1 Introduction

Rock bolting system has been widely used for ground control in both mining and general engineering construction. A rock bolting system generally consists of reinforcement (i.e., rebars) embedded into the rock mass and surface support (i.e., plate, strap, and mesh) that is in contact with the rock mass surface. The effectiveness and capacity of the rock bolting system depend on the performance of both the reinforcement and the surface support as well as the compatibility of each of the elements of the system. In unfavorable ground conditions, the damage of a rock bolting system normally initiates from the weakest link of the system (Barrett and Player 2002; Charette and Bennett 2017). It is therefore critical to guarantee a good connection between each element of the system, especially the surface support which is normally the weakest part of the system. The surface support is used to constrain broken rock masses between rock bolts and distribute its load between the rock bolts resulting from the deformation and failure of the rock mass (Hadjigeorgiou and Potvin 2011). If the surface support fails prematurely, the additional load capacity of the rock bolt could be a waste of resources (Kaiser et al. 1996).

Understanding the mechanical behavior and failure mechanisms of each component (reinforcement and surface support) is of great importance to the success of ground control. Tremendous efforts have been devoted to understanding the performance and failure mechanisms of rock bolts under varying loading conditions by means of theoretical analysis (Stimpson 1989; Guan et al. 2007; Oreste and Cravero 2008), laboratory testing (Farmer 1975; Kaiser et al. 1992; Grasselli 2005; Karanam and Dasyapu 2005; Jalalifar et al. 2006; Kang and Wang 2007; Kang et al. 2013), in-situ testing and numerical modeling (Marenče and Swoboda 1995; Chen et al. 2009; Deb and Das 2011, 2014). The performance of surface support, however, is not well understood. Limited attention has been devoted to the understanding of the performance of meshes as a surface support component in the rock bolting system. This is partly because that surface support is simply a passive retaining support for the dead-weight of small broken rock mass between rock bolts (Ortlepp 1983). Due to the lack of knowledge of mechanical behavior of surface support, the design and selection of surface support in a rock bolting system mainly rely on field experience or empirical design charts associated with a rock mass classification system (Potvin and Hadjigeorgiou 2016). Some testing rigs have been developed to evaluate the mechanical performance of meshes under static or quasi-static loading in South Africa (Ortlepp 1983), Canada (Pakalnis and Ames 1983; Tannant 1995, 2004), Australia (Thompson et al. 1999; Morton et al. 2007; Shan et al. 2014; Whiting 2017), India (Watson et al. 2017), and most recently United State (Dolinar 2006; Batchler et al. 2018). A comprehensive review of these testing rigs and experimental results can be found in (Baek 2018), are thus, is not described herein. These studies have enhanced our understanding of the behavior of meshes. These testing rigs have provided useful tools to quantify the working and ultimate load and displacement capacities of surface support (mainly welded wire mesh). Since there is no standardized way to construct and conduct such a test, designing the test rigs and testing method is kind of arbitrary. There are many differences among these test rigs in terms of loading strategy, loading rate, sample size, boundary condition etc. Most of these tests were performed on welded wire meshes and used peripheral boundary conditions which do not necessarily reflect actual constraint conditions in underground mines where the meshes are constrained by rock bolts and other surface support components such as steel joints and straps (Morton et al. 2007).

In this study, a sophisticated testing facility is developed and used to test the mechanical performance of surface support under quasi-static loading. Special concern is given to the boundary conditions of tested meshes to reflect realistic constraint conditions. Three different types of mesh, welded wire mesh, chain link mesh, and geogrid mesh are tested in conjunction with three different surface components including rock bolt plate, rebar-welded strap, and W-strap.

2 Development of Test Facility and Test Configuration

A sophisticated quasi-static testing facility was exclusively designed and built by the China Coal Research Institute for testing surface supports including straps and meshes, as shown in Fig. 1. The stiff load-bearing frame rests on the two bearing beams which rest on the support frame. Tested straps and meshes are mounted on the sample frame and loaded by the hydraulic loading ram on top. The ram can move along the load-bearing frame to apply loading to different positions of tested straps and meshes. Three sets of monitoring systems, namely displacement sensor, pressure sensor and boundary tension sensor, are arranged in the testing facility to monitor the displacement, load and lateral tension of the tested straps and meshes under vertical load, respectively. A console is used to control the experiments, and the voltage signal collected by the sensors is transformed into a digital signal for outputting the experimental results. The loading ram has a maximum load capacity of 300 kN and is digital servo-controlled. The stiffness of the load-bearing frame and the bearing beam is 5 MN/mm that is sufficiently stiff to eliminate the influence of energy stored in the loading frame on the tests.

