Introduction

Coal mining has raised significant environmental concerns. China’s raw coal production reached 4.71 billion tons of total energy consumption in 20231,2. The effects of coal mining on rock strata stability, surface subsidence, and solid waste accumulations have become increasingly significant3,4,5. This has resulted in environmental issues like damage to buildings and farmland, water resource depletion, and land desertification. For instance, the area affected by coal mining subsidence in China has reached 20,000 square kilometers and increases by 700 square kilometers annually6,7,8,9. Additionally, coal mining produces 500–800 million tons of gangue annually, while coal-fired plants generate 600 million tons of fly ash per year10,11,12. These solid wastes occupy land and severely pollute the environment. Therefore, it is crucial to optimize the mining method for sustainable mining in terms of avoiding subsurface subsidence and utilizing mining wastes.

Underground mining methods have become increasingly diverse, encompassing unsupported methods (e.g., room and pillar), supported methods (e.g., cut and fill), and caving methods (e.g., block caving). The long-wall caving mining method, which causes rock strata collapse and surface subsidence, remains the primary coal mining technique. Recently, backfill mining, which refills the goaf with backfill materials, has gained popularity due to its ability to prevent subsidence and recycle mining waste. Benzaazoua et al.13 developed a desulfurization technology for cemented paste backfill, while Deng et al.14 introduced long-wall cemented backfill mining for ultra-thick seams, effectively protecting aquifers by controlling water-conducting fracture zones. Sun et al.15 analyzed coal pillar stability in long-wall roadway backfill mining, and Jia and Zhang16 studied the stability of backfill during long-wall mining, determining the minimum backfill width. Zhu et al.17 proposed a layered backfill mining method for thick seams, showing that cemented backfill using solid waste controls surface subsidence. Feng et al.18 optimized backfill areas to reduce costs by analyzing potential roof fracture zones. Although long-wall backfill mining improves control over both overlying strata and the surface, it has some limitations: it is most effective for thick, gently inclined seams; the simultaneous backfilling and mining operations can interfere with each other; and the rapid creation of large roof spaces demands high early strength for backfill materials. In addition, the backfill time under the influence of the flowability of the backfill material should also be considered19, as it has a significant impact on the backfill effectiveness.

Short-wall mining, characterized by short coal faces, is a highly adaptable and efficient method known for its high safety and resource recovery rates20,21. It is suitable for irregular block segments and complex geological areas22. However, it faces challenges due to differences in face layout, process procedures, and strata pressure compared to longwall mining. Liu et al.23 used RFPA2D to study coal pillar stability, confirming its effectiveness in controlling surface subsidence. Wang et al.24 examined roof pressure under full caving in short-wall continuous mining, demonstrating its safety benefits. Tan et al.25 analyzed the engineering stability of Wongawilli coal pillars, recommending a pillar width-to-mining height ratio between 0.16 and 0.75 for a mining height of 2–6 m. Zhang et al.26 proposed full caving at corners for coal recovery and established criteria for optimal pillar widths. Zhang et al.27 introduced short-wall block backfill mining to control water-conducting fractures and optimize pillar retention width and backfill density. Zhu et al.28 suggested backfilling without coal pillars in short-long wall mining, highlighting its economic benefits and potential for energy reservoir creation. Zhang et al.29 investigated heavy metal contamination in mine water during segmented block backfill mining, emphasizing pollution control strategies. Zhao et al.30 explored the impact of backfill parameters on overlying strata, identifying backfill cohesion as the primary influencing factor.

Literature reviews indicate that the short-wall backfill mining method has garnered attention due to its feasibility and effectiveness in land protection and coal waste utilization. However, most studies focus on specific coal mine situations, with a lack of a general theory and design protocol, hindering its safe application and promotion.

This paper presents a short-wall roadway backfill method (SRBM) and its corresponding design protocol to address safety and environmental concerns. An analytical model is developed to predict overlying strata failure in the short-wall backfill stope. The variation of basic roof tensile stress under different mining parameters is examined. Mining parameters, including backfill material mixtures, are determined for a coal mine. A three-dimensional numerical model, considering both mining and backfilling processes, is established to analyze stress and deformation evolution. Finally, the method is applied in a coal mine, with stress and deformation monitored during implementation.

