Study on a Transparent Similar Rock-Anchoring Structure under Impact Tests and Numerical Simulation Tests

Abstract

Interface slip is one of the main failure forms of an anchor structure, but the debonding slip process of rockbolt cannot be directly observed under impact load, so it is important to study the failure mechanism of the interface debonding of an anchor structure. For this purpose, a kind of transparent similar rock-anchoring structure and a new rockbolt impact load test system were developed. The debonding process of the anchor structure was carried out based on the system and Particle Flow Code (PFC) simulation test. The main conclusions are as follows: Axial force of rockbolt decreases progressively from the load end to the other end, but the shear stress of the interfaces increases at the early stage and then decreases. The majority of shear stress is at I interface compared to shear stress at II interface. The crack in the transparent sample occurs first in the middle of the sample, then extends to the local position of the two interfaces and finally penetrates all cracks, and the debonding failure mode is dual-interface mixed. According to laboratory tests and PFC simulation tests, the crack growth process can be divided into three stages. The results of the PFC simulation tests are similar to those of the laboratory impact tests of rockbolt.

1. Introduction

With the increase in coal mining depth in China, deep roadways face complex conditions, such as high ground stress or disturbing force, which often lead to the occurrence of rock or coal burst [1,2,3,4,5,6]. A large number of exposed rockbolts can cause the destruction of the roadway support structure, resulting in serious casualties and economic losses [7,8,9,10]. There are two important reasons for the failure of a rockbolt anchor structure under impact load. First, the surrounding rock of the deep roadway can store a significant amount of strain energy due to the effect of the high initial stress of the original rock; the rockbolt is then in a high-bearing stress environment after the deformation of the surrounding rock. Second, rockbolts are often affected by dynamic loads. These disturbances carry energy and spread in the form of stress waves and become the major inducement of rockbolt failure in deep engineering.

Many experts have carried out field and laboratory experiments and theoretical analyses on rockbolts [11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26]. Previous research on the dynamic response of rockbolt mostly focuses on non-damage detection technology, analyzing the vibration characteristics of the anchorage system under various excitation forces. According to the strain rate during loading, acoustic stress wave propagation in an anchorage system can be placed in the static load category. In practical engineering, blasting vibration, mine shock, rock burst and other dynamic loads are mostly in the category of medium and high strain rate. Debonding failure modes of anchor structures are distinctive with different loading modes. Rockbolt and anchoring agents are pulled out as a whole during the course of the pull-out test. Due to the opacity of the sample, the shear slip trace of the anchoring agent cannot be directly observed and judging the debonding failure only from the end failure is not convincing. It may be a single interface debonding failure mode, or a double interface debonding failure mode. These studies have contributed significantly to our understanding of the mechanical behavior of rockbolts under static or quasi-static tensile conditions, but without reference to the influence of dynamic loads. Under dynamic loads, the rockbolt in high stress surrounding rock shows different mechanical responses from static load [27,28,29]. Under certain conditions, even a small disturbance may induce the failure of rockbolt support in high stress engineering, resulting in the occurrence of deep rock mass disaster. The impact resistance of a rockbolt supporting structure is a research hot spot [30,31,32]. Many test methods have been extensively studied in the laboratory to verify the dynamic performance of rockbolts [21,22,31,33]. The large amount of test data provided by dynamic impact test equipment are very valuable and have greatly promoted the development of rockbolt supports. The cumulative dynamic absorption energy of D rockbolt and J Rockbolt are measured by the canmet-MMSL testing machine, and their values reach about 46.5 kJ [15,16,34,35].

In recent years, the development of numerical simulation has improved the research status of the dynamic performance of rockbolt. A numerical model of fully grouted rockbolt was made and analyzed by FLAC3D in order to investigate the influence of various parameters in the double shear test [36]. The stress characteristics of end-anchored rockbolt could be divided into three stages with the impact time [37]. The rockbolts of the dynamic test results are simulated and verified by LS-DYNA [38]. In contrast to the above, rockbolts, or end-anchored rockbolts, absorb the kinetic energy and fracture energy of rock mass by their free deformation.

