In this study, during the continuous retreat mining of the upper section working face, the upper section gradually transforms into a goaf (mined-out area), with the immediate roof above it collapsing and separating from the main roof (in Figure 1). The fixed support edge of the main roof in the upper section working face is located on the solid coal side. Fracturing of the coal mass primarily occurs in the lateral coal mass, with the fracture line above the coal seam serving as the central axis. This results in rotation or bending subsidence, forming a hinged structure comprising rock blocks A, B, and C. The formation of rock block B arises from the collapse of the main roof at the end of the working face (Yang, 2019).

As shown in Figure 1, the semi-arch articulated structure comprises rock block A, arc-shaped triangular block B, and rock block C (Yang et al., 2021). Block A represents the main roof rock beam above the working face of this section; block B corresponds to the arc-shaped triangular block on the goaf side of the upper section working face; and block C is the fractured block in the upper section goaf. The stability of this semi-arch articulated structure is directly influenced by the fracturing of the main roof rock beam and the collapse of the immediate roof. Under the combined effects of advanced and lateral support pressure, within the extraction influence range of the lower section working face, rock block B rotates and subsides, altering its position. This movement leads to compression and subsidence of the gangue beneath rock block C, while also causing compression and subsidence of the coal mass and roof associated with rock block A. Rock block B plays a pivotal role in maintaining the stability of large-scale structures within the overlying rock mass during gob-side entry driving.

The formation and stability of large structures in gob-side entry driving are influenced by the thickness and mechanical properties of the immediate roof and main roof, as well as the extent of coal seam extraction. The key block B, specifically the main roof beam in the overlying rock structure, is characterized by four fundamental parameters: (1) the fracture position S ₀ of the key block in the coal mass; (2) fracture length L ₁ (C _i) of the main roof along the working face advancement direction; (3) the lateral fracture span L ₂ of the main roof beam; and (4) thickness h of the old roof beam (i.e., thickness of the main roof beam).

The fracture position S ₀ of the key block in the coal mass is calculated as Equation 1: S 0 = 2 S m M E C E γ E k γ H (1)

Where S _m is the advance distance of the working face (m); M _E is the maximum thickness of the fractured main roof beam (m); and C _E is the periodic weighting step distance of the fractured main roof beam (m).

The lateral fracture span L ₁ of the main roof beam represents the periodic weighting step distance of the main roof.

The lateral fracture span of the main roof beam L ₂ is calculated as Equation 2: L 2 = 2 L 1 17 10 L 1 L 2 2 + 102 − 10 L 1 L 2 (2)

Where L ₀ is the length of the working face (m).

Grouting reinforcement technology for roadway excavation along the goaf is an important method to improve coal pillar stability and bearing capacity. Its core mechanisms are reflected in numerous aspects (Zhang P. et al., 2018; Zhang H. et al., 2018; Guo et al., 2017), such as enhancing the self-supporting capacity of surrounding rock, creating a network skeleton effect through grouting consolidation, reducing the loosening zone, preventing weathering and water seepage, and improving the stress state of anchor cables.

Grouting methods are selected according to geological conditions and the penetration capacity of the slurry, and can generally be divided into filling grouting, permeation grouting, and splitting (fracturing) grouting. For gob-side roadways supported by small coal pillars, where a large number of mining-induced fractures are developed, filling and permeation grouting are the main techniques adopted. According to flow–sedimentation theory, when grout enters a fissure, its velocity and pressure decrease rapidly with increasing distance from the injection point. Once the flow velocity falls below a critical value, solid particles begin to settle under gravity, which reduces the effective flow cross-section. This further alters the local pressure gradient and velocity distribution, while the remaining liquid phase continues to migrate along the fracture until the voids are gradually filled.

Figure 2 illustrates the simulated diffusion and spatial distribution of grout under the same grouting pressure (3 MPa) but with two different water–cement ratios (0.5:1 and 1:1). The simulations show that grout migration is strongly controlled by the pre-existing fracture network of the rock mass. In both cases, the slurry preferentially propagates along interconnected joints and cavities, whereas penetration into intact rock blocks is negligible. When the water–cement ratio is increased to 1:1, the diffusion radius and total injected volume both increase significantly compared with the 0.5:1 case, indicating that higher water content improves fluidity and thus promotes percolation into finer or deeper fractures.

The numerical analyses were carried out using UDEC (Universal Distinct Element Code), which is well suited to modeling fractured rock masses as discrete block systems. In the model, the strata were represented by deformable blocks separated by joints, and hydro–mechanical coupling was introduced to simulate grout flow within these discontinuities. Grouting was modeled by prescribing a constant injection pressure of 3 MPa at the borehole and tracking the transient evolution of fluid pressure and flow velocity throughout the joint network. Joint aperture, spacing, and connectivity were calibrated using field fracture surveys and laboratory permeability tests to ensure that the model reproduced the actual fracture characteristics of the coal–rock mass. The grout was treated as a Bingham-type fluid, with viscosity and yield stress parameters determined from rheological tests at different water–cement ratios. Diffusion was therefore simulated as a coupled process in which grout penetration simultaneously improved joint stiffness due to consolidation.

