysis for two test methods using the internal and external sensors on the soil-structure interface shear test, as shown in Figure 7. By establishing the force balance, we can get:

(1)$f_{\mathrm{out}}=f_{\mathrm{r}}+T_{\mathrm{out}}-f_{\mathrm{b}}$

(2)$f_{\mathrm{in}}=f_{\mathrm{CPC}}+T_{\mathrm{in}}$

where f_out represents the shearing load of the interface measured by the external sensor; f_in represents the shearing load of the interface measured by the internal sensor; f_r is the friction between the horizontal shear rod and the sidewall of the water bath; f_b represents the friction force between the sample box and the underlying structure; T_out is the counting of the external sensor; T_in is the counting of the internal sensor; f_CPC is the friction of the linear slide rail, the previous test results show that the friction of the linear slide rail is too small and negligible. It can be considered that the measurement results of the internal sensor are equal to the shearing force of the interface, f_in = T_in.

The friction between the horizontal shear rod and the sidewall of the water bath f_r mainly affects the accuracy of the traditional test method with an external sensor. In other words, if we use the counting of the external sensor as the shearing force of the interface, we will underestimate the shear strength of the soil-structure interface (f_out > T_out). Therefore, it is necessary to calibrate the direct shear test system friction and correct the measurement results of the traditional soil-structure interface shear test. We measure the friction of the direct shear test system in two ways: (i) pushing with a handheld dynamometer or (ii) loading with a pulley. An improved horizontal shear rod is employed in this test, as shown in Figure 8. The friction of the pulley, f_pulley, is measured through the three groups of tests, and the average of test results is 9.72 N as the f_pulley.

The results of the two calibration tests are shown in Table 2. The test results show that the f_r of approach 1 is 22.6 N, and the f_r of approach 2 is 21.59 N. The difference between them is approximately 1.01 N. Because the sensitivity of the external sensor is 1.02 N/mv, the error of the two test methods is within a sensitivity range. We can consider that the test results are relatively close, and the average value of the two is 22 N as the f_r. The f_r is about 25~50% of the measurement results of the external sensor in the previous interface shear test. It has a significant impact on the measurement results and should be corrected, and the correction results are shown in Figure 6.

From Figure 6b, we can find that although the test results of the external sensor have been corrected with the interface shear system friction, the peak shear strength of the tests of the internal sensor is still greater than those of the external sensor. And the residual strength measured by the internal sensor is less than the test results of the external sensor. There is only one reason to explain this problem: the shear stress distribution in the shear zone is uneven. The shear stress at the centre of the contact surface is greater than the average shear stress of the whole contact surface.

In summary, by combining the literature and the interface direct shear test characteristics, we discuss the effects of the uneven distribution of the shear stress on the test results during the shearing process.

4.2. Analysis of the Shear Stress Distribution

Hu and Pu [9] and Yin et al. [10] investigate the soil-concrete interface behaviours by the finite element method (FEM) and laboratory test, respectively. Both studies [9,10] have found that the distribution of the stress and strain on the interface was uneven. Hu and Pu [9] found the right contact units are tensioned horizontally at the beginning of shearing and are the first to fail. With the increase of shear displacement, the failure gradually develops to the interior of the interface. And when the shear displacement reaches 10 mm, the development law of shear stress remains unchanged with the shear displacement increasing. It presents that all contact units are destroyed, and the last failed contact unit is on the right of the interface centre. However, the shear stress of the left contact units grows relatively large in the early shearing stage. This is mainly because the normal stress is higher than the average due to the dilation of the soil sample under the CNS test conditions. At this moment, the corresponding stress ratio (τ/σ_n) of the left contact units is the highest in all contact units, so the failure also expands into the interface centre from the left edge. Yin et al. [10] observed the soil deformation near the contact surface by the interface direct shear tests with the “micro-periscopes”. They found that the left soil slides first, and the soil does not slide under minor shear stress in other positions. And with the increase of the shear stress, the centre soil gradually starts to slide. They considered that the failure of the contact surface is a gradual development process from the left edge to the inside, and the shear stress distribution at the centre is the largest. Guo et al. [28] also observed the shear stress and normal stress uneven distribution phenomenon in the soil-polymer interface by the FEM.