Short-wall roadway backfill mining method and design protocol

Short-wall roadway backfill method

The SRBM method divides the mining area into sections and arranges working faces based on the short-wall mining production system. A complete transportation system is constructed to ensure connectivity between the belt gate, air-return gate, and open-off cut, forming a negative-pressure ventilation system.

The working faces are linked by crosscuts that connect the upper and lower headings, with several crosscuts excavated to recover coal resources. Different crosscut sequences are designed to coordinate mining and backfill operations. Upon completing mining in the first crosscut, backfilling with coal gangue and fly ash begins immediately, while the continuous miner moves to the next crosscut. This creates a “time–space” coordinated cyclic operation system. This method not only recovers coal but also addresses waste disposal, surface subsidence, damage to surface structures, and environmental pollution. The tunnel dimension parameters are shown in Table 1. The principle of SRBM technology is shown in Fig. 1 (The figure shows a short-wall working face in operation, with the coal body between the backfill body being the coal seam to be mined in the next stage).

Table 1 Tunnel parameters.
Full size table
Fig. 1
figure 1

Principles of SRBM.

Full size image

Working face layout and operation processes

The working face layout adopts the single-directional longwall mining method. Initially, the upper and lower gates (belt gate with air-return gate) and the open-off cut are excavated. The short-wall working face is divided into crosscuts, with each crosscut measuring 50–100 m in length and 5 m in width (The crosscut width is based on the cutting width of the continuous miner). These crosscuts are further divided into equally wide support lanes along the horizontal axis. During mining, a continuous miner advances from the belt gate toward the air-return gate, extracting coal in the crosscuts. A crawler-type hydraulic support frame moves with the miner, providing timely roof support. Upon reaching the maximum roof control distance, a bolting machine reinforces the face in the crosscuts. Coal is then transported by the loader to the crushing machine, after which it is conveyed away from the working face via a conveyor belt. Once all coal is extracted, backfill material is transported through the air-return gate to fill the crosscuts. The working face layout is shown in Fig. 2.

Fig. 2
figure 2

Working face layout.

Full size image

Backfill material preparation occurs on the surface, involving gangue screening, crushing, and slurry preparation and pumping. After connecting the underground crosscuts, backfill walls are quickly erected at both ends to form a sealed space, reinforced with blocking planks and lined with filter cloth to prevent slurry leakage. External support is provided by single pillars. The filling slurry is transported through the main pipeline in the air-return gate and directed through unidirectional dual pipes along the roof of the crosscut, with outlets at 1/3 and 2/3 of the lane length. The filling process occurs in three stages, with each stage allowing the backfill to set before proceeding to the next. This is illustrated in Fig. 3.

Fig. 3
figure 3

Preparation and filling process schematic.

Full size image

SRBM employs a “skip mining” method, where coal is mined by skipping every two crosscuts. After excavating the initial crosscut, the next is mined by skipping two subsequent crosscuts. As each crosscut is excavated, the previously mined crosscut is backfilled. This iterative process continues until the entire working face is mined. The mining sequence is shown in Fig. 4.

Fig. 4
figure 4

The mining sequence.

Full size image

Design procedure for mining parameters

Figure 5 illustrates the SRBM mining parameter design process. First, the mining process is defined based on the proposed method. A mechanical analytical model is used to identify key factors affecting roof strata fracturing and analyze their impact. Next, criteria for fracturing and a prediction model are developed. Geological conditions and mining parameters are input to assess whether fracturing occurs. If fracturing is detected, parameters are adjusted; if not, preliminary parameters are set. 3D numerical simulations are then conducted to evaluate stability. If unstable, parameters are revised and recalculated; if stable, the parameter design is finalized to ensure both ecological protection and safe, efficient mining.

Fig. 5
figure 5

Design procedure for mining parameters.

Full size image

Stability analysis of rock strata structure

Basic principle of rock strata control in SRBM

Traditional caving methods disrupt the equilibrium stress state of the overlying strata during coal extraction, causing stress redistribution in both horizontal and vertical directions. When the stress exceeds the rock mass’s failure strength, mining-induced fractures occur, leading to deformation and collapse of the overlying strata, which compromises the safety of the working face. In contrast, SRBM employs backfill support directly on the roof, reducing roof bending and subsidence. This creates a stable backfill-overlying strata structure, minimizing mining-induced damage and controlling strata movement and deformation. Under backfill support, the overlying strata bend and sink layer by layer from the immediate roof upwards. Over time, the deformation is transmitted to the surface, causing only slight subsidence31,32. The basic principle diagram of strata control in SRBM is shown in Fig. 6.