The simulation results indicate that, when rock splitting or hydraulic fracturing is not considered, grout mainly diffuses radially around the injection point and propagates outward along the most continuous fracture paths. Flow velocity is highest near the borehole and in shallow, highly fractured zones, but it decays rapidly with increasing distance because of pressure attenuation and the gradual narrowing of fractures. This behavior is consistent with theoretical predictions for flow–sedimentation processes in fractured–porous media. The comparison of the two water–cement ratios further demonstrate the dominant role of grout rheology: the 1:1 mixture, having lower viscosity, achieves a larger diffusion range and better coverage, but its final strength is relatively lower; by contrast, the 0.5:1 mixture produces a denser filling zone near the borehole while exhibiting a smaller penetration radius. These observations provide a useful basis for optimizing grouting parameters in field applications, especially in extra-thick coal seams where fracture development is complex and strongly anisotropic.

Figure 3 shows the attenuation curves of grouting pressure in both the radial and tangential directions of the roadway for the 1:1 mixture. In the radial direction, the grouting pressure drops sharply near the rock surface and then more gradually at greater depths, and the attenuation can be described by a quadratic relationship with diffusion distance. In the tangential direction, the grouting pressure exhibits a nearly symmetrical attenuation pattern about the grouting hole, reflecting the lateral diffusion characteristics of the slurry along the roadway.

Grouting materials play a pivotal role in achieving an effective grouting outcome. A nano-modified two-component grouting material has been developed. This material is primarily composed of ultra-fine inorganic mineral powder with a D90 particle size under 10 μm, which is one-sixth that of #42.5 cement. It exhibits strong wettability, high injectability, rapid setting, early strength development, and strong adhesion, making it suitable for both low- and high-pressure grouting processes. The material achieves a 100% consolidation rate and slight expansibility, effectively preventing post-solidification shrinkage. The weight ratio of the two components (material A and material B) is 1:1. For well-developed fractures, the recommended water-cement ratio ranges from 0.5:1 to 0.8:1, while for less-developed fractures, it ranges from 0.8:1 to 1:1.

Figure 4 illustrates the bleeding rates of material A and B grouts under varying water-cement ratios. As shown, both grouts demonstrate an initial increase in bleeding rates with time, stabilizing later. At a water-cement ratio of 0.6:1, neither grout exhibits bleeding, and the bleeding rate remains unaffected by time. For the material A grout, the bleeding rate reaches a maximum of 1.52% at a water-cement ratio of 1.5 after 120 min of settling. When the water-cement ratio is ≤ 1.0, the bleeding rates are below 1%. Similarly, the material B grout exhibits no bleeding at water-cement ratios of 0.6:1 and 0.8:1, while at 1.5:1, the bleeding rate is 1.6% after 120 min. At ratios <1.5, bleeding rates remain below 2%.

Figure 5 highlights the compressive strength characteristics of the grouting material at different water-cement ratios. A lower water-cement ratio correlates with higher compressive strength of the solidified body, while a longer curing period results in further strength gains. For example, at a ratio of 0.6:1, the 2-h compressive strength reaches 20.1 MPa, whereas at 1.5, it drops to 2.5 MPa. From a strength perspective, field grouting is recommended with water-cement ratios between 0.6:1 and 1.3:1 to meet engineering requirements.

Table 1 demonstrates a comparison of mechanical properties between two grouting materials. Material I (nano-modified two-component grouting material) is characterized by a micro-fine particle size (D90 = 7.8 μm), whereas Material II (low-viscosity two-component polyimine resin) is liquid without a defined particle size. The micro-scale of Material I is consequential: it promotes diffusion into microcrack networks and coal cleats that are inaccessible to coarser slurries or neat resins, improving grout curtain continuity and sealing effectiveness in low-permeability media.

In terms of kinetics, Material I exhibits a fast yet controllable setting profile (initial set 2–5 min; final set 40 min). The broader initial-set window provides practical adjustability for field operations—reducing the risk of premature line blockage while still limiting washout—and the longer final-set time allows deeper migration before immobilization. By contrast, Material II reaches initial and final set at ∼2 min and 30 min, respectively, offering less operational latitude during injection.

Mechanical testing shows that Material I develops adequate early strength (2 h compressive strength 9.5 MPa) and reaches 20.3 MPa at 7 d, with bond strengths of 1.5 MPa (coal) and 2.13 MPa (rock). Although these values are lower than those of Material II (21/36/42 MPa at 2 h/24 h/7 d and 3.2/3.5 MPa adhesion), the moderate stiffness of Material I can be advantageous in deformable coal–rock masses: it lowers the likelihood of stress concentration and secondary cracking at the grout–matrix interface, supporting energy dissipation rather than brittle localization.

