薄喷涂层(TSL)包覆支护对砂岩剥落破坏特性的影响

https://doi.org/10.1016/j.ijrmms.2025.106054获取权利和内容
全文访问

抽象

通过使用一种称为薄喷涂衬垫 (TSL) 的新型支撑材料,可以提高地下结构的安全性和稳定性。这种材料利用其卓越的粘合性能和抗拉强度,在岩石表面形成连续而坚固的保护层。然而,物模型在采矿扰动引起的动态加载条件下的性能需要进一步研究。为了解决这个问题,对具有不同涂层长度和厚度的 TSL 和砂浆涂层试样进行了分体式霍普金森压力棒 (SHPB) 剥落测试。采用高速摄影详细捕捉试件剥落破坏过程。实验结果表明,随着 TSL 涂层厚度的增加,TSL 与岩石之间的结合力减小。与未涂层砂岩试件相比,TSL 涂层试件的剥落强度较低。然而,剥落强度随着 TSL 厚度和长度的增加而增加。在涂层和岩石之间的界面处观察到初始开裂。砂浆涂层试样也发现了类似的模式,尽管它们表现出不同的支撑机制和更高的层开裂强度。此外,利用 FLAC3D-PFC3D 耦合模型来验证实验结果。数值模拟结果与相同冲击载荷下的实验数据非常吻合。但当冲击载荷增加时,TSL 涂层试样的剥落强度随着 TSL 涂层厚度和长度的增加而降低。研究结果有助于优化地下巷道支护结构的设计,为评价其稳定性提供参考。

关键字

动载荷
薄喷涂层 (TSL)
Bondig 力
剥落破坏
PFC3D-FLAC3D 耦合

1. 引言

剥落是地下硬岩矿中常见的动态拉伸破坏现象。1,2当巷道开挖引起的应力达到或超过其抗拉强度时,围岩将发生剥落破坏,对施工安全和地下矿山的稳定性构成重大危害。3因此,研究岩石的剥落破坏尤为重要。Split Hopkinson Pressure Bar (SHPB) 技术已被广泛用于研究岩石材料的剥落破坏特性。4,5以往的研究利用霍普金森压力杆探讨了岩石材料在预围压下的不同剥落特性,6,7温度8和原位应力。9,10结果表明,剥落作为一种渐进式破坏过程,会影响工程结构的长期稳定性。因此,在道路开挖后立即应用混凝土喷射混凝土支护,以减少围岩的暴露时间。这种方法有效地帮助控制和稳定爆破对围岩造成的扰动,从而提高其稳定性和支护能力。11,12然而,随着开挖深度的增加,传统的喷射混凝土支护方法面临施工效率低、粉尘浓度高、回弹率高等一些限制。13此外,在较大的动载荷下,喷射混凝土的高刚度和不足的韧性无法吸收岩石损伤的能量,导致其无法防止岩石失稳。这种破坏可能会导致喷射混凝土和岩石之间的分层,从而导致喷射混凝土板的形成并损害整个支撑系统。为了应对这些挑战,研究人员开发了一种新型柔性薄喷涂层材料 (TSL)。TSL 相对较薄 (2-5 mm),可快速凝固,并提供出色的粘合性能和高拉伸强度。14,15与传统的喷射混凝土材料不同,TSL 还表现出显着的渗透性,使它们能够穿透岩体内的裂缝和节理,从而增强周围岩石的稳定性。161718
目前,国内外研究人员通过室内实验和数值模拟对被称为薄喷涂层 (TSL) 的新型支撑材料进行了广泛研究,以更好地了解其性能和支撑机制。19202122 然而,大多数现有研究都集中在静态条件上。例如,Qiao 等人232425 研究了 TSL 的机械性能,包括抗拉强度、剪切强度和粘结强度,并得出结论 TSL 对岩石表现出优异的支撑性能。Shi et al.26,27通过实验探索了 TSL 材料的承载能力,发现 TSL 抑制了岩石变形。Chen 等人。28使用压缩、弯曲和拉伸测试研究了 TSL 在高湿度下的机械性能。结果表明,湿度增加会降低 TSL 的机械和密封性能,但会提高其粘合强度。伊尔马兹29,30还检查了 TSL 的拉伸强度和剪切粘合强度。此外,TSL 被认为是一种弹性材料,其机械性能受固化时间的影响。奥兹图尔克25使用实验室测试和数值模拟研究了固化时间对两种不同 TSL 材料弹性性能的影响。结果表明,随着固化时间的延长,TSL 的抗压强度和 Young 模量增加,而其变形能力降低。伊尔马兹31和古纳32研究了固化时间对 TSL 拉伸强度和机械性能的影响,发现拉伸强度和 Young 模量都随着固化时间的增加而增加,而断裂伸长率降低。Li 等人。33使用现场粘附测试来研究 TSL 和煤之间的粘合强度,结果表明,完整和断裂煤的粘合强度随着固化时间的延长而增加。然而,完整煤的粘结强度显著高于裂隙煤。
As a supporting structure, TSL adheres to the rock surface to form a continuous film, providing practical support. Existing studies and engineering practices have demonstrated that this membrane-like support can prevent or inhibit the occurrence of spallation damage.34 Currently, there are two primary mechanisms for preventing the movement or destruction of surrounding rock: the flexible thin coating support mechanism and the rigid thick coating support mechanism. Flexible thin coating support, which includes TSL, is characterized by its superior deformation ability. It primarily prevents the expansion of surrounding rock through its bonding and tensile properties, while shear performance playing a relatively minor role (see Fig. 1a). Conversely, rigid support, such as thick shotcrete, is distinguished by its high stiffness. It primarily inhibits the expansion of the surrounding rock through its shear performance, while its tensile properties are relatively less significant (see Fig. 1b).
Fig. 1
  1. Download: Download high-res image (598KB)
  2. Download: Download full-size image

Fig. 1. (a) Supporting properties of TSL (b) Supporting properties of thickly coated concrete.