Based on the above, we can explain why the interface peak strength measured by the internal sensor is more significant than that measured by the external sensor from the shear stress distribution. It is mainly caused by the uneven distribution of shear stress on the shear plane, which is the difficulty and limitation in the interface shear test. It objectively proves the scientificity and rationality of the test method proposed in this paper.

4.3. Normal Stress Effect on the Interface Shear Strength

As we all know, scholars usually use the Mohr-Coulomb criterion to characterize the mechanical properties of the soils, such as Equation (3). The peak strength and its corresponding normal stress for each set of tests are recorded in Figure 9. And the circle points are the test and analysis result of the external sensor, which corresponds to the test result of the traditional test method; the rectangle points are the test and analysis result of the internal sensor, which corresponds to the test result of the improved test method mentioned in this paper.

(3)${\tau }_f={\sigma }_3\times \tan \delta +c$

where τ_f is the maximum shear stress of the shear stress vs. shear displacement curves; σ₃ is the normal stress; c is the sand and steel contact surface cohesion.

From Figure 9, it can be found that the relationship between the peak shear stress and the normal stress is linear with the Mohr-Coulomb criterion, regardless of whether the measurement method uses the external sensor or our internal sensor. The friction angle obtained by our improved method is more significant than that obtained by the traditional interface test, in which the friction angle obtained by the improved method is δ = 40°, while by the conventional way, the friction angle δ’ = 29°. This is mainly due to the boundary effects on the test results, which result in different positions of the shear planes. If the failure occurs inside the soil, more soil particles participate in the anti-friction motion, leading to greater friction force. And if the failure occurs on the contact surface, the relative movement of soil particles will be less, and the friction force will be less. The test method proposed in this paper belongs to the former one. It is less affected by the boundary effect, failure occurs inside the soil, and the friction angle is more significant than the conventional method.

5. DEM Analysis

For further study of the effectiveness of the interface test method proposed by this paper, this section combines the discrete element method (DEM) to explore the discontinuous deformation process of the soil-structure interaction from the mesoscopic view. We can effectively simulate the interface shear band’s stress and deformation variation law. The last section captured two soil-structure interaction parameters: the friction coefficient of 0.84, which was measured in the centre of the interface, and 0.56, which was measured externally. The maximum stress ratio (μ_max = τ_max/σ_n) corresponds to the shear band’s failure, representing the soil’s failure (soil particles themselves interaction). And the last one is the average maximum stress ratio (μ_aver = τ_aver/σ_n), representing the average resistance coefficient between soil particles in the shear band and the structure. Thus, the parameters will be adopted into DEM analysis to simulate particle-to-particle and structure interactions.

5.1. DEM Test Scheme

The particle flow method is the most widely used DEM numerical analysis method for discontinuous media. Popular DEM numerical analysis software, such as PFC^2D/^3D, usually takes the rigid disk (for 2D) or rigid ball (for 3D) as the essential solution element. Due to its smaller computer memory space and faster calculation speed than the other methods, it is suitable for solving the soil-structure interaction’s discontinuous deformation problem. Therefore, combined with the 2D particle flow numerical analysis software PFC^2D, this paper conducts the numerical direct shear tests on the discontinuous deformation problem of the soil-structure interface. We use the linear contact model to characterize the mechanical contact relationship between the soil particles and structures, as well as the soil particles themselves. The details are as follows:

• The soil particles are assumed to be balls, of which the diameter ranges from 0.15 mm to 1.2 mm, responding to the actual grain gradation of the Fujian standard sand;

• The particle sample is generated by the radius expansion method, and the initial porosity of the sample is 0.12;

• The consolidation stress and normal stress are loaded by continuously adjusting the moving rate of the boundary wall through the private server control function;

• Combining with the mesoscopic parameters obtained from the literature by Huang et al. [51] and the calibration test, the numerical interface direct shear test of the soil-structure interaction by DEM is carried out. The mesoscopic parameters are shown in Table 3.

The DEM simulation adopts three structure models to study the shear stress distribution on the contact surface and compare the method proposed in this paper with the traditional approach. The DEM model is shown in Figure 10, in which the structure type I is employed to investigate shear stress distribution on the contact, dividing the bottom wall (acts as the structure) into seven parts, of which the middle five divisions (W2~W5) correspond to the different points of the contact surface between the soil and structure; the structure type II represents the traditional interface test method, employing a whole wall to simulate the structure; and the structure type III represents the method proposed in this paper, using a central wall and two connected side walls to show the internal test scheme. Due to limited space, this paper merely cites the interface shear test with a normal stress of σ₃ = 100 kPa and the shear rate of v = 1 mm/min as typical to illustrate the shear stress distribution law, as shown in Figure 11.