Fig. 6
figure 6

Mechanism of overlying strata movement under different mining methods. (a) Traditional caving method. (b) Backfill mining

Full size image

Establishment of the mechanical model

To study overburden movement and deformation during the SRBM process, this paper examines the horizontal segmentation and vertical layering of the overburden elastic foundation during mining and backfilling. Based on the Winkler foundation hypothesis33, the basic roof above the backfill is modeled as the research object. The load on the upper roof rock beam is represented by q1(n), q2(n), while the lower direct top, backfill, and solid coal are treated as an elastic foundation. A composite beam mechanical model for the elastic foundation is established. The overburden and foundation of the stope are as shown in Fig. 7. Assume that the thicknesses of the rock layers above the backfill are h1, h2,…. The total thickness of the bedrock in the ith layer above the backfill is \(H_{n} = \sum\nolimits_{n = 1}^{n} {h_{n} } \;\;\left( {n = 1,2, \ldots n} \right)\).

Fig. 7
figure 7

The overlying strata of the stope and the ground foundation structure.

Full size image

Assuming that the center of the stope is at the original rock stress point O, then the load q1(n) above the interface between the backfill and the solid coal at the side of the stope is:

$$q_{1\left( n \right)} = \frac{{\left( {1 - f} \right)q_{\left( n \right)} }}{{l_{2} }}x + fq_{\left( n \right)} + \frac{{L\left( {f - 1} \right)q_{\left( n \right)} }}{{2l_{2} }}$$
(1)

where f is the stress concentration factor; L is the length of the backfill; l2 is the range of lateral support stress influence in the solid coal; q(n) is the original rock stress of the nth layer.

$$q_{2\left( n \right)} = \gamma \left( {M - H_{n - 1} } \right)$$
(2)

where γ is the unit weight of the rock layer; M is the potential collapse height of the bedrock.

Let the deflection of the basic roof rock be w(x). Using the displacement function w(x) as the unknown variable, establish the mechanical model as shown in Fig. 8a.

Fig. 8
figure 8

Mechanical model of stope composite structure. (a) Stope mechanical model. (b) Simplified stope mechanical model

Full size image

In this model, the basic roof, the underlying rock layers, the solid coal, and the backfill can be considered as a composite structure. After simplification, the mechanical model is as shown in Fig. 8b. The elastic foundation coefficients are as follows:

$$\left\{ {\begin{array}{*{20}l} {\frac{1}{{k_{m} }} = \frac{1}{{k_{a} }} + \frac{1}{{k_{1} }}} \hfill \\ {\frac{1}{{k_{n} }} = \frac{1}{{k_{b} }} + \frac{1}{{k_{1} }}} \hfill \\ \end{array} } \right.$$
(3)

where ka, kb, and k1 represent the elastic foundation coefficients of the solid coal, the backfill, and the immediate roof, respectively.

The differential equation for the deflection curve of different segments of the basic roof is:

$$\left\{ {\begin{array}{*{20}l} {E_{0} I\frac{{d^{4} w_{(x)1} }}{{dx^{4} }} + k_{n} w_{(x)1} = q_{2\left( n \right)} ,} \hfill & {\quad 0 \le x \le \frac{L}{2}} \hfill \\ {E_{0} I\frac{{d^{4} w_{(x)2} }}{{dx^{4} }} + k_{m} w_{(x)2} = q_{1\left( n \right)} ,} \hfill & {\quad \frac{L}{2} \le x \le \frac{L}{2} + \, l_{2} } \hfill \\ \end{array} } \right.$$
(4)

where E0I is the stiffness of the beam.