Due to the close contact between the TSL and the rock surface, many researchers regard the bonding performance of TSL as a critical factor in assessing its support effectiveness.15 Bonding the TSL to the rock surface plays a vital role in preventing rock fracture failure. Therefore, it is necessary to study the key factors influencing the interfacial bonding performance between the TSL and the rock surface. Previous studies35, 36, 37, 38 have identified rock strength, surface roughness, surface integrity, and TSL coating thickness as crucial parameters affecting bonding performance. Ozturk39 proposed a method for testing the bond strength of TSL and examined the effect of coating thickness on TSL bond strength. The findings indicated that the bond strength of TSL decreases with the square root of the coating thickness. Furthermore, Ozturk40 explored the influence of rock properties and environmental conditions on TSL bond strength, highlighting that chemical reaction between the rock and the TSL are crucial for enhancing bond strength, whereas rock surface roughness has a minimal effect. Li et al.41 conducted laboratory adhesion tests to investigate the effects of rock strength and surface roughness on TSL bond strength, concluding that higher rock strength and greater surface roughness significantly improve TSL bond strength.
As a novel supporting material, the TSL forms a stable bond with the rock surface due to its excellent bonding ability and tensile properties, providing reliable support for underground engineering structures. While most previous studies have concentrated on the bonding performance of TSL under static loads, with less attention paid to its behavior under dynamic loads. Given that engineering environments such as underground roadways often encounter sudden dynamic loads, including earthquakes or blasting vibrations, these external forces potentially weaken the bond between TSL and rock, thereby impacting the stability of the overall supporting structure. Understanding the support characteristics of TSL under dynamic conditions is crucial for ensuring roadway safety. Since the bonding properties of TSL are directly related to its support effectiveness,15,42 it is essential to investigate these properties under dynamic loading conditions. TSL can be classified into rigid and flexible types based on their mechanical properties. Rigid TSLs are more effective in preventing rock fractures and movement, while flexible TSLs offer better support under small deformations. Flexible TSLs are particularly valued for their robustness, good deformability, bonding force, and superior energy absorption capacity.16,32 In this paper, we focus on examining the influence of TSL wrapping support on the spalling failure characteristics of sandstone under dynamic loading through SHPB test device. A method for determining the bonding force between TSL and rock under dynamic loading is proposed, and the influence of TSL thickness and length on the bonding force between TSL and sandstone is analyzed. Additionally, by comparing the spalling characteristics of mortar coating under dynamic load and employing PFC3D-FLAC3D numerical simulation, the supporting mechanism of TSL coating under dynamic load is further explored.

2. Test scheme

2.1. Specimen preparation

The TSL material used in this study consists of polymer powder, which includes quartz powder, titanium dioxide, white cement, binders, and other components. The manufacturer recommends a polymer powder-to-water weight ratio of 2:1, with a setting time not exceeding 40 min. When the thickness of TSL is less than 10 mm, it can fully cured within 48 h. For the cement mortar, ordinary Portland cement (Type 42.5) was used as the cementitious material, and river sand as the fine aggregate, with a particle size ranging from 0.8 mm to 3.0 mm. The mortar mixture ratio is provided in Table 1. The basic physical and mechanical parameters of the materials, as measured in the laboratory, are listed in Table 2. The rock used in this study is a homogeneous sandstone, with a length of 500 mm and a cross-sectional area of 35 mm × 35 mm, as shown in Fig. 2a. To achieve uniform TSL and mortar coatings of 3 mm and 5 mm thicknesses around the rock specimen, rectangular plastic molds with lengths of 100 mm and 150 mm and cross-sectional areas of 41 mm × 41 mm and 45 mm × 45 mm were prepared, respectively. The specimen schematic is shown in Fig. 2b.

Table 1. Mortar mixture ratio.

Sand/(kg)Cement/(kg)Water/(kg)Water cement ratio
5341701070.63

Table 2. Basic physical-mechanical parameters of material.

MaterialDensity/(kg/m3)Elastic modulus/(GPa)Compressive strength/MPaTensile strength/MPaWave velocity/(m/s)
Sandstone232620.00675.53000
TSL13651.06.5481419
Mortar21507.2374.751800
Fig. 2
  1. Download: Download high-res image (1MB)
  2. Download: Download full-size image

Fig. 2. Coating specimen and strain gauge pasting position (a) Sandstone specimen (b) Size of coating specimen (b) Schematic diagram of strain gage placement (d) Specimen preparation process (e) TSL-coated specimen (f) Mortar-coated specimen.

The process of preparing TSL and mortar is demonstrated in Fig. 2d. During the mixing process, to prevent air bubbles and lumps, the TSL powder is gradually added to the water and stirred thoroughly for 3–5 min before being poured into the prepared plastic molds. The molds were removed after 24 h, and then the specimens were placed in a curing box at 25 °C. Guner32 found that the strength and Young's modulus of flexible TSL after 7 days of curing can reach 70 %–80 % of their values after 28 days (complete curing). Considering the urgency and frequency of underground construction, the curing time for the flexible TSL material used in this paper is 7 days. For the mortar coating specimens, it is challenging to press them into the mold due to the thin thickness of the coating and the coarse aggregates in the cement mortar. Therefore, the evenly mixed cement mortar was poured directly onto the rock specimens, followed by vibration and smoothing. After one day, the mold was removed, and the specimen was placed in the standard maintenance room for curing. After 7 days of curing, the specimens were ready for testing. The finished TSL and mortar coating specimens are shown in Fig. 2e and f. In underground engineering, spraying is typically used for support. However, it is not convenient to create specimen from underground engineering, as this could damage the uniformity of the rock and TSL. Additionally, spraying makes it difficult to achieve a uniform thickness for the TSL coating. Therefore, this study employed molding technology to prepare the TSL coating specimens. It is also noted that after the excavation of an underground roadway, the in-situ stress on the surrounding rock boundary becomes zero. Thus, in this study, TSL was applied to long strip rock rod using a mold.
The TSL and mortar coating are placed 50 mm away from one end of the free surface of the specimen. This arrangement is made because the focus of the study is on investigating the effect of the TSL coating on the spalling characteristics of the sandstone specimens. A partial blank space is left at one end from the free surface to prevent the TSL coating from influencing the tensile waves generated by the reflection of stress waves at the free surface. Additionally, prior impact tests on pure sandstone specimens showed that cracks do not form 50 mm from the free surface. This indicates that leaving a 50 mm blank will not affect the failure mode of the specimen.