5.2. DEM Analysis Results

From Figure 11, we can see that the shear stress curves vary across different sections. The shear stress increases gradually from left to right: the shear stress in the left area (W2) is the least and reaches its maximum in the right section (W6), while the inner cells (W3~W5) are similar. A fascinating phenomenon is that the shear band above W3~W5 gradually thickens first and then thins, with the thickest shear band presents above W4. The main reason for this phenomenon is that the boundary effect on the two sides simultaneously affects the soil particles. During the shearing process, the soil particles above the W5, on the shear band, move along the shear direction, forming a dense area. This particle flow law leads more soil particles to engage in anti-sliding, so the shear resistance is much higher than the average level and exhibits a hardening curve. But for the soil particles above the W6, except for the soil particles on the shear band moving along the shear direction, the remaining soil particles tumbled slightly upward along the shear direction due to the constraining of the boundary conditions, resulting in a decrease in the normal stress and exhibiting a softening curve. Meanwhile, the soil particles above the W2 and W3 areas collapse downward during the shearing process, forming a loose scope. With the increase of shear displacement, the number of collapsed particles varies across different sections, and the shear resistance weakening degree varies. Due to the strong boundary effect, the failure of the soil-structure interaction occurs on the contact surface, and the thickness of the shear band is approximately zero (W2). The area near the centre of the contact surface (W3 and W5) is less affected by the boundary effect (loose or squeeze) than the boundary area (W2 and W6), and there are evident disparities. For the W4 part at the centre of the sample, the failure gradually extends into the soil, and the shear band grows thicker. It means that more soil particles are disturbed in this area, and the contact area of the soil particle will increase, resulting in higher shear resistance. And with the shearing displacement increase, the particle movement gradually becomes stable, decreasing the shear stress and forming a softening curve.

The DEM shear stress analysis results of this paper’s proposed method and the traditional test method are shown in Figure 12. The shear stress law indicates that the basic feature of the DEM numerical test results is consistent with that of the MIDST test results. However, two noticeable differences exist between the DEM and the test results. First, for the peak shear stress, the centre block’s shear stress is slightly larger than the average, mainly due to the shear stress’s uneven distribution on the contact surface. In fact, the difference between the two methods obtained by DEM is much smaller than that of the test. This is mainly because the system friction of the whole test system has not been completely eliminated, even though friction between the horizontal shear rod and the container sidewall has been made. Due to the dynamic friction transformation to static friction, the average shear stress will be undervalued before the failure, resulting in the type of the interface average response curves not being similar. Second, for the residual strength of the interface, since the method proposed in this paper is aimed at a certain point on the contact surface to measure the shear stress, the mechanism is similar to the continuous penetration process of the pile. In addition to the strength softening of the interface, the friction fatigue caused by continuous shear will also reduce the residual strength [24,52].

5.3. Mechanism of Different Interface Test Methods

DEM results show that, in addition to the obvious shear band on the contact surface (Figure 13a), two vertical shear bands are near the sample box’s sidewalls (Figure 13b), exhibiting a vertical “soil arched” effect. The boundary constraints mainly cause their formation, and the sample is divided into four parts based on the particle’s movement direction (Figure 13c). For the left part, the soil particles collapse, and soil particles in the right area roll up. The particles in the bottom part move in a shear direction, and the upper particles move in the opposite direction. This phenomenon illustrates that except for the relative displacement of the soil-structure, there is also soil deformation during shearing. It is consistent with the conclusion of Yin et al. [10], which proves the rationality of the DEM verification method in this paper.