By taking the characteristic coefficients \(\alpha _{1} = \sqrt[4]{{\frac{{k_{n} }}{{4E_{0} I}}}}\), \(\alpha _{2} = \sqrt[4]{{\frac{{k_{m} }}{{4E_{0} I}}}}\), the deflection of different segments of the basic roof is obtained as follows:

$$\left\{ {\begin{array}{*{20}l} {w_{(x)1} = e^{{\alpha_{1} x}} \left( {A_{1} \cos \alpha_{1} x + B_{1} \sin \alpha_{1} x} \right) + e^{{ - \alpha_{1} x}} \left( {C_{1} \cos \alpha_{1} x + D_{1} \sin \alpha_{1} x} \right) + \frac{{q_{2\left( n \right)} }}{{k_{n} }},} \hfill & {\quad 0 \le x \le \frac{L}{2}} \hfill \\ {w_{(x)2} = e^{{\alpha_{2} x}} \left( {A_{2} \cos \alpha_{2} x + B_{2} \sin \alpha_{2} x} \right) + e^{{ - \alpha_{2} x}} \left( {C_{2} \cos \alpha_{2} x + D_{2} \sin \alpha_{2} x} \right) + \frac{{q_{1\left( n \right)} }}{{k_{m} }},} \hfill & {\quad \frac{L}{2} \le x \le \frac{L}{2} + \, l_{2} } \hfill \\ \end{array} } \right.$$
(5)

The relationship between the elastic foundation coefficients and the thickness of the solid coal, rock layer, and backfill is:

$$k_{1} = \frac{{E_{1} }}{{h_{1} }},\;\;k_{a} = \frac{{E_{a} }}{{h_{b} }},\;\;k_{b} = \frac{{E_{b} }}{{h_{b} }}$$
(6)

where E1, Ea, and Eb represent the elastic moduli of the immediate roof, solid coal, and backfill, respectively. h1 and hb represent the heights of the immediate roof and the backfill (mining height of the coal seam), respectively.

The relationships between the rotation angle θ(x), bending moment M(x), shear force Q(x), and deflection w(x) at any cross-section of the rock beam are as follows:

$$\left\{ {\begin{array}{*{20}l} {\theta_{( x )} = \frac{{dw_{(x)} }}{dx}} \hfill \\ {M_{( x )} = - E_{0} I\frac{{dw^{2}_{(x)} }}{{dx^{2} }}} \hfill \\ {Q_{( x )} = - E_{0} I\frac{{dw^{3}_{(x)} }}{{dx^{3} }}} \hfill \\ \end{array} } \right.$$
(7)

By further solving, the bending moment for different segments of the basic roof is obtained as follows:

$$\left\{ {\begin{array}{*{20}l} {M_{(x)1} = - E_{0} I\left[ {\begin{array}{*{20}l} {\alpha_{1}^{2} \exp \left( {\alpha_{1} x} \right)\left( {A_{1} \cos \frac{{\pi \alpha_{1} x}}{180} + B_{1} \sin \frac{{\pi \alpha_{1} x}}{180}} \right)} \hfill \\ { + \alpha_{1}^{2} \exp \left( { - \alpha_{1} x} \right)\left( {C_{1} \cos \frac{{\pi \alpha_{1} x}}{180} + D_{1} \sin \frac{{\pi \alpha_{1} x}}{180}} \right)} \hfill \\ { + \frac{{\alpha_{1}^{2} \pi \exp \left( {\alpha_{1} x} \right)\left( {B_{1} \cos \frac{{\pi \alpha_{1} x}}{180} - A_{1} \sin \frac{{\pi \alpha_{1} x}}{180}} \right)}}{90}} \hfill \\ { - \frac{{\alpha_{1}^{2} \pi \exp \left( { - \alpha_{1} x} \right)\left( {D_{1} \cos \frac{{\pi \alpha_{1} x}}{180} - C_{1} \sin \frac{{\pi \alpha_{1} x}}{180}} \right)}}{90}} \hfill \\ { - \frac{{\alpha_{1}^{2} \pi^{2} \exp \left( {\alpha_{1} x} \right)\left( {A_{1} \cos \frac{{\pi \alpha_{1} x}}{180} + B_{1} \sin \frac{{\pi \alpha_{1} x}}{180}} \right)}}{32400}} \hfill \\ { - \frac{{\alpha_{1}^{2} \pi^{2} \exp \left( { - \alpha_{1} x} \right)\left( {C_{1} \cos \frac{{\pi \alpha_{1} x}}{180}} \right) + D_{1} \sin \frac{{\pi \alpha_{1} x}}{180}}}{32400}} \hfill \\ \end{array} } \right],} \hfill & {\quad 0 \le x \le \frac{L}{2}} \hfill \\ {M_{(x)2} = - E_{0} I\left[ {\begin{array}{*{20}l} {\alpha_{2}^{2} \exp \left( {\alpha_{2} x} \right)\left( {A_{2} \cos \frac{{\pi \alpha_{2} x}}{180} + B_{2} \sin \frac{{\pi \alpha_{2} x}}{180}} \right)} \hfill \\ { + \alpha_{2}^{2} \exp \left( { - \alpha_{2} x} \right)\left( {C_{2} \cos \frac{{\pi \alpha_{2} x}}{180} + D_{2} \sin \frac{{\pi \alpha_{2} x}}{180}} \right)} \hfill \\ { + \frac{{\alpha_{2}^{2} \pi \exp \left( {\alpha_{2} x} \right)\left( {B_{2} \cos \frac{{\pi \alpha_{2} x}}{180} - A_{2} \sin \frac{{\pi \alpha_{2} x}}{180}} \right)}}{90}} \hfill \\ { - \frac{{\alpha_{2}^{2} \pi \exp \left( { - \alpha_{2} x} \right)\left( {D_{2} \cos \frac{{\pi \alpha_{2} x}}{180} - C_{2} \sin \frac{{\pi \alpha_{2} x}}{180}} \right)}}{90}} \hfill \\ { - \frac{{\alpha_{2}^{2} \pi^{2} \exp \left( {\alpha_{2} x} \right)\left( {A_{2} \cos \frac{{\pi \alpha_{2} x}}{180} + B_{2} \sin \frac{{\pi \alpha_{2} x}}{180}} \right)}}{32400}} \hfill \\ { - \frac{{\alpha_{2}^{2} \pi^{2} \exp \left( { - \alpha_{2} x} \right)\left( {C_{2} \cos \frac{{\pi \alpha_{2} x}}{180}} \right) + D_{2} \sin \frac{{\pi \alpha_{2} x}}{180}}}{32400}} \hfill \\ \end{array} } \right],} \hfill & {\quad \frac{L}{2} \le x \le \frac{L}{2} + \, l_{2} } \hfill \\ \end{array} } \right.$$
(8)

The continuity conditions for the deflection w(x), bending moment M(x), rotation θ(x), and shear force Q(x) at the simplified rock beam at the interface between the solid coal and the backfill are as follows:

$$\left\{ {\begin{array}{*{20}l} {w_{{\left( \frac{L}{2} \right)1}} = w_{{\left( \frac{L}{2} \right)2}} ,\;\;\theta_{{\left( \frac{L}{2} \right)1}} = \theta_{{\left( \frac{L}{2} \right)2}} } \hfill \\ {M_{{\left( \frac{L}{2} \right)1}} = M_{{\left( \frac{L}{2} \right)2}} ,\;\;Q_{{\left( \frac{L}{2} \right)1}} = Q_{{\left( \frac{L}{2} \right)2}} } \hfill \\ \end{array} } \right.$$
(9)

The deflection curve and bending moment curve of the rocking beam above the backfill are symmetrically distributed. Assuming the rock beam is a fixed-end long beam, the boundary conditions of the rock beam can be obtained as follows:

$$\left\{ {\begin{array}{*{20}l} {w_{(0)1} = 0,\;\;M_{(0)1} = 0} \hfill \\ {w_{{\left( {\frac{L}{2} + l_{2} } \right)2}} = 0,\;\;M_{{\left( {\frac{L}{2} + l_{2} } \right)2}} = 0} \hfill \\ \end{array} } \right.$$
(10)

By substituting the values, we obtain the parameters A1, B1, C1, D1, A2, B2, C2 and D2. Using these parameters can further derive the bending moment curve of the basic roof beam.

Fracture criteria of the basic roof

The maximum tensile stress σmax of the basic roof beam can be obtained by solving the following equation:

$$\sigma_{\max } = \left| {\frac{{6M_{\max } }}{{h^{2} }}} \right|$$
(11)

where h is the height of the basic roof beam, Mmax is the maximum bending moment of the beam.

Assuming the tensile strength of the basic roof beam is [σ], according to the first strength theory34, the condition for the basic roof fracture can be obtained as:

$$\sigma_{\max } \ge \left[ \sigma \right]$$
(12)

The fracture mechanism of overlying strata in SRBM stope

The main controlling factors of overlying strata fracture

Based on the particularity of the stope working face layout and the principles of rock strata control, the key factors affecting the breakage of the overlying strata are:

  1. 1.