2.2. Test method

This article employs a SHPB device with a diameter of 50 mm to conduct spalling tests on TSL-coated specimens, aiming to examine the influence of TSL-wrapped support on the spall failure characteristics of sandstone. The experimental device includes a spindle-shaped striker, an incident rod, a track, ultra-dynamic strain gauges, a DL850 digital oscilloscope, a high-speed camera, and high-intensity lighting. The specifications of the device are detailed in Reference.43 Dynamic loading is controlled by a high-pressure nitrogen gas tank,44,45 and the schematic diagram of SHPB device is shown in Fig. 3. When the stress wave reaches the free surface, it reflects as a tensile wave. If the tensile stress exceeds the tensile strength of the specimen, it results in spalling failure. The maximum tensile stress at this point is considered the spalling strength of the specimen. As the stress wave continues, secondary or multiple spalling events may occur following the initial spalling.46
Fig. 3
  1. Download: Download high-res image (690KB)
  2. Download: Download full-size image

Fig. 3. SHPB experimental device schematic diagram.

Before the test begins, lubricant is applied evenly to the contact surface of the specimen and the incidence rod to reduce the friction between the contact surface. The punch of the SHPB system is spindle-shaped, which allows for the generation of an ideal half-sine wave, effectively minimizing the inertial effects during the test. After each impact, the striker is returned to the starting position to ensure consistent impact conditions. The propagation of the stress wave is monitored by attaching strain gauges to the surface of the specimen, with symmetrical placement as shown in Fig. 2c. Symmetric attachment of the strain gauges helps eliminate the influence of any surface defects on the strain signal. Additionally, due to the influence of stress wave propagation superposition and initial spallation range, the strain sticking positions are shown in Fig. 2c.
The test is conducted under an impact air pressure of 0.40 MPa. Given the short dynamic loading duration of approximately 250 μs, a high-speed camera is used to capture the crack and fracture process of the specimen. The high-speed camera has a resolution of 384 × 256 pixels and a frame rate of 50,000 frames per second fps, meaning one photograph is captured every 20 μs.

3. Test results

3.1. Adhesion calculation

When the striker impacts the incident bar, the stress wave is transmitted to the TSL-coated specimen, leading to spalling failure. The stress wave propagation within the specimen was monitored by attaching strain gauges to its surface. The loading waveform of specimen T-1-5 is shown in Fig. 4. From the figure, it is evident that the waveforms recorded by strain gauges B and C are nearly identical. However, during the propagation of the compression wave, there is an unavoidable delay in the signal recorded by strain gauge C. Given the close proximity of strain gauges B and C (see Fig. 2c), there is no significant loss in the superposition and propagation of the stress wave. The delay is attributed to the compressive deformation of the TSL caused by the deformation of the underlying rock, while in the region of the tensile wave, the delay between the two phenomena is considerably reduced. When the reflected tensile wave propagates, the undergoes tensile deformation earlier, and due to the excellent homogeneity of the TSL material, the delay in TSL tensile deformation is not significant. Since the strain gauge C is attached directly to the TSL material, and given material's homogeneity and its tight wrapping around the sandstone, we can reasonably assume that the strain measured by strain gauge C corresponds to the strain of the rock at that position. Furthermore, it is observed from the voltage signal of the pure sandstone specimen in the figure that the tensile stress of the pure sandstone specimen after the superposition of compression wave and tensile wave is greater in the pure sandstone specimen than in the TSL-coated specimen.
Fig. 4
  1. Download: Download high-res image (666KB)
  2. Download: Download full-size image

Fig. 4. Loading waveform diagram of sandstone and TSL-coated specimen.

To obtain the bonding force of the TSL, the strains measured by strain gauges B and C must be converted into stresses. The reflected tensile stresses for specimen TL-1-3 are shown in Fig. 5a. By subtracting the reflected tensile stress, we can determine the differences in tensile stress. The tensile stresses for T-1-3, T-1-5, TL-1-3, and TL-1-5 are 2.42 MPa, 1.61 MPa, 2.04 MPa, and 1.84 MPa, respectively. The measured stress can be converted into force according to Eq (1):(1)P=σ×a2where P is the bonding force, σ is the tensile stress, and a is the side length of the cross section.
Fig. 5
  1. Download: Download high-res image (333KB)
  2. Download: Download full-size image

Fig. 5. Reflected tensile stress and bonding of TSL-coated specimens.

The bonding force of TSL-coated specimens with different thickness and length values can be using Eq (1). Table 3 shows that the bonding forces for T-1-3, T-1-5, TL-1-3, and TL-1-5 are 2.964 kN, 1.972 kN, 2.499 kN, and 2.254 kN, respectively. From Fig. 5b, it is clear that the bonding force of the specimens decreases as the TSL wrapping thickness increases, and the specimens exhibit better bonding performance at a TSL wrapping thickness of 3 mm. A study by Ozturk23 demonstrated that the bond strength of TSL is inversely proportional to the square root of TSL thickness. Since the contact area between TSL and rock remains constant across different thicknesses, the bond strength is directly proportional to the bond force for a fixed contact area. As the TSL thickness increases, the bond force decreases, aligning with the observed results. This confirms that the calculation method proposed in this study to determine the bonding force is consistent with previous research and validates the feasibility of this approach.

Table 3. The spalling test results of TSL coating specimen.