Thus, we propose the mechanism of the sand-structure interaction with the test results and the DEM results. During shearing, if we push the lower shear box to the right, the extrusion force of the upper shear box will first load on the right end of the sample. And when the specimen is in equilibrium with the horizontal, the horizontal reaction force can only be provided by the friction of the structure lying in the lower shear box. Because the soil sample is a non-rigid body, the displacement of the soil particles is different during the shearing. The squeezed side soil first has a horizontal displacement and is subjected to the structure friction. With the increase of the relative displacement, the relative dislocation between the soil sample and structure gradually develops from the right to the left. The shear force on the contact surface at the right end first reaches the maximum and fails, which is most affected by the boundary condition of the upper shear box, and the failure occurs on the contact surface. With the shear displacement increasing, the left soil is greatly affected by the particle behaviour compared with the other locations, and the shear stress ratio (τ/σ_n) is under a comparatively high state. Thus, the progressive failure is from left to right and combined with the right boundary effect, and we can determine that the final failure occurs near the centre on the contact surface. Meanwhile, the boundary effect does not affect the central soil on the contact surface because it is far from the sample box edges. It is constrained by the elastic boundary conditions rather than the rigid boundary conditions, so its stress state is closer to reality. And the failure is more likely to occur inside the soil (mainly affected by normal stress). The shear behaviours of the central soil-structure interface are affected by multiple interface interactions, including the rolling interaction between the soil particles and themselves and the biting or sliding interaction between the soil particles and the structure. While the test method proposed in this paper aims at this key problem, modifying the traditional interface test method with the inter-sensor inside the soil-structure interface to realize the accurate measurement of the soil-structure interaction.

From the above, we demonstrate the rationality of the results of the MIDST apparatus with the previous research results and the mechanical analysis. The improved interface shear test system breaks through the difficulties and limitations of the existing research methods on interface shear behaviour. Using the MIDST apparatus, we can more accurately obtain the actual interface characters and eliminate the influence of the uneven shear stress distribution and the system friction on the test results. In addition, MIDST results reveal that the interface shear test is not an element test but a boundary value problem, which leads to some influence on the test results by the specimen (shear box) and soil particle size. This study qualitatively evaluates the limitations of the traditional interface shear test using the MIDST method, assuming that the sample size employed is not affected by the scale effect based on previous research. However, this aspect of the influence of the scale effect on the MIDST method has not been verified. Therefore, our further research plan will consider the scale effect on the MIDST interface shear characteristics and determine the appropriate soil sample size with different particle sizes and gradations. We will also conduct more laboratory tests, consider the evolution of interface parameters of various soils and structural materials under different loading conditions, and combine theoretical calculations with the results of large-scale model pile tests and full-scale installation tests to establish a correlation between laboratory test results and field test results, which can quantitatively give design parameters suitable for guiding engineering.

6. Conclusions

This paper proposes a novel interface direct shear apparatus to investigate the shear behaviour of the soil-structure interface. And to study the difference between the modified interface direct shear test and the traditional method, a series of soil-structure interface tests consisting of the Fujian standard sand and steel were conducted. The main conclusions are summarized as follows:

• A series of shearing tests validated the behaviours of the MIDST apparatus by comparing the results with the traditional direct shear test results. For the Fujian standard sand, the results measured by the internal sensor were the softening curves, while the results measured by the external sensor were the hardening curves. The peak strength measured by the internal sensor was more significant than that measured by the external sensor, while the residual strength measured by the internal sensor was lower than that measured by the external sensor.

• The system friction dramatically influences the results of the traditional interface shear test. The system friction had more than a 25% impact on the conventional interface direct shear test results, which show that the shear strength of the interface would be underestimated by adopting the traditional interface direct shear methods.

• The boundary effect brings out the uneven distribution of the shear stress on the shear plane, which caused the difference between the novel MIDST interface shear test and the traditional interface direct shear test. The results obtained by the conventional interface direct shear tests were the average shear stress of the interface, but the MIDST interface shear tests were the real shear stress of the shear plane (zone).

• Compared with the traditional interface shear test, the internal sensor (MIDST) measurement method can eliminate the influence of the uneven shear stress distribution, the fixed shear plane, the boundary effect, and the system friction on the test results. The relationships between the peak shear stress and the normal stress, tested by novel and the conventional method conform to the Mohr-Coulomb criterion. The friction angle obtained by the novel method is φ = 40°, which is greater than that obtained by the conventional method which is φ′ = 29°. The results measured by the novel test method were closer to the actual shear behaviour of the soil-structure interaction.

• PFC^2D was used to investigate the process of the shear test and describe the related mechanical mechanisms. The numerical simulation results were in good agreement with the test results.

Author Contributions

X.H., Data curation, Formal analysis, Investigation, Methodology, Validation, Writing—original draft. L.Y., Conceptualization, Methodology, Supervision, Visualization, Writing—review & editing. Y.H., Writing—review & editing. Q.Y., Supervision, Funding acquisition, Writing—review & editing, Resources. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data will be made available on request.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. Limitations of traditional interface direct shear test and the modification of the new method.