Fig. 1
figure 1

Testing facility used for surface support test

Full size image

The mesh sample is restrained within the stiff sample frame that rests on the support frame, see Fig. 2. The mesh sample is bolted to the bottom part of the sample frame by 22-mm-diameter rock bolts. The sample frame allows square pattern bolting configurations with varying spacing from 0.8 to 1.2 m with an interval of 0.1 m, representing common installation patterns in underground coal mines in China. In this study, a spacing of 1.2 m is used for all the tests.

Fig. 2
figure 2

Sample frame and constraint boundary used for mesh panel test

Full size image

Considering that a full-scale of mesh is not tested, the boundary condition of the mesh should be carefully designed to represent a realistic stress condition subjected by a mesh. In real-world practice, meshes are generally overlapped along the axial direction of a roadway with an overlap width of around 0.1–0.2 m. A row of rock bolts is installed at the overlap to constrain the meshes. The overlapped meshes are clamped together using steel wires. To represent this condition in the test, the tested meshes are overlapped along the four sides of the mesh with an overlap width of 0.1 m. The other sides of the mesh are linked to the sample frame using shackles as shown in Fig. 3, representing the far-field boundary condition of the overlapped mesh on the other side. The shackles are tightened using the tightening handles to improve clamping and ensure that all the wires are loaded.

Fig. 3
figure 3

Fixed boundary condition and mesh overlap setup used in the mesh test

Full size image

After the mesh has been installed and pre-tightened, it was then loaded perpendicularly using the hydraulic loading ram connected to the load-bearing frame.

The load was applied to the mesh by pushing a circular flat platen (diameter 30 cm) downward through the center of the mesh. To eliminate the rigid steel-to-steel loading effect, a circular rubber with a diameter of 35 cm and a thickness of 4.5 cm was placed between the dome and the mesh. The load was applied in displacement control mode, with a loading rate of 10 mm per minute. Even though this loading rate was much higher than the realistic loading rate suffered by meshes in field, it was still believed to be significant slow so that the structure deforms very slowly and therefore the inertia force can be ignored. It was thus regarded to be ‘quasi-static loading’ to distinguish it from another commonly used testing method called drop-weight which is a typical ‘dynamic loading’. It was hypothesized that the dome would load the mesh in a manner similar to a failing roof in coal measures rocks (Shan et al. 2014). The applied load was measured using a 300 kN internal load cell with an accuracy of ± 0.05%F.S. The displacement of the mesh under the load was monitored using the Linear Variable Differential Transformer (LVDT) with an accuracy of ± 0.05 cm. The loading and recording of the load and displacement were conducted using a computer.

Three different types of meshes were tested, welded wire mesh, diamond mesh (i.e., chain-link mesh), and geogrid mesh. The diameter of the wire in the welded wire mesh, the chain-link mesh, and the geogrid mesh is 6 mm, 4 mm, and 4 mm, respectively. The welded wire mesh has a mesh wire spacing of 100 × 100 mm and the geogrid mesh has a mesh wire spacing of 40 × 40 mm. The diamond mesh has a diamond side length of 50 mm. For each mesh type, the mesh is installed in conjunction with three different constraint patterns, bearing plate alone, rebar-welded strap + bearing plate, and W-strap + bearing plate, Fig. 4. The bearing plates were domed shape, grade 4, 150 mm by 150 mm in size, with a thickness of 10 mm. The rebar-welded strap is made by welding two parallel 12-mm diameter steel using 1.2-m long rebars with a spacing of 0.08 m. The W-strap is 1.2 m long, 300 mm wide and 4 mm thick.

Fig. 4
figure 4

Different constraint patterns installed in conjunction with rock bolts and meshes in the tests. a Bearing plate alone, b Rebar-welded strap + bearing plate, and c W-strap + bearing plate

Full size image

3 Test Results

3.1 Load–Displacement Response

Figure 5 shows the load–displacement curves of the meshes installed in conjunction with different constraint patterns. During the initial loading, the mesh showed a slow increase in load with displacement. This is because the mesh laid sub-horizontally with respect to the load direction and could not resist the load. This phenomenon is especially pronounced for the geogrid mesh and the diamond mesh. As the load increased, the mesh became more vertical and thus resisted the load more effectively, leading to a substantial increase in the stiffness of the mesh.