Although many studies on the mechanical response mechanism of rockbolt under medium and high strain rate of dynamic load have shown successful results, the mechanical properties of rockbolt under dynamic load need to be further studied. For example, the difference between the axial stress distribution characteristics and failure mode of the interface between rockbolt and anchor agent under dynamic load and static load needs to be explored. In view of these problems, Section 2 exhibits the test device of a transparent rock sample and anchor structure under impact load, Section 3 analyzes the distribution law of rockbolt stress and failure mode based on the experimental results, and Section 4 presents the debonding process of rockbolt from micro crack to macro cracks, based on PFC.

In this article, the bearing capacity and failure mode of an anchor structure are studied under dynamic load. Furthermore, the crack evolution process and failure mechanism of the dynamic impact test of rockbolt, which will help to clarify a roadway support design for a dynamic disaster in deep mines, are revealed.

2. Impact Load Test System of Transparent Similar Rock-Anchoring Structure

2.1. Test Device and Working Principle

1.

Test device

The test device is composed of impact device, fixed device and monitoring device, as shown in Figure 1.

Impact device: The developed single-stage electromagnetic gun is used as the impact device, which is composed of a capacitor box, launcher and bullets. The launcher has a barrel length of 1200 mm and an inner diameter of 42 mm. The bullet has a length of 400 mm, diameter of 40 mm and a weight of 4 kg.

Fixed device: In order to avoid deviation in the sample during the impact process, the anchoring support is welded with steel plates of 10 mm thickness, and its size is 100 mm × 100 mm × 300 mm. The steel plate behind the support with a drilled hole of 50 mm diameter, plays a buffer role, and two rockbolt holes are reserved on two sides in order to fix them, as shown in Figure 2.

Monitoring device: The electromagnetic gun drives the bullet to impact the rockbolt on the anchoring support, the dynamic strain tester records the data and a high-speed camera captures the whole impact test process, as shown in Figure 3. A DH5922N model of the dynamic signal test analysis system (32 channels) is used to collect strain gauge data. A high-speed camera system composed of Phantom-MIRO-M310 Camera and Phantom Camera Control software was used. When the voltage of the capacitor box is 5000 V, the corresponding velocity of the bullet is 20 m/s.

2.

Working principle

The basic principle of the electromagnetic gun: the capacitor is charged by the boost circuit and discharges to the coil. The coil generates a strong magnetic field and an electromagnetic force fires the bullet out, as shown in Figure 4.

2.2. Test Scheme

(1)

Production process of the measuring force rockbolt

In order to measure the axial force of the rockbolt in the impact load test, measuring force rockbolt is made. The production process is as follows:

• Notched rockbolts with a diameter of 20 mm and yield strength of 335 MPa of a left-handed rebar anchor are selected, symmetric slots are made along the length of the rockbolt direction; its size is 8 mm in width and 3 mm in depth, and the grooves must be straight and without obvious steps, as shown in Figure 5.
• Measuring point Weld-free strain gauge of BF120-3AA type is adopted. Ten strain gauges are pasted in the slot of the rockbolt. Ten strain gauges are pasted 3 mm away from the II interface, as shown in Figure 6.

In order to ensure the uniform distribution of the measured data, measuring points were located 5 mm away from the left end and the other measuring points were arranged from end to end of the rockbolt with 800 mm spacing.

(2)

Production process of a similar transparent rock sample

The transparent similar rock of the sample is composed of six mechanical glass plates with a size of 300 mm × 100 mm × 100 mm. A rockbolt hole with a dimeter of 26 mm is drilled in the front and back of the glass plate, and two fairleads with a 3 mm diameter are drilled on the top and bottom of the rockbolt hole, as shown in Figure 7.