Figure 2. Schematic of the Modified Interface Direct Shear Test (MIDST) apparatus.

Figure 3. Schematic of the interface-measurement block.

Figure 4. Friction measurement of the linear slide rail.

Figure 5. Specimen preparation.

Figure 6. Shear behaviour of the interface shear verification tests under different normal stress levels: (a) shear force and (b) shear stress curves.

Figure 7. Load analysis for two test methods, both for those using internal and external sensors.

Figure 8. Calibration of fr.

Figure 9. Peak shear stress vs. normal stress.

Figure 10. Numerical modelling of soil-structure interface direct shear test with the DEM method.

Figure 11. Relationship between the shear stress and displacement at different positions on the soil-structure interface: (a) shear stress of different blocks and (b) particle movement mechanism.

Figure 12. A comparison of different test methods between DEM and experimental results: (a) DEM simulation results and (b) laboratory test results.

Figure 13. Displacement nephogram of the soil particle obtained by PFC2D: (a) x-direction; (b) y-direction; (c) particles move direction.

Figure 13. Displacement nephogram of the soil particle obtained by PFC2D: (a) x-direction; (b) y-direction; (c) particles move direction.

Table 1

Resistance of the slide rail after assembly under different normal forces.

Normal Force (N)	f1 (N)	f2 (N)	f1–f2 (N)
0	35.07	34.81	0.26
25	38.44	38.11	0.33
50	41.53	41.14	0.39

Table 2

The results of the two calibration tests (unit: N).

Method 1						Method 2
Weight	Sensor	Friction	f pulley	f r,1	Average1	Dynamometer	Sensor	f r,2	Average2
37.50	4.08	33.42	9.72	23.70	22.60	49.81	29.58	20.23	21.59
62.49	30.60	31.89		22.17		26.84	4.08	22.76
87.48	56.10	31.38		21.66		34.14	13.26	20.88
112.47	79.56	32.91		23.19		46.89	24.48	22.41
137.46	105.06	32.40		22.68		41.28	20.40	20.88
162.45	130.56	31.89		22.17		44.81	22.44	22.37

Table 3

Mesoscopic parameters of the DEM test.

Contact Stiffness		Ball-Ball Friction Coefficient	Ball-Facet Friction Coefficient
kn (N/m)	ks (N/m)	fric b-b	fric b-f
2 × 108	1 × 108	0.84	0.56

References

1. Randolph, M.F. Science and empiricism in pile foundation design. Géotechnique; 2003; 53, pp. 847-875. [DOI: https://dx.doi.org/10.1680/geot.2003.53.10.847]

2. DeJong, J.T.; Westgate, Z.J. Role of Initial State, Material Properties, and Confinement Condition on Local and Global Soil-Structure Interface Behavior. J. Geotech. Geoenviron. Eng.; 2009; 135, pp. 1646-1660. [DOI: https://dx.doi.org/10.1061/(ASCE)1090-0241(2009)135:11(1646)]

3. Desai, C.S.; Drumm, E.C.; Zaman, M.M. Cyclic Testing and Modeling of Interfaces. J. Geotech. Eng.; 1985; 111, pp. 793-815. [DOI: https://dx.doi.org/10.1061/(ASCE)0733-9410(1985)111:6(793)]

4. Lemos, L.J.L.; Vaughan, P.R. Clay-interface shear resistance. Géotechnique; 2000; 50, pp. 55-64. [DOI: https://dx.doi.org/10.1680/geot.2000.50.1.55]

5. Li, C.H.; Kong, G.Q.; Zhang, X.R.; Liu, H.L.; Wang, L.H. Thermomechanical properties of sand-structure interface using temperature-controlled triaxial instrument. Can. Geotech. J.; 2021; 58, pp. 248-260. [DOI: https://dx.doi.org/10.1139/cgj-2020-0026]

6. Martinez, A.; Stutz, H.H. Rate effects on the interface shear behaviour of normally and overconsolidated clay. Géotechnique; 2019; 69, pp. 801-815. [DOI: https://dx.doi.org/10.1680/jgeot.17.P.311]