NumberingStrain gage placement
Empty CellBCEmpty Cell
Reflection stress/MPaReflection stress/MPaAdhesive force/kN
T-1-35.54 MPa3.12 MPa2.964 kN
T-1-55.81 MPa4.2 MPa1.972 kN
TL-1-36.35 MPa4.31 MPa2.499 kN
TL-1-56.44 MPa4.6 MPa2.254 kN

3.2. Spalling strength and failure characteristics of TSL-coated specimens

The maximum reflected tensile signal, measured by strain gauges on the specimen's surface can be used to calculate the maximum tensile stress generated by the reflected tensile wave at different moments. This maximum tensile stress serve as the spalling strength value of the specimen. The spalling results for the TSL-coated specimens are recorded in Table 4, showing the spalling strengths of T-1-3, T-1-5, TL-1-3, and TL-1-5 are 5.54 MPa, 5.81 MPa, 6.35 MPa, and 6.44 MPa, respectively. For the pure sandstone specimens, the spalling strength is more prominent, with cracks forming closer to the free surface. This indicates that the reflected tensile stress generated by the stress wave reaching the free surface is more significant, quickly reaching the specimen's tensile strength, which results in cracks near the free surface. In contrast, the spalling strength for the TSL-coated specimens was significantly reduced, likely due to the combined effect of the bonding and tensile deformation resistance of the TSL. TSL is tightly wrapped around the rock, and its bonding force constrains the specimen, reducing the tensile deformation and preventing tensile failure in the wrapped section. This action consumes part of the energy and reduces the amplitude of the reflected tensile stress, thereby preventing the wrapped section of TSL from meeting the spalling conditions. Additionally, the tensile performance of TSL contributes to the inhibition of specimen deformation when the reflected tensile wave reaches the wrapped section of TSL. The counteracting tensile force generated by the TSL reduces the amplitude of the reflected tensile wave, preventing the wrapped section from experiencing spalling failure. It is worth noting that the spalling strength of TSL-coated specimens increases with the increase in TSL coating thickness and length. One possible explanation is that the larger bonding force causes a more pronounced reduction in the reflected tensile stress, leading to a lower spalling strength of the specimen. Another possible explanation relates to stress wave superposition within the specimen. The wavelength of the stress wave in sandstone is approximately 60 cm, and the fracture range is typically 15–30 cm. When the TSL coating length is 10 cm, the incident compression wave still exists due to the wavelength when the reflected tensile wave reaches the fracture. This causes the reflected tensile wave to be superimposed with the incident compression wave, reducing its amplitude. However, when the TSL coating length is increased to 15 cm, the strain gauge monitoring position is farther from the fracture, when the stress wave reaches the TSL coating length of 10 cm fracture, the reflected tensile wave continues to increase and propagate backward. At this moment, the incident compression wave is relatively small. This interaction results in an increased reflected tensile stress, which in turn raises the spalling strength of the specimen.

Table 4. Spalling results of different TSL coating specimens.

NumberingSpalling strength/MPaLayer number of spallingDistance from each spalling surface to the free end/cm
S-113.0113.3
T-1-35.54115.2
T-1-55.81115.3
TL-1-36.35120.1
TL-1-56.44120.3
In this study, the interface between the rock at a distance of 5 cm from the free surface and TSL, was defined as the first interface, while the interface at a distance of 15 cm or 20 cm from the free surface was defined as the second interface. As shown in Fig. 6, spalling cracks in TSL-coated specimens initiate at the second interface under the conditions of TSL coating thickness and length used in this study, whereas the pure sandstone specimen fails at 13.3 cm. This phenomenon may be attributed to the bonding and tensile capacities of TSL, which constrain the specimen and inhibit the expansion of tensile cracks, making the TSL-coated section being less prone to spalling failure. However, when the reflected tensile wave reaches the second interface between TSL and sandstone, the bonding and tensile capacity of the TSL material becomes ineffective, and the reflected tensile stress surpasses the tensile strength of the specimen, causing cracking at the second interface. These findings demonstrate that the TSL material provides significant support by enhancing the tensile strength of the rock, thereby reducing its susceptibility to tensile failure.
Fig. 6
  1. Download: Download high-res image (415KB)
  2. Download: Download full-size image

Fig. 6. Spalling failure characteristics of TSL-coated specimens.

3.3. Spalling strength and failure characteristics of mortar-coated specimens

The spalling results of mortar coating specimens are summarized in Table 5, showing that the spalling strength gradually decreases with an increase in mortar thickness and length, reaching its minimum in the SL-1-5 specimens. As illustrated in Fig. 1, mortar, being a brittle supporting material, primarily inhibits the expansion of surrounding rock through its shear resistance. It often requires a greater thickness to prevent rock damage and movement when acting on underground structures. With increasing mortar thickness, more energy is dissipated during stress wave propagation, leading to a reduction in the reflected tensile stress and the spalling strength of the specimen. This phenomenon is corroborated by the reflected tensile signals measured by strain gauge B, which indicate a decrease in tensile stress after the stress wave reflects off the free surface. Mudau47 has demonstrated that increasing the thickness of the concrete linings reduces spalling failure. Fig. 9 displays the loading waveforms of specimens S-1-5. Fig. 7 shows the loading waveforms of specimens S-1-5. The figure shows that there is almost no reflected tensile signal measured by the strain gauge attached to the mortar, which further proves that the mortar primarily prevents the damage and movement of the rock through its shear capacity. Moreover, as a brittle material with weak tensile deformation, the bonding effect of mortar is likely limited to a thin layer near the rock-mortar interface, rendering its overall thickness less impactful on bonding performance. Table 5 also reveals that the spalling strength of the mortar coating specimen remains lower than that of the pure sandstone specimen, indicating that mortar still offers reliable support as a supporting material. However, compared to TSL-coated specimens, mortar appears less effective in reducing spall strength. This difference may stem from the distinct mechanisms by which the two materials operate: while mortar relies on shear resistance to inhibit rock damage, spalling failure primarily arises from tensile stress. As a result, mortar allows for less energy dissipation during stress wave propagation, leading to greater tensile stress reflected from the free surface and, consequently, a higher spalling strength.

Table 5. Spalling strength of different mortar coating specimens.