    Geological conditions: This includes the lithology, interlayer structure, coal seam conditions, and other geological and hydrological factors. These conditions determine the original stress state and strength of the overlying strata, directly influencing their breakage after mining and backfilling.

  2. 2.

    Mining height: Mining height significantly affects the breakage of the overlying strata. Different heights alter the dimensions of the mining space, impacting the stress balance and potentially causing local stress concentrations that lead to breakage.

  3. 3.

    Stope length: Like mining height, stope length affects the size of the mined space and the stress distribution in the overlying strata. An appropriately designed stope length can reduce the risk of breakage.

  4. 4.

    Backfill mechanical properties: The stiffness of the backfill in the goaf is critical for preventing overlying strata breakage. The backfill’s mechanical properties, which vary with composition, can effectively support the strata and prevent failure.

In summary, several factors control overlying strata breakage in SRBM stopes. While geological conditions are generally fixed, adjusting mining height, stope length, and backfill materials can effectively mitigate strata breakage and ensure safe mining operations.

Breakage patterns of overlying strata

Using the geological parameters of a certain mine in Yulin as an example, design a short-wall working face. The coal seam has an average thickness of 5 m and a burial depth of approximately 150 m. The abutment pressure zones on both sides of the working face extend to 30 m in length, with the stress concentration factor of ranging from 1 to 335,36, as determined by experimental studies. The immediate roof consists of 6.85 m thick feldspathic sandstone. Physical parameters of the overlying strata are shown in Table 2.

Table 2 Example of overburden parameters in working face.
Full size table

Based on the key factors affecting overlying strata fracture, various calculation schemes with different parameters were designed, as shown in Table 3. Using these parameters in the mechanical model, the basic roof bending moments were calculated for different mining heights, stope lengths, and backfill elastic moduli. The results are presented in Table 4.

Table 3 Calculation scheme.
Full size table
Table 4 Calculated bending moment values.
Full size table

The bending moments from Table 4 were substituted into Eq. (11) to calculate the corresponding tensile stress values. The variations in tensile stress under different factors are illustrated in Fig. 9. Figure 9a shows that the tensile stress increases in an “S” shape with the mining height. For mining heights of 2, 3, 4, and 5 m, the tensile stress is below the allowable value. However, at 6 m, the stress reaches 3.39 MPa, causing fracture. Figure 9b indicates a linear increase in tensile stress with stope length. The stress remains below the allowable limit for stope lengths up to 900 m, but rises to 3.59 MPa at 1000 m, leading to fracture. Figure 9c reveals a logarithmic relationship between tensile stress and backfill elastic modulus. A higher backfill modulus results in lower tensile stress and better control of the overlying strata.

Fig. 9
figure 9

Variation of tensile stress under the influence of different controlling factors. (a) Height mining. (b) Length of stope. (c) backfill’s elastic modulus

Full size image

Prediction method of basic roof tensile stress

Based on the aforementioned fitting analysis, SPSS software was used to conduct a multivariate nonlinear fitting analysis on the relationship between the basic roof bending moment and the main controlling factors, establishing a prediction model.

$$\sigma = 4.71{\text{e}}^{0.078H} + 36.56E^{ - 0.446} - 68.02L_{S}^{ - 0.046}$$
(13)

where H is the mining height, E is the elastic modulus of backfill, and Ls is the length of stope.

The goodness of fit for Eq. (13) is R2 = 0.96, indicating a close relationship between the variables in the regression equation. This suggests that the regression model is highly accurate.

Figure 10 shows the sensitivity of tensile stress to these factors. As seen, parameter E exhibits the steepest slope, followed by Ls, and then H. Thus, the sensitivity ranking is E > Ls > H. The model indicates that a combination of multiple parameters influences basic roof tensile stress. Excessive tensile stress, exceeding the allowable limit for the basic roof, increases the bending moment of the roof beam, causing fracture. This undermines the safety of the mining backfill and significantly affects its stability. Therefore, in mine face design and application, mining height, stope length, and backfill elastic modulus can be adjusted based on mine conditions and production needs.