7. Yang, Q.; Zheng, Z.; El Naggar, M.H.; Kong, G.Q.; Zhang, S.X.; Yang, G.; Ren, Y.B. Experimental study of the interface evolution behavior and softening mechanism of structure-marine soil. Soils Found.; 2023; 63, 101324. [DOI: https://dx.doi.org/10.1016/j.sandf.2023.101324]

8. Zheng, Z.; Naggar, M.H.E.; Wang, J.L.; Yang, Q. Coupling effect of roughness and normal pressure on mechanical characteristics and failure mode of deep-sea sediment interface. Ocean Eng.; 2023; 267, 113255. [DOI: https://dx.doi.org/10.1016/j.oceaneng.2022.113255]

9. Hu, L.M.; Pu, J.L. Testing and modeling of soil-structure interface. J. Geotech. Geoenviron. Eng.; 2004; 130, pp. 851-860. [DOI: https://dx.doi.org/10.1061/(ASCE)1090-0241(2004)130:8(851)]

10. Zong-Ze, Y.; Hong, Z.; Guo-Hua, X. A study of deformation in the interface between soil and concrete. Comput. Geotech.; 1995; 17, pp. 75-92. [DOI: https://dx.doi.org/10.1016/0266-352X(95)91303-L]

11. Gan, F.; Ye, X.; Yin, K.; Li, M.; Yang, J.; Xiao, Y. Experimental study on progressive failure of soil-structure interfaces based on a new measuring method of local stress and displacement. Chin. J. Rock Mech. Eng.; 2018; 37, pp. 734-742.

12. Buckley, R.M.; Jardine, R.J.; Kontoe, S.; Parker, D.; Schroeder, F.C. Ageing and cyclic behaviour of axially loaded piles driven in chalk. Géotechnique; 2018; 68, pp. 146-161. [DOI: https://dx.doi.org/10.1680/jgeot.17.P.012]

13. Chen, R.P.; Peng, C.Y.; Wang, J.F.; Wang, H.L. Field Experiments on Cyclic Behaviors of Axially Loaded Piles Jacked in Soft Clay. J. Geotech. Geoenviron. Eng.; 2021; 147, 04020176. [DOI: https://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0002452]

14. Gong, W.; Li, L.; Zhang, S.; Li, J. Long-term setup of a displacement pile in clay: An analytical framework. Ocean Eng.; 2020; 218, 108143. [DOI: https://dx.doi.org/10.1016/j.oceaneng.2020.108143]

15. Heidari, M.; Hesham El Naggar, M. Design guideline for extended pile shaft incorporating nonlinear pile-soil interaction. Ocean Eng.; 2022; 243, 109371. [DOI: https://dx.doi.org/10.1016/j.oceaneng.2021.109371]

16. Igoe, D.; Gavin, K. Investigation of Cyclic Loading of Aged Piles in Sand. J. Geotech. Geoenviron. Eng.; 2021; 147, 04021011. [DOI: https://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0002451]

17. Rimoy, S.P.; Jardine, R.J.; Standing, J.R. Displacement response to axial cycling of piles driven in sand. Proc. Inst. Civ. Eng.-Geotech. Eng.; 2013; 166, pp. 131-146. [DOI: https://dx.doi.org/10.1680/geng.12.00052]

18. Chao, Z.M.; Fowmes, G. Modified stress and temperature-controlled direct shear apparatus on soil-geosynthetics interfaces. Geotext. Geomembr.; 2021; 49, pp. 825-841. [DOI: https://dx.doi.org/10.1016/j.geotexmem.2020.12.011]

19. Donna, A.D.; Ferrari, A.; Laloui, L. Experimental investigations of the soil–concrete interface: Physical mechanisms, cyclic mobilization, and behaviour at different temperatures. Can. Geotech. J.; 2016; 53, pp. 659-672. [DOI: https://dx.doi.org/10.1139/cgj-2015-0294]

20. Doherty, P.; Gavin, K. Cyclic and Rapid Axial Load Tests on Displacement Piles in Soft Clay. J. Geotech. Geoenviron. Eng.; 2012; 138, pp. 1022-1026. [DOI: https://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0000655]

21. Li, Z.; Bolton, M.D.; Haigh, S.K. Cyclic axial behaviour of piles and pile groups in sand. Can. Geotech. J.; 2012; 49, pp. 1074-1087. [DOI: https://dx.doi.org/10.1139/t2012-070]