NumberingSpalling strength/MPaLayer number of spallingDistance from each spalling surface to the free end/cm
S-113.0113.3
S-1-311.8115.3
S-1-511.3215.2, 17.1
SL-1-38.84120.0
SL-1-57.47120.1
Fig. 7
  1. Download: Download high-res image (339KB)
  2. Download: Download full-size image

Fig. 7. Loading waveform diagram of mortar-coated specimen.

Fig. 8a illustrates the spalling failure characteristics of the mortar-coated specimens, showing that the initial spalling crack occurs at the second interface and is larger in length compared to that of the pure sandstone specimen. The result suggests that the mortar coating effectively delays the formation of layer cracks. By wrapping the rock, the mortar provides sufficient resistance to breaking and movement, reducing the likelihood of spalling failure in the coated section due to reflected tensile waves. Spalling occurs at the second interface when the stress wave propagates beyond the mortar-wrapped section. Fig. 8b shows the spalling failure process of the S-1-5 specimen captured by using a high-speed camera, with red lines indicating the location of cracks. To facilitate the analysis, several representative spalling diagrams of the specimens were selected. The diagrams reveal that S-1-5 exhibits two spalling cracks, located 15.2 cm and 17.1 cm from the free end, respectively. In contrast, the T-1-5 specimen, which has the same thickness and length, displays only one spalling crack at the interface of 15.3 cm from the free end. This comparison highlights that TSL-coated specimens exhibit fewer spalling layers and lower spalling strength under identical conditions. These findings indicate that the superior tensile and bonding strength of TSL materials enables underground structures to provide more effective support.
Fig. 8
  1. Download: Download high-res image (462KB)
  2. Download: Download full-size image

Fig. 8. Spalling failure characteristics of mortar-coated specimens.

4. Numerical simulation

4.1. PFC3D built-in model and parameter calibration

In PFC3D, forces and moments are transmitted through particle-to-particle contacts, making it essential to set the appropriate microscopic parameters and select a suitable contact model to realistically simulate the fracture behavior of materials such as rock. PFC3D handles various built-in contact models, including planar joint contact model (FJM), parallel bond model (PBM), and smooth joint model (SJM). Among these, the parallel bonding model has been demonstrated to be particularly suitable for simulating rock due to its ability to accurately capture the material's mechanical properties and structural characteristics.48,49 Therefore, this study adopts the parallel bonding model to represent the contact behavior in rocks and similar materials. In contrast, since TSL is a flexible material with ductile characteristics compared to rock, the flat-joint contact model is deemed more appropriate for investigating the mechanical properties of TSL.
In PFC3D, the macroscopic properties of materials are primarily determined by adjusting microscopic parameters, such as particle radius, friction coefficient, and contact modulus, rather than directly employing macroscopic parameters. Accurately simulating mechanical properties and damage characteristics of materials using the discrete element method (DEM) necessitates the selection of appropriate microscopic parameters. Previous studies have established the relationship between microscopic and macroscopic parameters through PFC.50,51 For parameter calibration, commonly used techniques include the trial-and-error method52 and sensitivity analysis,53 with the trial-and-error method being the most widely adopted. Accordingly, this paper employs the trial-and-error method for parameter calibration. By iteratively optimizing and calibrating microscopic parameters, the numerical model is refined to accurately reproduce the macroscopic mechanical properties of the material. The calibration process involved simulating uniaxial compression tests, and the results are presented in Fig. 9. Table 6 lists the microscopic parameters of the numerical model used to replicate the macroscopic mechanical behavior of the specimen.
Fig. 9
  1. Download: Download high-res image (705KB)
  2. Download: Download full-size image

Fig. 9. Numerical simulation results in uniaxial compression (a) Stress-strain curve (b) Failure characteristics.

Table 6. Micro-parameters in numerical simulation.

Parallel bond modelFlat-joint model
Micro-parametersSandstoneMortarMicro-parametersTSL
Particle density (kg/m3)27002150Particle density (kg/m3)1365
Minimum radius of particle (mm)1.31.3Minimum radius of particle (mm)1.2
Maximum radius of particle (mm)1.81.8Maximum radius of particle (mm)1.6
Effective modulus of both bond and particle (GPa)167Effective modulus of both bond and particle (GPa)1
Frictional angle (°)4530Frictional angle (°)30
Stiffness ratio of both and parallel-bond and particle2.02.0Stiffness ratio of both and parallel-bond and particle2.0
Shear strength of parallel-bond (MPa)30 ± 218 ± 2Shear strength of bond (MPa)6 ± 2
Tensile strength of parallel-bond (MPa)56 ± 226 ± 2Tensile strength of bond (MPa)10 ± 2

4.2. Modeling of PFC3D-FLAC3D coupling

PFC3D software is developed based on the DEM and is typically used to simulate the micromechanical behavior of discrete media. In contrast, FLAC3D software is based on the Finite Difference Method (FDM) and is primarily applied to model the macroscopic mechanical behavior of continuous media. This paper employs a coupling method between PFC3D and FLAC3D to model the SHPB system. This approach can significantly improve computational efficiency compared to traditional discrete element methods.54,55 The specimen is modeled using PFC3D, which is well-suited for simulating material failure, such as in rocks, as well as the initiation and propagation of cracks. Meanwhile, the elastic model in FLAC3D is ideal for modeling SHPB devices, as it better satisfies the one-dimensional stress wave assumption.56 Consequently, FLAC3D was used to construct the metal incident rod. The numerical model is presented in Fig. 10a. The dimensions of the incident rod and specimen align with the experimental setup, and the position of measurement circle corresponds to the location of strain gauge. The DEM represents the specimen using particles, where the particle size range determines the number of particles under a given model size. The number of particles directly influences the stability and computational efficiency of the model. The selected particle size for this model includes 23,532 to 24,329 particles and 79,233 to 82,465 contacts. During the modeling process, a wall was generated at the interface between the specimen and the incident rod using the "wall-zone" command, and the "skip errors" command is employed to bypass any errors when stitching wall facets from zone faces. The coupling analysis of DEM-FDM is implemented through interface coupling. The workflow of the coupled PFC3D-FLAC3D numerical model is illustrated in Fig. 10b.
Fig. 10
  1. Download: Download high-res image (789KB)
  2. Download: Download full-size image

Fig. 10. Numerical simulation (a) Numerical model of SHPB system (b) FLAC3D-PFC3D coupling numerical simulation flowchart (c) Comparison of simulated and experimental waveform.