Fig. 10
figure 10

Different controlling factors’ sensitivity to the primary roof tensile stress.

Full size image

Engineering verification test

Geological overview

The northeast coal mine in Yulin City, Shaanxi Province, has geological reserves of 7.29 million tons. The ___location of the mining area is shown in Fig. 11. Geological conditions are stable, with dip angles less than 1° and no significant faults. The No. 3 coal seam, the main mining seam, has an average thickness of 5 m, a dip angle of about 0.5°, and a burial depth of approximately 150 m. It is a stable, super-thick coal seam with a simple structure and no gangue intercalations. The specific mechanical parameters of the coal-rock layers are shown in Table 5, which provide parameters for subsequent numerical simulations.

Fig. 11
figure 11

Geographical ___location of the mining area.

Full size image
Table 5 The physical and mechanical parameters of the coal-rock layers.
Full size table

The 3101 working face of the mine is designed as a short-wall mining area. The working face is 100 m long, with an advancing length of 855 m. The average coal thickness is 5 m, and the coal seam has a dip angle of approximately 0.5°. The coal seam is divided into 171 crosscuts, each with a width of 5 m, along the advancing direction.

Cemented backfill material

Selecting appropriate aggregates and cementitious materials, and designing suitable proportions of backfill materials, is especially crucial. Based on the mine’s conditions, gangue is selected as the coarse aggregate, and fly ash and cement as cementitious materials. The cement content in the backfill material directly affects the deformation of the overlying strata37. Experimental determination of the backfill material parameters is conducted to select suitable proportions that meet the mine’s requirements. The raw materials and design testing process are shown in Fig. 12.

Fig. 12
figure 12

Backfill materials and testing methods.

Full size image

The backfill material’s slurry concentration is 78%, with four gangue particle sizes and three solid material mass percentages, resulting in 12 combinations. Workability and pumpability were assessed through standard slump tests. Cubic backfill specimens (70.7 × 70.7 × 70.7 mm) were prepared and cured for 14 days at a temperature of 20 ± 1 °C and relative humidity of 90 ± 5%, followed by mechanical property testing.

Table 6 presents the slump values of the backfill slurry, ranging from 126 to 228 mm. The slump value reflects the flow characteristics of the material: a larger slump indicates better flow. In mine engineering, concrete slurry pipelines should have a slump greater than 150 mm38. Thus, all mixtures, except for X12, meet this requirement.

Table 6 Slump test value.
Full size table

Figure 13 presents the stress–strain curves of specimens with varying fly ash proportions under the A2 gangue particle size grouping. The curve is divided into five stages: OA (pore compression and closure), AB (elastic), BC (plastic deformation), CD (yielding), and DE (post-peak). Each specimen exhibited tensile cracks extending from the bottom to the top or from the top to the bottom, resulting in tensile failure. Table 7 shows the uniaxial compression strengths and elastic modulus for the specimens. It can be seen that the UCS varies from 3.30 to 4.32 MPa, and the elastic modulus varies from 0.44 to 0.62 GPa. The slump values and uniaxial compressive strength of the backfill material in Tables 6 and 7 are compared with the borehole values39.

Fig. 13
figure 13

Complete stress–strain curves of backfill specimens.

Full size image
Table 7 The UCS and elastic modulus of backfill specimens.
Full size table

Selection of engineering parameters

Based on the mine’s geological conditions, the basic roof thickness is 6.85 m with a tensile strength of 3.3 MPa. To optimize production costs and economic benefits, the mining height and stope length are designed to their maximum values of 5 m and 855 m, respectively. Substituting these parameters into Eq. (13), the minimum elastic modulus of the backfill is found to be 0.592 GPa. According to the data in Table 7, the proportions of X2, X7, and X10 meet the requirements, with X2 chosen for its costefficiency in gravel crushing and cement. The gravel particle size distribution is 25% (0.0–5.0 mm), 50% (5.0–10.0 mm), and 25% (10.0–16.0 mm), with solid material mass percentages of cement, fly ash, and gravel being 15%, 35%, and 50%, respectively.