22. Ye, C.L.; Cao, F.X.; Sun, H.; Wu, Z.X.; Zhang, T.; Sun, T.C. Experimental Study on the Shear Characteristics of the Interface between Marine Soft Clay and Jacked Pile. J. Mar. Sci. Eng.; 2024; 12, 1549. [DOI: https://dx.doi.org/10.3390/jmse12091549]

23. Kaynia, A.M.; Hebig, J.; Pein, T.; Shin, Y. Numerical model for dynamic installation of large diameter monopiles. Soil Dyn. Earthq. Eng.; 2022; 161, 107393. [DOI: https://dx.doi.org/10.1016/j.soildyn.2022.107393]

24. Moriyase, S.; Kobayashi, S.; Matsumoto, T. Experimental study on friction fatigue of vibratory driven piles by in situ model tests. Soils Found.; 2018; 58, pp. 853-865. [DOI: https://dx.doi.org/10.1016/j.sandf.2018.03.010]

25. Tsetas, A.; Tsouvalas, A.; Metrikine, A. A non-linear three-dimensional pile-soil model for vibratory pile installation in layered media. Int. J. Solids Struct.; 2023; 269, 112202. [DOI: https://dx.doi.org/10.1016/j.ijsolstr.2023.112202]

26. Zhou, J.; Liu, C.J.; Zhu, K.F.; Shen, P.P. Macroscopic behaviour and mesoscopic mechanism of shear strength characteristics of pile-soil interface under cyclic loading. Ocean Eng.; 2024; 297, 116998. [DOI: https://dx.doi.org/10.1016/j.oceaneng.2024.116998]

27. Zhou, W.J.; Wang, L.Z.; Guo, Z.; Liu, J.W.; Rui, S.J. A novel t-z model to predict the pile responses under axial cyclic loadings. Comput. Geotech.; 2019; 112, pp. 120-134. [DOI: https://dx.doi.org/10.1016/j.compgeo.2019.04.027]

28. Guo, C.; Lin, P.; Guo, P.; Wang, F.; Zhao, L. A Weibull-based damage model for the shear softening behaviours of soil-structure interfaces. Geotech. Res.; 2021; 8, pp. 54-63. [DOI: https://dx.doi.org/10.1680/jgere.20.00043]

29. Jitsangiam, P.; Pra-ai, S.; Boulon, M.; Jenck, O.; Chen, X.B.; Techavorasinsakul, S. Characterization of a soil-rough structure interface using direct shear tests with varying cyclic amplitude and loading sequences under a large cyclic testing cycle condition. Acta Geotech.; 2022; 17, pp. 1829-1845. [DOI: https://dx.doi.org/10.1007/s11440-021-01289-4]

30. Miller, G.A.; Hamid, T.B. Interface direct shear testing of unsaturated soil. Geotech. Test. J.; 2007; 30, pp. 182-191. [DOI: https://dx.doi.org/10.1520/GTJ13301]

31. Potyondy, J.G. Skin Friction between Various Soils and Construction Materials. Géotechnique; 1961; 11, pp. 339-353. [DOI: https://dx.doi.org/10.1680/geot.1961.11.4.339]

32. Desai, C.S.; Rigby, D.B. Cyclic Interface and Joint Shear Device Including Pore Pressure Effects. J. Geotech. Geoenviron. Eng.; 1997; 123, pp. 568-579. [DOI: https://dx.doi.org/10.1061/(ASCE)1090-0241(1997)123:6(568)]

33. Evgin, E.; Fakharian, K. Effect of stress paths on the behaviour of sand-steel interfaces. Can. Geotech. J.; 1996; 33, pp. 853-865. [DOI: https://dx.doi.org/10.1139/t96-116-336]

34. Kishida, H.; Uesugi, M. Tests of the interface between sand and steel in the simple shear apparatus. Géotechnique; 1987; 37, pp. 45-52. [DOI: https://dx.doi.org/10.1680/geot.1987.37.1.45]

35. Tsubakihara, Y.; Kishida, H. Frictional behaviour between normally consolidated clay and steel by two direct shear type apparatuses. Soils Found.; 1993; 33, pp. 1-13. [DOI: https://dx.doi.org/10.3208/sandf1972.33.2_1] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39982334]

36. Uesugi, M.; Kishida, H. Frictional Resistance at Yield Between Dry Sand and Mild Steel. Soils Found.; 1986; 26, pp. 139-149. [DOI: https://dx.doi.org/10.3208/sandf1972.26.4_139] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39972515]