Fig. 5
figure 5

Load–displacement response of meshes tested with different constraint patterns. The dots indicate the peak load at which a substantial load drop occurs due to the wire breakage. a Welded wire mesh, b Geogrid mesh, and c Diamond mesh

Full size image

The load–displacement response of the meshes in the latter loading stage can be characterized by a “saw-toothed” pattern in which a series of sudden drops were followed by gradual increases in the load with displacement. The smaller load drops in the load–displacement response are generally associated with the slippage of the mesh wires underneath the surface reinforcement (e.g., bearing plate, steel strap, and W-strap) or the loading platen. This phenomenon is particularly pronounced for the diamond mesh where numerous small load drops occurred with displacement. The slippage of mesh wire ultimately produced a higher peak displacement and overall softer loading response. The substantial load drops are essentially associated with failure of single or multiple wires. (Tannant 1995) defined the "peak load" as the first major peak in the load prior to "a significant drop in load invariably caused by failure of a mesh wire”. This definition is adopted in this study. With the peak-load point, the entire load–displacement response of the meshes can be divided into two sages, the working stage and the failing stage (Fig. 6). The working stage refers the load–displacement response prior to the peak load, and the failing stage refers to the load–displacement response after the peak load. The stiffness of the mesh \(K\) is calculated as the slope of a line from the peak load to a point at 25% of peak load using the following equation (Dolinar 2006):

$$ K = \frac{{L_{p} - L_{25} }}{{D_{p} - D_{25} }}, $$
(1)
Fig. 6
figure 6

Load–displacement curve for a test on a welded mesh showing key parameters used for evaluating the mesh performance

Full size image

where \(L_{p}\) and \(L_{25}\) are the peak load and the load at 25% peak load, respectively, unit kN; \(D_{p}\) and \(D_{25}\) are the displacement at peak load and at 25% peak load, respectively, unit mm.

Table 1 summarizes the peak loads, displacement at peak load, displacement at final failure, and failure patterns along with the calculated stiffness for each test.

Table 1 Summary of the test result of different meshes in conjunction with different constraint patterns
Full size table

It is worth noting that the welded wire mesh can still withstand large deformation in the failing stage, even though the resistance fluctuated remarkably. In contrast, the diamond mesh and the geogrid mesh lose its capacity quickly in the failing stage. This is attributed to the different structures of the three meshes and the associated load transfer mechanics. For the welded wire mesh, the welded joints separate the continuous load-bearing structure of single wires and divide the mesh into uniformly distributed load-bearing zones. The rapture of a wire in a load-bearing zone has insignificant influence on the bearing capacity of other load-bearing zones. The mesh thus, can still withstand high load and deformation. For the diamond mesh and the geogrid mesh, the load-bearing structure of a single wire is continuous, the rupture of the wire will result in a substantial drop of the load-bearing capacity of the entire mesh.

3.2 Mesh Failure Pattern

Figure 7 presents the failure patterns of all the tested meshes. Two failure patterns were generally observed, shear rupture at the bearing plate and tensile rupture at the “bagging” area. The failure pattern of a mesh seems to be largely dependent on the constraint pattern of the mesh. All the meshes installed with bearing plate alone failed through wire (strand) shear rupture at the bearing plate, regardless of the mesh type. All the meshes installed with rebar-welded strap + bearing plate failed through wire (strand) rupture at the “bagging” area. All the meshes installed with W-strap + bearing plate also failed through wire (strand) rupture at the “bagging” area. For the welded mesh, shear rupture generally occurred in the heat-affected zone, see Fig. 8. No failure occurred at the welded joint. The strength between the welded joint and the heat-affected zone is contradictory. To achieve a high strength of the welded joint, high welding heat must be involved. This will cause a series of reactions happened to the area around the welded joint including embrittlement and softening, resulting in a significant reduction in the tensile strength and shear strength of the mesh. On the other hand, without high welding heat, the welded joint will be incompetent due to insufficient welding and fail in shear or tension, leading to premature failure of the mesh.