Small rubber blocks containing transparent quartz sand and transparent solution are configured. Due to the fluidity of epoxy resin glue solution, the accurate placement of a strain gauge cannot be realized. In order to obtain accurately measured data, the casting process is repeated three times. The specific preparation steps are as follows:

• Before initial pouring, a 10 mm transparent glue block is prepared and randomly poured into the mold to the height of the sticking strain gauge, and vacuumed transparent glue is slowly drained into the mold with a glass rod until the small glue block is submerged. Then, it is left to stand for 24 h, as shown in Figure 7a,b.
• After initial setting, the strain gauges are pasted and numbered in the predetermined position. A silica gel tube with an outer diameter of 26 mm is inserted into the reserved hole of the mold, as shown in Figure 7c,d. Then, some 10 mm transparent blocks are poured to the height of the strain gauge on the second layer.
• After the second strain gauge is laid, the mold is filled with small rubber blocks and then the transparent glue is poured. After standing for 24 h and removing the silica, the gel tube is left to stand for another 24 h. Then, it can be removed from the mold, as shown in Figure 7e,f.
• Repeating the above steps three times, as shown in Figure 7, 48 h before the test, the model without the anchoring agent is placed in −40 °C cryogenic refrigerator. Then, 24 h before the test, the specimen is taken out and the anchoring agent and rockbolt are installed.

3. Analysis of Test Results

3.1. Distribution Law of Rockbolt Axial Force and Interface Shear Stress

Figure 8a shows the distribution of rockbolt axial force along the anchoring depth. It can be seen from the figure that the maximum axial force of rockbolt is 180 kN, and the minimum axial force of rockbolt is 30 kN. The axial force of rockbolt shows a decreasing trend from the dynamic load end to the other end. Figure 8b shows the distribution of shear stress at I and II interface with anchorage depth. In general, shear stress at I and II interface increases first and then decreases. Maximum value is about 18 MPa, and minimum value is about 6.2 MPa at I interface. Maximum value is about 14.3 MPa, and minimum value is about 9.2 MPa at II interface. Compared with I interface, shear stress fluctuation of II interface is relatively small.

3.2. Failure Mode Analysis

The high-speed camera captured the debonding failure process of rockbolt, and five pictures with obvious changes were selected, as shown in Figure 9.

During debonding with the transparent rock sample, white spots were produced on the anchorage agent, indicating that it slipped. Combined with the schematic diagram of debonding failure process, the process is divided into three stages:

In the crack initiation stage (459.7 ms), the first part of the rockbolt suffers from the bigger impact load, but the affected zone of impact load is lesser. Maximum shear stress occurs at I interface. At the local position of the anchoring agent faint white spots occur, which indicates a debonding slip at I interface. The anchoring agent suffers from shear crack in the middle of the rockbolt, as shown in Figure 9b.

In the crack coalescence stage (459.7 ms~461.2 ms), when the impact load transfers to the middle of the sample, the affected zone of impact load spreads to both sides, and the whole sample bears the impact load, which results in cracks in the middle of the anchoring agent. The increase in white spots on II interface is especially obvious. At the back end of the sample, the rockbolt and the anchoring agent move slightly, and the shear cracks of the anchoring agent in the middle of II interface gradually increase and coalesce. There is no large slip in the first half of the interface of the anchorage agent, and local position has sporadic slip marks; there are also obvious shear slip marks in the second half of the anchorage agent, and the scratches are serration, as shown in Figure 9c.

In the debonding slip stage (461.2 ms), when the load transfers to the left end, white spots at the anchoring agent of II interface gradually become bigger, and a few more white spots appear. The crack on the anchoring agent expands gradually, and the rockbolt and anchorage agent experience a significant slip, indicating debonding at both interfaces. The tear zone of the anchoring agent in the middle of II interface further increases, as shown in Figure 9d. The white spots on the II interface continue to increase, and the relative slip of the rockbolt and anchorage agent gradually increase; the tearing area of the anchoring agent also reaches its maximum. Due to the mechanical bite between the transverse rib of the rockbolt and the anchoring agent, it is difficult to observe a relative debonding slip. However, the intine of the reserved hole is smooth, and the rockbolt-anchoring agent appears to slip relative to the surrounding rock sample. Therefore, the rockbolt and anchoring agent are rushed out in the latter part, as shown in Figure 9e.