To verify the model reliability, the incident stress waves monitored at the same position in both the experiment and simulation were compared, as shown in Fig. 10c. The results indicate that the simulated waveform closely matches the experimental waveform, demonstrating the feasibility of using PFC3D-FLAC3D coupled simulation for the SHPB system. This section primarily focuses on modeling the failure process of TSL-coated specimens. For each combination of TSL coating thickness and length, a corresponding test model is constructed. Additionally, a sophisticated crack monitoring mechanism is implemented to track the evolution of crack development throughout the simulation.

4.3. Numerical simulation results and analysis

To analyze the characteristics of spalling failure, the propagation path of the stress wave in specimen T-1-5 was numerically simulated, and the corresponding stress cloud diagram is presented in Fig. 11a. As shown in this figure, when the stress wave reaches the TSL-wrapped section, the stress in the TSL lags behind that in the rock, as shown in Fig. 11a4 and 11a5. Subsequently, the stresses in the two materials become nearly identical, as shown in Fig. 11a6 and 11a7. This behavior aligns with the pattern of changes recorded by experimental strain gauges, confirming the validity of the proposed hypothesis and demonstrating the effectiveness of the method for calculating bonding forces. From the above observations, it can be concluded that at the same impact velocity, the pure rock specimen fractures at a distance of 13.3 cm from the free surface, whereas the rock specimens wrapped in TSL only fracture after the reflected tensile wave exits the TSL, i.e., the spalling fracture damage occurs at 15 cm and 20 cm from the free surface. According to the one-dimensional stress wave assumption of the SHPB, the stress wave does not propagate into the TSL, as corroborated by Fig. 11a. However, the TSL exhibits excellent bonding and tensile properties, tightly wrapping the rock. When the reflected tensile wave propagates to the TSL-wrapped section, the TSL deforms along with the rock deformation, consuming part of the reflected tensile wave's energy. This reduces the amplitude of the stress wave and prevents spalling fracture damage from occurring at the original location. Moreover, the strong bond between the TSL and the rock interface further resists tensile failure, absorbing additional energy and lowering the reflected tensile stress. As a result, spalling cracks do not initiate within the TSL-wrapped section. When the reflected tensile wave exits the TSL-wrapped section, the reflected tensile stress at this point exceeds the tensile strength of the rock, causing spalling cracking damage to occur immediately in the TSL-coated specimen, as shown in Fig. 11a7.
Fig. 11
  1. Download: Download high-res image (1MB)
  2. Download: Download full-size image

Fig. 11. Numerical simulation (a) Stress cloud diagram of T-1-5 specimen (b) Spalling failure characteristics of TSL-coated specimens.

Fig. 12a presents the stress time history curve of the TSL-coated specimen. The spalling strengths of specimens T-1-3, T-1-5, TL-1-3, and TL-1-5 are 5.65 MPa, 5.94 MPa, 6.56 MPa, and 6.85 MPa, respectively. These results indicate that the spalling strengths increase with the increase of TSL thickness and length, achieving the maximum spalling strength in specimen TL-1-5. Fig. 12b shows the spalling strength of TSL-coated specimens obtained from experimental and numerical simulations. While the numerically simulated spalling strengths are slightly higher than the experimental values, the two sets of results are in good agreement. Specifically, both the experimental and numerical data show a clear trend of increasing spalling strength with greater TSL thickness and length.
Fig. 12
  1. Download: Download high-res image (431KB)
  2. Download: Download full-size image

Fig. 12. Simulation results of TSL-coated specimens under impact loading.

Fig. 11b illustrates the spalling failure characteristics of the TSL-coated specimens. The spalling cracks were observed to occur at the second interface, with only one spalling crack forming, consistent with the experimental results. Although only a spalling crack appears at the second interface, PFC3D's built-in FISH programming language allows for recording detailed information about microcracks within the model. By invoking relevant FISH functions, comprehensive crack data can be retrieved. As shown in Fig. 11b, the number of microcracks in the specimens is presented. The microcracks provides an indication of the extent of specimen failure, where a higher number of microcracks corresponds to more severe failure. Specifically, the microcrack counts for T-1-3, T-1-5, TL-1-3, and TL-1-5 are 307, 333, 401, and 438, respectively. These results indicate that the number of microcracks increases with the thickness and length of the TSL, suggesting that the severity of spalling failure also increases with these parameters. The severity of spalling failure can be correlated with the spalling strength, as more severe spalling failure generally corresponds to higher spalling strength. This finding aligns with the earlier experimental observations, further validating the consistency between numerical simulations and experimental results.