Fig. 7
figure 7

Failure pattern of the three different types meshes installed with three different constrain patterns

Full size image
Fig. 8
figure 8

Shear rupture generally occurred in the heat-affected zone (HAZ) of the welded wire mesh

Full size image

3.3 Effect of Mesh Constraint Pattern

The mesh constraint pattern appears to have a significant influence on the performance in both the working stage and failure stage. For each type of mesh, the mesh installed in conjunction with W-strap + bearing plate exhibited the highest peak load, followed by the mesh installed with rebar-welded strap + bearing plate. The mesh installed with bearing plate alone showed the lowest peak load. The mesh with W-strap + bearing plate also exhibited the greatest stiffness, followed by the mesh with rebar-welded strap + bearing plate and then bearing plate alone. This finding is consistent with the result of (Dolinar 2006) who tested the performance of welded wire mesh installed with different size plates and found that larger bearing plates increased both the peak load and stiffness of the mesh. When the mesh is installed with rebar-welded strap or W-strap, the load applied on the mesh center is distributed more uniformly to more wires than the bearing plate alone due to the larger coverage across the wires. The strap can provide better control over the slippage.

4 Discussion

For a clear comparison of the performance between different types of meshes, the load–displacement curves in Fig. 6 are replotted in groups of constraint patterns, see Fig. 9. Figure 10 shows the column plots of the test results listed in Table 1 to present a direct comparison of the mechanical properties of the tested meshes with different constraint patterns. It is very clear that the welded wire mesh is substantially stiffer than the geogrid mesh and the diamond mesh, regardless of the constraint pattern. The geogrid mesh and the diamond mesh do not show much difference in stiffness when rebar-welded strap or W-strap is installed with bearing plate. Under the condition of bearing plate alone, the geogrid mesh is more flexible than the diamond mesh.

Fig. 9
figure 9

Comparison of the performance of different types of meshes in conjunction with different constraint patterns. The dots indicate the peak load at which a substantial load drop occurs due to the wire breakage. a Bearing plate, b Rebar-welded strap + bearing plate, and c W-strap + bearing plate

Full size image
Fig. 10
figure 10

Comparison of the mechanical properties of different types of meshes in conjunction with different constraint patterns. a Peak load, b Stiffness, c Deformation at peak load, and d Deformation at failure

Full size image

In the case of bearing plate alone, the diamond mesh can withstand higher load and substantial larger deformation than the welded wire mesh. This result is consistent with (Ortlepp 1983; Tannant 1995; Morton et al. 2007) who also used bearing plate alone to hold tested meshes. When rebar-welded strap or W-strap is considered, the welded wire mesh can bear higher load than the diamond mesh. To the author’s best knowledge, this finding has not been reported before. As discussed previously, in the case of bearing plate alone, all the three types of meshes failed through shear rupture at the bearing plate as a result of the plate cutting through the wires or as a result of the wires cutting each other, which is a typical failure pattern in underground coal mines, as shown in Fig. 11. For the welded wire mesh, much fewer wires were constrained by the bearing plate than the diamond mesh. The premature failure of the single wires resulted in unpredictable loss of structural integrity of the welded wire mesh (Bhagwat and King 2014), causing a substantial drop of its load-bearing capacity. The high-tensile strength of the wires was not fully mobilized. When the meshes were installed with steel strap or W-strap, much more wires were constrained at the boundary of the welded wire mesh. The structural integrity of the welded wire mesh trended to be contained even when single wires ruptured. This finding suggests that the high bearing capacity of welded wire mesh can only be fully mobilized when rebar-welded strap or W-strap is incorporated.

Fig. 11
figure 11

Welded wire mesh failed at the bearing plate as a result of the plate cutting through the wires

Full size image

The geogrid mesh exhibited the lowest working capacity among the three types of mesh in all the cases of constraint pattern.