To sum up, the debonding failure process of the anchoring agent under impact load is dual-interface mixed, as shown in Figure 10.

4. Numerical Simulation

4.1. PFC Brief

Unlike the continuum mechanics methods, PFC attempts to study the mechanical properties and behavior of media from the perspective of the microstructure. The program is more suitable for describing mesoscopic-macroscopic crack propagation, failure accumulation and impact damage in solid materials. PFC2D includes a contact bonding model and a parallel bonding model. For rock-like materials, a parallel bonding model is an ideal model.

4.2. Numerical Modeling

The surrounding rock and anchor agent are composed of randomly generated particles, represented by blue and green particles, respectively. Rockbolt and bullet are composed of regularly arranged particles, represented by black particles, and a total of 23,393 particles generated, as shown in Figure 11a. The axial force of rockbolt is recorded by using a large circular measurement, and the shear stress of I interface and II interface are recorded using a small circular measurement, which generated 319 measuring circle; its distribution is shown in Figure 11b. The axial force of rockbolt and the shear stress of the anchoring agent are obtained by stresssxx and stressxy order, respectively. The mechanical parameters of the anchored surrounding rock are obtained through uniaxial compression test and Brazilian splitting test, and the mechanical parameters of the anchor rod and the anchoring agent are provided by the manufacturer.

Table 1. Main parameters of PFC2D.

Parameter	Surrounding Rock	Anchoring Agent	Bolt
ball radius (mm)	5.0 × 10−1–7.5 × 10−1	6.0 × 10−1–8.0 × 10−1	1.0
density (kg/m3)	2.1 × 103	2.0 × 103	1.2 × 103
normal stiffness	2.4 × 109	1.2 × 109	40.0 × 109
stiffness ratio	1.9	1.0	2.8
parallel bond normal stiffness	3.8 × 109	1.0 × 109	33.0 × 109
parallel bond normal-to-shear stiffness ratio	3.9	2.0	2.8
dashpot normal critical damping ratio	0.5	0.5	0.5
parallel bond tensile strength	2.9 × 107	6.0 × 107	1.0 × 1010
parallel bond cohesion	6.1 × 107	2.4 × 107	10.0 × 1010

shows the microscopic mechanical parameters, and the model parameters are consistent with the results of the laboratory test.

4.3. Comparative Analysis of Numerical Simulation and Impact Test Results

Figure 12a–c, respectively, expressed the axial force curve of rockbolt and the shear stress curve of I and II interface in PFC2D. As can be seen from the figure, the axial force of rockbolt gradually decreases from the load end to the other end, which is similar to the result in the laboratory test. Peak values appear at the shear stress curves of I and II interface in the latter half of the anchoring structure, and the shear stress is slightly larger than the laboratory data. Because strain gauge pasted numbers are few in the lab test, the measurement positions are adjacent to the interface, making it impossible to accurately measure the shear stress of interfaces. However, numerical simulations can obtain the axial force and shear stress of measured circles. The particles of the adjacent measured circle may exist at different bonding states, rotation capacity, contact force and so on, which affects shear stress. Shear stress fluctuation in PFC simulation is slightly large, and there is some error with laboratory test, but the curve obtained by PFC is more comprehensive, and variation tendency is similar to the result of the laboratory test.