5. Discussion

The test results indicate that the spalling crack in the specimens occur exclusively at the second interface, with no evidence of multiple spalling events. To further analyze the spalling failure characteristics of TSL and mortar-coated specimens from a microscopic perspective, numerical simulations were conducted for specimens with varying lengths and thicknesses under increased impact loads. This approach aimed to provide a detailed investigation of the underlying failure mechanisms.
When the impact load reaches 45 MPa, the spalling failure characteristics of the TSL-coated specimens are illustrated in Fig. 13a. Multiple laminar cracking are observed in the specimens. For a TSL coating thickness of 3 mm, the initial laminar cracking occurs at the first interface, whereas for a TSL coating thickness of 5 mm, the initial laminar cracking occurs at the second interface. Notably, no damage is observed in the rock or TSL material within the TSL wrap region, likely due to the superior supporting performance of the TSL. When the rock undergoes slight deformation, the TSL effectively protects protect the rock structure. Wang et al.57 demonstrated that the absorption performance of TSL under dynamic loading increases with thickness, significantly raising the energy required to fracture sandstone with a TSL coating. Under the action of a one-dimensional stress wave, the stress wave only propagates within the rock. However, TSL's excellent bonding and deformation capabilities allow it to reduce tensile deformation in the rock while dissipating the energy of the stress wave. Furthermore, the TSL is tightly wrapped around the rock surface, which increases the energy required to break the rock and reduces the likelihood of spalling failure in the TSL-wrapped region. For a TSL thickness of 5 mm, the increased thickness enhances its tensile capacity. Under higher load impacts, the specimen consumes more energy during tensile and compressive deformation, preventing the tensile wave reflected from the free surface failing to exceed the tensile strength of the rock over a short propagation distance. Consequently, no fracture occurs at the first interface. However, as the reflected tensile wave reaches the second interface, the net tensile stress surpasses the rock's tensile strength, leading to the generation of initial laminar cracking, followed by secondary laminar cracking. In contrast, when the thickness of TSL is 3 mm, although the bonding force is significant, the reduced thickness leads to less energy consumption during tensile deformation. As a result, the tensile wave reflected from the free surface more easily reaches the rock's tensile strength, causing initial spalling occurs at the first interface. After the initial spalling occurs, the subsequent reflective stresses diminish, the strong bond provided by the 3 mm TSL prevents further damage to the rock within the TSL-wrapped region. Fig. 13a also shows that the number of spalling layers varies among the specimens, with specimen T-1-3 exhibiting the highest number of spalling cracks—four in total. This variation is attributed to the influence of impact load on the bonding force and tensile capacity of the TSL. TSL primarily protects the rock through its bonding force and tensile strength. With a smaller TSL thickness, the bonding force is more significant, but the tensile capacity is reduced. Conversely, with a greater TSL thickness, the tensile capacity increases while bonding force decreases. At lower impact loads, the bonding force of the TSL dominates, resulting in less damage to specimens with a 3 mm TSL coating. However, as the impact load increases, the role of the tensile capacity of the TSL becomes more significant, leading to reduced spalling damage in specimens with thicker coatings, such as TL-1-5.
Fig. 13
  1. Download: Download high-res image (630KB)
  2. Download: Download full-size image

Fig. 13. Spalling results of TSL-coated specimens under impact loading (a) Spalling failure characteristics (b) Stress time history curve.

Fig. 13b illustrates the spalling strength of TSL-coated specimens under numerical simulation. The spalling strengths of specimens TL-1-5, T-1-5, TL-1-3, and T-1-3 increase sequentially, with the maximum spalling strength observed in specimen T-1-3. This finding aligns with the analysis presented earlier. Thicker TSL-wrapped specimens demonstrate enhanced deformation capacity, allowing them to effectively absorb energy during tensile and compressive deformation under higher impact loads. This energy absorption reduces the spalling strength of thicker specimens. Conversely, thinner TSL coatings are less effective at dissipating energy, resulting in higher spalling strengths. Consequently, specimen T-1-3 exhibits the highest spalling strength, corresponding to more severe spalling failure.
When the impact load is 45 MPa, the spalling failure characteristics of the mortar-coated specimens are shown in Fig. 14a. The results indicate that the spalling failure is less severe in specimen SL-1-5, it is more pronounced in specimenS-1-3 specimen. The degree of spalling failure increases as the mortar thickness decreases. As previously discussed, the mortar wrapping layer primarily stabilizes the specimen by providing shear capacity, which helps protect the surrounding rock from failure. During stress wave propagation within the specimen, the mortar's shear capacity dissipates part of the energy, thereby mitigating the severity of spalling failure. Fig. 15a presents the shear stress monitored by the measuring circle within the mortar. The data reveal that shear stress decreases as the thickness of the mortar increases. This observation may be attributed to the increased shear resistance of the mortar as its thickness increases. Enhanced shear resistance leads to greater energy dissipation by the stress wave during its propagation, resulting in a reduction in shear stress. Additionally, as shown in Fig. 15b, the shear stress monitored on the TSL decreases with increasing TSL thickness and is consistently lower than that measured on the mortar. This difference can be explained by the distinct supporting mechanisms of TSL and mortar. TSL primarily provides support through its bonding force and tensile properties, while mortar relies on its shear resistance. Since spalling failure is fundamentally a tensile failure, TSL's superior tensile capacity enables greater energy dissipation of stress waves during propagation. This results in significant stress attenuation within the specimen. Furthermore, the results of the reflected tensile stress measured on the TSL are shown in Fig. 15c, it can be seen that the tensile stress measured under the TSL wrapping with a thickness of 5 mm is less than the tensile stress under the TSL wrapping with a thickness of 3 mm. This finding aligns with the experimental results, corroborating the conclusion that thicker TSL layers more effectively reduce tensile stress in specimens. With the increase of TSL thickness, it improved deformation ability effectively absorbs the energy released during the tensile and compressive deformation of the specimen. This results in a reduction in the reflected tensile stress, thereby mitigating the degree of spalling failure. In specimen SL-1-3, three spalling cracks are observed: one at the first interface, one at the second interface, and one within the mortar-wrapped section. When the impact load increases, the tensile stress reflected from the free surface also increases. Upon propagating to the first interface, if the spalling condition is met, an initial spalling crack forms. This behavior aligns with the initial spalling cracks observed in the SL-2-3 specimen during experiments (see Fig. 14b). However, due to the lower impact load used in the experiment, the secondary crack appears at a distance of 30 cm from the free surface. Once initial spalling occurs, part of the stress wave energy is dissipated within the fracture section, while the remaining wave propagates further. Because the initial spalling location is close to the free surface, the un-incident stress waves treat the fracture surface as a new free surface. This creates additional reflected tensile stresses, which satisfy the spalling condition in the mortar-wrapped section, leading to the formation of a second crack layer. The third crack layer then develops at the second interface due to continued stress wave propagation and reflection. As shown in Fig. 14a, the failure of specimen S-1-5 is less severe than that of specimen SL-1-3. Under higher load impacts, the enhanced shear resistance of the mortar significantly improves its ability to protect the rock from damage. The energy absorbed within the specimen increases with the thickness of the mortar wrapping. This greater energy dissipation helps to maintain the stability of the surrounding rock and effectively prevents damage within the mortar-coated section.
Fig. 14
  1. Download: Download high-res image (586KB)
  2. Download: Download full-size image

Fig. 14. Spalling results of mortar coating specimens under impact loading (a) Numerical simulation (b) Spalling test.