In conjunction with rock bolts and other surface support components, meshes play different roles in surface support, depending on the ground conditions. In stable ground conditions, the mesh retains and holds loosed rocks in place. The mesh between the rock bolts must be given the necessary resistance to the dead-weight of the loosed rocks. Stiffness is not important in this case. Flexible mesh such as the diamond mesh with small holes is sufficient and appropriate. In the case of instability problems, stress-induced failure (e.g., splitting, spalling, or buckling) occurred in the immediate vicinity of an excavation, the surface support combined with rock bolts holds the fractured rock mass in position, forming a support arch by mobilizing the strength of fractured rock reinforced by rock bolts. This support arch then provides confinement and enhances the rock mass strength outside the fractured zone. In this case, it is crucial that the surface support is stiff so that it can provide immediate resistance after installation and sufficient resistance will form with little outward deformation of the damaged rock mass. Heavy gauge welded wire mesh with higher stiffness and load-bearing capacities is therefore suitable for this condition. For heavy gauge welded wire mesh, the weakest link is the welded joints, and the loading capacity is limited by the relatively low breaking strength of the welded joints. Another weak link is the heat-affected zone which is vulnerable to shear rupture, see Fig. 8. Shear strength of wires, as a general rule of thumb, is approximately 50% of the tensile strength (Bhagwat and King 2014). The heat-affected zone has a relatively lower shear strength. The bearing capacity of the welded wire mesh can be greatly improved if welds are not incorporated. For example, (Bhagwat and King 2014) proposed a FlexKnot mesh, where instead of using welds to join wires, the high-tensile strength wires are joined by a unique steel wire knot. The load-bearing capacity (i.e., peak load) of the FlexKnot mesh is 2.6–4.2 times that of the 10-gauge and 8-gauge welded wire mesh. Proper installation with sufficient mesh overlap is critical to the performance of welded wire mesh.

In extremely ground conditions (i.e., highly jointed rock mass), the surface support is used to confine the smaller broken rock masses between reinforcement elements (e.g., rock bolts and cables) and distribute its load between the reinforcement elements (Hadjigeorgiou and Potvin 2011). The surface support provides an important role in the overall ground support system. There existed a small tension zone between rockbolts at the surface of an underground opening. The blocks in the tension zone are loose and likely to fall-out. The fall-out of this small amount of loose rock blocks in the tension zone tends to start raveling of the whole mass due to the movement of certain key blocks. The installation of meshes on the free surface can largely eliminate fall-out of small rock blocks, and thus, enhance the bearing capacity of highly fractured rock masses.

In the burst-prone ground, it is critical for the surface support to transfer dynamic load effectively to yielding bolts (Cai and Kaiser 2018). Welded wire mesh is more suitable than geogrid mesh and diamond mesh not only because it is stronger, but also because it is stiffer. In heavy rock burst grounds, normal welded wire mesh may not provide sufficient load and deformation capacity, special meshes have been developed, such as the high energy absorption (HEA) mesh (Potvin 2009; Potvin and Heal 2010). Welded wire mesh can also be used with shotcrete to provide higher load and energy absorption capacities in burst-prone grounds (Kaiser and Tannant 2001; Maclsaac and Swan 2001). A proper design of surface support using meshes relies on the knowledge of the energy absorption capacity of meshes which can only be determined by performing drop tests under dynamic conditions. This is beyond the scope of this study but it is worth noting that reported testing facilities for such tests include the dynamic testing facility of the Western Australian School of Mines (Player et al. 2004) and the Geobrugg facility in Switzerland (Bucher et al. 2013). A more sophisticated testing facility for dynamic test on meshes is under development in the China Coal Research Institute.

5 Conclusion

Using the newly developed testing facility, three different types of mesh, welded wire mesh, chain-link mesh, and geogrid mesh were tested in conjunction with different surface components, including rock bolt plate, rebar-welded strap and W-strap. The following conclusions can be drawn.

  1. (1)

    The welded wire mesh is substantially stiffer than the geogrid mesh and the diamond mesh, regardless of the constraint pattern. When rebar-welded strap or W-strap is incorporated, the welded wire mesh exhibits the greatest load-bearing capacity, followed by the diamond mesh and then the geogrid mesh.

  2. (2)

    The surface constraint pattern has a significant effect on the performance of a mesh. When the mesh is installed with rebar-welded strap or W-strap, much more wires are constrained at the boundary of the mesh. The structural integrity of the mesh can be contained even when single wires rupture. This is particularly true for the welded wire mesh. The high bearing capacity of welded wire mesh can only be fully mobilized when rebar-welded strap or W-strap is incorporated.

  3. (3)

    The welded wire mesh can still withstand large deformation after peak load. In contrast, the diamond mesh and the geogrid mesh lose its capacity quickly after peak load.