Figure 13 shows the relationship between the loading step and the total number of cracks. With loading, the total number of cracks in the anchorage structure increase gradually. The total number of cracks includes old cracks and new cracks, and it is a cumulative number. Figure 14 shows the debonding evolution process of the simulation test. In order to highlight the cracks, cracks are presented as red. Figure 14b–e corresponds to Figure 10 of the laboratory test. The whole impact process can be divided into three stages:

Crack initiation stage (O–B), which is when the particles’ stress is greater than the particles’ bonding force, about 3000 steps, and the bonding force chain shows breakage. At this stage, the cracks generated are only 5% of the total number of cracks, and the growth rate of the crack is small and slow; II interface cracks are more than I interface, as shown in Figure 14b.

In the crack coalescence stage (B–C), the number of cracks accounts for about 75% of the total number of cracks, about 5000 steps, and the crack accumulation curve is nearly vertical in a short time. When the crack tip produces concentrated stress, some force chains are broken and new cracks appear. The number of cracks generated in the latter half is more than the former half, and the number of cracks generated in the II interface is more than I interface, as shown in Figure 14c.

In the debonding slip stage (C–E), the crack accumulation curve gradually depresses and the crack development rate slows down. When the loading step reaches 7500 steps, the crack accumulation curve is almost horizontal. This indicated that II interface generates many cracks, and the friction force cannot resist the impact load. The rockbolt-anchor agent moves along the impact load direction and fewer new cracks generate, as shown in Figure 14d. After loading 12,500 steps, a large displacement occurs in the rockbolt-anchoring agent, resulting in debonding failure; total cracks number is 793, as shown in Figure 14e.

The results of the rockbolt debonding simulation test are similar to those of the laboratory impact test, but the displacement of the rockbolt and anchorage agent is small. The reasons are as follows: the bullet accelerates through electromagnetic force, and the bullet experiences the process of acceleration, uniform speed and deceleration in the laboratory test. The bullet rate is constant in the simulation test, and the difference in bullet rate between the laboratory test and the simulation test results in the different impact energy of the rockbolt. In the laboratory test, a silica gel tube is used to reserve holes, and the interface between the anchoring agent and the surrounding rock is relatively smooth. The particles size of the simulated test is different, and the interface particles of the anchoring agent and surrounding rock interlock. The friction force is relatively large, and the particles are not prone to dislocation.

Figure 14 shows the final failure mode diagram. The rockbolt and anchoring agent at the left end are rushed out and cracks occur at I, II interface. The cracks generated in the latter part of II interface are denser than the former part, and denser than I interface. Therefore, the final debonding failure mode of the anchoring structure is dual-interface mixed mode, which is similar to the results of the laboratory test.

In conclusion, the PFC simulation test with calibrated micro-mechanical parameters could not only better reflect debonding failure characteristics and the process of a transparent anchor structure under impact load, but also reflect the distribution law of axial force and the shear stress of its interface.

5. Conclusions

A kind of transparent similar rock was developed using a special manufacturing process and the debonding process of an anchor structure was carried out based on a developed rockbolt impact load test system and PFC simulation test, which is beneficial for revealing the debonding failure mechanism of an anchor structure under impact load. The main conclusions are as follows:

(1)

Through the impact load test of the transparent anchoring structure, the axial force of rockbolt shows a downward trend, and the maximum and minimum values occur at the load end and the other end, respectively. Shear stress at two interfaces first increases and then decreases; the shear stress at I interface is more than shear stress at II interface.

(2)

There are obvious serrated slip traces of the anchorage agent in the latter half of transparent sample. There is no large slip surface in the former half of I interface, and only local positions have scattered scratches. In the former half of I interface, debonding failure occurs and in the latter half of II interface, debonding failure also occurs. The final debonding failure mode is dual-interface mixed.

(3)

Due to the different measuring methods, shear stress distribution of the interface is slightly different between the laboratory test and the PFC test, but the final failure mode is similar. According to the failure process of the anchoring structure based on the laboratory and simulation test, the crack evolution process of the anchoring structure is divided into three phases: crack initiation, crack coalescence, and debonding slip. The axial force curve of the rockbolt, shear stress distribution of two interfaces and failure mode of the numerical simulation are in good agreement with the laboratory test results.