Overall, the data indicate that Material I is optimized for scenarios where penetrability, controllable setting, and compatibility with weak or highly fractured matrices are prioritized over peak strength—e.g., sealing fine fissures, reducing permeability, and stabilizing coal measures without imposing excessive rigidity.

Working face #8121 is located in the western sector of the first mining district’s west wing. Its northern end connects to three main roadways in the 1070 west wing, and its southern end terminates at the F1532 fault boundary. The western side adjoins the goaf of working face #8121, whereas the eastern side is solid coal. Roadway #5121 serves as the return-air and transport roadway for working face #8121. It was excavated along the floor of coal seams No. 3-5, with a design length of 1660.647 m and a coal-pillar width of 10 m. The layout of working face #8121 is shown in Figure 6.

The reinforcement employed Φ22.4 mm hollow, high-pressure grouting cables fabricated from 1 × 9 steel strands. Each cable is 4,300 mm long; the drill bit diameter is 36 mm; and the ultimate tensile capacity is ≥ 450 kN. Two resin cartridges were used (MSK2860 and MSZ2860).

A high-strength bearing plate (300 × 300 × 16 mm) with a spherical washer was installed at each collar. Cables were arranged in a rib pattern of “2-0-2”, with 1,800 mm row spacing and 950 mm column spacing. All cables were installed vertically into the ribs and tensioned to 250 kN. The cross-sectional arrangement is shown in Figure 7.

After the baseline support (anchor cable-mesh-bolt plus shotcrete) was completed, high-pressure grouted cables were installed between steel bands #333 and #555 to reinforce the surrounding rock. A nano-modified, two-component microfine grout (D90 < 10 μm) was adopted. Key properties were: initial set 3–20 min, final set 5–30 min, 28-day compressive strength ≥30 MPa, and coal–grout bond strength ≥3.5 MPa. Component A and B powders were mixed at a 1:1 volumetric ratio with a water-to-cement ratio of 0.6:1–0.8:1, adjusted to site conditions.

Mixing and delivery used two-component synchronous electric pumps with high-speed mixers. Injection was performed through the hollow cables at a terminal pressure typically limited to 5–10 MPa and, where necessary to ensure grout take, briefly raised to 10–20 MPa. The locking device, mixing head, and plate matched those of the existing Φ21.8 mm cables. After installation, the plate, spherical washer, and lock were mounted and the cables tensioned to seal the collar and enable subsequent high-pressure injection via the grouting connector (Figure 8).

The grouting construction procedure comprised equipment setup, slurry preparation and ratio calibration, injection, shutdown and cleaning, and operational control. First, the pump and mixing tanks were installed on a level pad, and the suction and discharge lines were laid as straight as possible, avoiding twists, crossings and sharp bends. System priming and functional checks were conducted by circulating clean water for more than 1 min with the grouting valve open, followed by briefly running the pump dry to expel residual water. Slurry preparation involved adding the prescribed amounts of water to the A- and B-tanks, starting the mixers, and then introducing equal masses of A- and B-component powders under continuous agitation. After approximately 10 min of mixing, the pump suction was immersed in the blended slurries; with the relief valve closed and the grouting valve open, equal discharge at the A and B outlets (A:B = 1:1 by mass) was verified, after which the hoses were returned to the tanks for recirculation. For injection, the discharge hose was quickly connected to the grouting pipe, the grouting valve was opened with the relief valve closed, and pumping commenced. Upon completion of each hole, the relief valve was opened to depressurize the system and the grouting valve was closed before moving to the next hole. For batch changes or stoppages, the system was first depressurized (relief valve open, grouting valve closed), the suction line was transferred to clean water, and the pump was operated with the discharge valve open until the outflow ran clear; subsequently, the discharge valve was closed, the relief valve reopened, and circulation continued until clear water was observed at the relief outlet, while the mixing tanks were cleaned concurrently. During operation, leakage points were sealed immediately, using manual sealing for minor leaks and adjusting grout setting time in the case of large or continuous leakage to balance flowability and sealing effectiveness. A functional pressure gauge was mandatory, and any faulty gauge was replaced without delay. The injection rate was regulated within 10–25 L min^-1 according to the pressure response, and a bottom-to-top grouting sequence with a pre-planned hole order was strictly enforced, with shift-by-shift verification to prevent omissions.

During field implementation, 176 of the 240 designed hollow grouting cables were successfully injected; the remaining cables could not be grouted owing to localized cable-head failures and excessively short exposed lengths after shotcreting. The total grout consumption was 7.24 t, corresponding to an effective uptake of approximately 25–50 kg per cable. The relatively uniform grouting volumes along the reinforcement zone, together with the absence of abnormally high local grout takes, indicate that no large open fractures or cavities were encountered within the treated coal mass. Throughout construction, the grouting pressure was generally maintained below 10 MPa and stabilized at approximately 5 MPa under normal operating conditions. The on-site grouting process for the hollow cables is illustrated in Figure 9.