Fig. 15
  1. Download: Download high-res image (804KB)
  2. Download: Download full-size image

Fig. 15. Time-history curve of stress (a) Shear stress of mortar coating specimens (b) Shear stress of TSL coating specimens(c) Normal stress of TSL coating specimens.

Compared to the spalling failure characteristics of TSL-coated sandstone under the same impact level, mortar-coated sandstone exhibits more severe spalling failure. This difference can be attributed to the nature of the failure mechanism, which is primarily driven by tensile stress. TSL, as a supporting structure, helps prevent the rock damage through its bonding force and tensile properties. During stress wave propagation, the bonding force and tensile deformation ability of TSL effectively absorb energy, thereby reducing the reflected tensile stress and minimizing spalling damage. In contrast, as a brittle material, mortar mitigates rock damage through its shear capacity. However, the energy consumed during the stress wave propagation is limited, leading to higher reflected tensile stress and more severe spalling failure. These findings suggest that TSL provides superior support compared to mortar under the same thickness and length.
In actual underground mining operations, frequent dynamic disturbances can cause damage to the surrounding rock. As a brittle support material, mortar is prone to cracking when the surrounding rock experiences small deformations. Therefore, the "sacrificial support" method is often used to protect the surrounding rock.58 In contrast, the new supporting material TSL, offers better flexibility and can absorb energy more effectively under frequent dynamic disturbances, thereby protecting the surrounding rock from failure. Additionally, TSL can tightly adhere to rocks and even penetrate the cracks, forming interlocks that enhance the bonding strength and provide better protection against rock damage. When subjected to frequent dynamic disturbances, TSL can prevent rock failure caused by small deformations due to its superior deformation ability, bonding strength, and energy absorption capacity. Even if the surrounding rock undergoes slight movement, TSL still tightly adheres to the surrounding rock, exhibiting good support performance and effectively reducing the extent of rock failure. This approach is fundamentally different from the "sacrificial support" method used with mortar. In summary, TSL offers better protection for the surrounding rock, reducing the likelihood of damage during frequent underground disturbances.

6. Conclusion

This study combined laboratory tests and FDEM numerical simulation to investigate the effectiveness of TSL wrapping support in mitigating spalling failure in sandstone under dynamic loading. High-speed photography was utilized to observe the spalling failure characteristics of sandstone coated with TSL. A method for determining the bonding force between TSL and rock under dynamic loading was developed, and the influence of TSL thickness and length on this bonding force were analyzed. The support advantages of TSL materials under frequent dynamic loads were further validated by comparing the spalling strength and failure mode of mortar-coated specimens under identical conditions. The main conclusions of this study are listed as follows.
  • (1)
    This paper proposed a method to determine the bond force between TSL and rock under dynamic loading. The results showed that the bond force decreased as the thickness of the TSL increased.
  • (2)
    The excellent adhesion and tensile capacity of TSL effectively prevented tensile damage to the sandstone. Under impact loading (at lower levels), the spalling strength of TSL-coated specimens increased with the TSL's thickness and length. Specimen TL-1-5 exhibited the highest spalling strength among the tested specimens, although its spalling strength was still significantly lower than that of pure rock specimens. In contrast, the strength of sandstone spalling under mortar wrapping decreased as the mortar thickness and length increased, as the mortar primarily prevents sandstone damage through its shear capacity. Consequently, the spalling strength of mortar-wrapped specimens was greater than that of those wrapped in TSL. Furthermore, neither the TSL-coated nor mortar-coated specimens showed initial spalling at the wrapped section, with the first spalling fracture occurring at the second interface.
  • (3)
    The simulation results demonstrated that both tensile and shear capacities were enhanced as the thickness TSL and mortar wrapping increased. As the impact load increased, the tensile properties of TSL were fully utilized, consuming more energy during the tensile and compressive deformation of the specimen, thereby reducing the degree of spalling failure. In contrast, the shear action of the mortar resulted in greater energy dissipation with thicker wrapping, further mitigating spalling failure. Overall, under dynamic loading, the bonding force and tensile ability of TSL significantly outperformed its shear performance when compared to traditional mortar. This improved performance made TSL particularly well-suited for supporting underground roadways subject to frequent dynamic disturbances. TSL not only provided superior support but also adhered firmly to the surrounding rock, effectively safeguarding it from potential damage.
In conclusion, this paper provided valuable insights into the design of TSL support and reinforcement for underground tunnels, offering a foundational understanding of its stability evaluation. However, here were areas that could have been improved: 1) The differences between the TSL support method used in laboratory tests and the actual TSL spray support in the field may have resulted in slightly varied support effects; 2) Previous studies indicated that the typical thickness of TSL ranges from 3 mm to 5 mm. However, this study did not explore the impact of thicker TSL layers or examine how the age of TSL affected the spalling failure characteristics of sandstone. Therefore, future research will focus on investigating the influence of TSL age on its performance in relation to spalling failure.

CRediT authorship contribution statement

Shiming Wang: Writing – review & editing, Validation, Resources, Methodology, Investigation, Conceptualization. Yunfan Bai: Writing – original draft, Methodology, Investigation, Data curation. Wentao Long: Validation, Data curation. Qiuhong Wu: Investigation, Data curation. Chuanqi Li: Writing – review & editing. Jian Zhou: Writing – review & editing, Investigation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The work was supported by the National Natural Science Foundation of China (Grant Nos. 42177164, 52474121, 52374089 and 51604109) and the Scientific Research Foundation of Hunan Province Education Department (Grant Nos.22B0507 and 22A0351).

Data availability

Data will be made available on request.

References

Cited by (0)

View Abstract