37. Eid, H.T.; Al-Nohmi, N.M.; Wijewickreme, D.; Amarasinghe, R.S. Drained Peak and Residual Interface Shear Strengths of Fine-Grained Soils for Pipeline Geotechnics. J. Geotech. Geoenviron. Eng.; 2019; 145, 06019010. [DOI: https://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0002131]

38. Liu, Z.L.; Shi, J.Y.; Zhang, Y.C.; Xu, G.J.; Shu, S.; Li, Y.P.; Lei, G.H. A Novel Multifunctional Ring Shear Apparatus for Investigating the Interface Strength Characteristics of Liner Systems. Geotech. Test. J.; 2023; 46, pp. 20-43. [DOI: https://dx.doi.org/10.1520/GTJ20220031]

39. Martinez, A.; Frost, J.D. Undrained Behavior of Sand-Structure Interfaces Subjected to Cyclic Torsional Shearing. J. Geotech. Geoenviron. Eng.; 2018; 144, 04018063. [DOI: https://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0001942]

40. Stark, T.D.; Williamson, T.A.; Eid, H.T. HDPE Geomembrane/Geotextile Interface Shear Strength. J. Geotech. Eng.; 1996; 122, pp. 197-203. [DOI: https://dx.doi.org/10.1061/(ASCE)0733-9410(1996)122:3(197)]

41. Yoshimi, V.; Kishida, T. Friction between sand and metal surface. Proceedings of the 10th International Conference on Soil Mechanics and Foundation Engineering; Stockholm, Sweden, 15–19 June 1981.

42. Brumund, W.; Leonards, G. Experimental Study of Static and Dynamic Friction Between Sand and Typical Constuction Materials. J. Test. Eval.; 1973; 1, pp. 162-165. [DOI: https://dx.doi.org/10.1520/JTE10893J]

43. Corfdir, A.; Lerat, P.; Vardoulakis, I. A Cylinder Shear Apparatus. Geotech. Test. J.; 2004; 27, pp. 447-455. [DOI: https://dx.doi.org/10.1520/GTJ11551]

44. Dumitrescu, A.I.; Corfdir, A.; Frank, R. Influence of the Anisotropy of Confining Stress on the Sand/steel Interface Behaviour in a Cylinder Shear Apparatus. Soils Found.; 2009; 49, pp. 167-174. [DOI: https://dx.doi.org/10.3208/sandf.49.167]

45. Liu, J.W.; Cui, L.; Zhu, N.; Han, B.; Liu, J. Investigation of cyclic pile-sand interface weakening mechanism based on large-scale CNS cyclic direct shear tests. Ocean Eng.; 2019; 194, 106650. [DOI: https://dx.doi.org/10.1016/j.oceaneng.2019.106650]

46. ASTM D4253-16e1 Standard Test Methods for Maximum Index Density and Unit Weight of Soils Using a Vibratory Table; ASTM International: West Conshohocken, PA, USA, 2016.

47. ASTM D3080/D3080M-11 Standard Test Method for Direct Shear Test of Soils Under Consolidated Drained Conditions; ASTM International: West Conshohocken, PA, USA, 2011.

48. Wood, D.M. Experimental Observations of Shear Band Patterns in Direct Shear Tests; University of Cambridge Engineering Department: Cambridge, UK, 1982.

49. Jewell, R.A.; Wroth, C.P. Direct shear tests on reinforced sand. Geotechnique; 1987; 37, pp. 53-68. [DOI: https://dx.doi.org/10.1680/geot.1987.37.1.53]

50. Stone, K.J.L.; Wood, D.M. Effects of Dilatancy and Particle Size Observed in Model Tests on Sand. Soils Found.; 1992; 32, pp. 43-57. [DOI: https://dx.doi.org/10.3208/sandf1972.32.4_43] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39982283]

51. Huang, W.-C.; Sung, C.-Y.; Liao, H.-Y.; Chu, S.-S. Micromechanical behavior of granular materials in direct shear modeling. J. Chin. Inst. Eng.; 2015; 38, pp. 469-480. [DOI: https://dx.doi.org/10.1080/02533839.2014.998286]

52. API. 2GEO: Recommended Practice 2GEO–Geotechnical Foundation Design Considerations; American Petroleum Institute: Washington, DC, USA, 2014.