Numerical simulation

Model establishment and methodology

FLAC3D has been proved a powerful and flexible engineering software widely used in the simulation research of mining and backfilling in mines14,17. FLAC3D allows for large deformations within the model, with the Lagrangian algorithm providing a reliable framework for assessing the stability of the crosscut. The model is simulated and analyzed using the Mohr–Coulomb strength criterion. A simulated advance length of 360 m is chosen, with a working face width of 100 m. The model dimensions are 460 m × 200 m × 160 m (length × width × height), and the crosscut width is 5 m. It is discretized into 1,609,031 nodes and 1,541,000 elements, with 50 m coal pillars at the front, rear, and lateral boundaries. Gravity is applied throughout the model, with no additional loads at the upper boundary. The vertical stress is 3.75 MPa, and the horizontal stress is 2.81 MPa. Displacement boundaries are set at the front, rear, left, right, and bottom, while the upper surface is free. The coal seam comprises 72 roadways, with mining and filling operations performed in three stages. The model is shown in Fig. 14.

Fig. 14
figure 14

Numerical computing model.

Full size image

The stress transfer patterns in the surrounding rock of the stope at different stages

As shown in Fig. 15, the vertical stress distribution in the coal and backfill bodies exhibits distinct characteristics during different mining and backfilling stages. Initially, the working face is supported jointly by the coal body and the backfill body, with the overburden load primarily transferred through the coal body. During the subsequent stage, although the support system remains a combination of coal and backfill, the backfill body assumes an increased share of the load. Notably, the maximum stress in the coal body and first backfill round escalates by 0.19 MPa and 0.007 MPa, respectively. By the final stage, the backfill body assumes full load-bearing responsibilities, with stress in the first and second backfill rounds rising by 0.61 MPa and 2.007 MPa, respectively. These findings indicate that mining and backfilling stages significantly influence stress transfer in the surrounding rock, with the maximum vertical stress in the working face vicinity displaying a progressive increase as mining advances.

Fig. 15
figure 15

Stress characteristics of surrounding rock mass under different stages.

Full size image

The pattern of surface movement varies at different stages

During the process of SRBM, surface movement and deformation occur, as shown in Fig. 16 for the movement, and deformation of various points in the rock layer after mining. Table 8 shows the classification of surface movement deformation and its relationship with the building protection levels.

Fig. 16
figure 16

The movement and deformation of rock layers at various points after mining.

Full size image
Table 8 Classification of surface movement deformation and its relationship with building protection levels40,41.
Full size table

Figure 17 illustrates surface displacement in the X, Y, and Z directions at different stages. The maximum surface displacement in the X direction is observed to increase from 3 to 11 mm across the three mining stages. Similarly, in the Y direction, the maximum displacement increases from 1.3 to 3.9 mm. Notably, subsidence in the Z direction rises progressively from 39 to 58 mm. These results demonstrate a continuous increase in surface displacement across all directions with advancing mining and backfill stages.

Fig. 17
figure 17

Displacement in various directions at different stages on the ground surface.

Full size image

The surface deformation parameters, including inclination, curvature, and horizontal deformation, and their relation to the critical threshold for building protection can be determined, as summarized in Table 9. The analysis reveals that these parameters increase progressively with advancing mining and backfilling operations across the three stages. Notably, all measured deformation values remain significantly below the threshold for Grade I building protection.

Table 9 The relationship between surface movement deformation parameters and the critical threshold.
Full size table

In summary, the SRBM technique effectively mitigates overlying strata movement deformation, reduces surface movement impact, and minimizes damage to surface ecology and structures.

Engineering application effectiveness

The on-site support parameters are shown in Table 10. On-site monitoring of surrounding rock stability at the working face was conducted using roof subsidence meters and stress meters, with the monitoring setup shown in Fig. 18 (based on excavation of crosscut No. 62). Four monitoring stations were established within the excavation, spaced 25 m apart. Roof subsidence meters were installed on the roof, and stress meters were placed 2 m deep on both sides of the crosscut.

Table 10 Support parameters.
Full size table
Fig. 18
figure 18

Monitoring deployment plan.

Full size image
  1. (1)

    Analysis of roof subsidence

Data from monitoring stations 1 and 2 were retrieved at intervals of 1 h, 3 h, 24 h, 30 days, and 90 days, as shown in Fig. 19. The results indicate that during excavation, subsidence rates at the two points were 2.76 mm/day and 11.52 mm/day, respectively. After backfilling, these rates decreased to 0.17 mm/day and 0.19 mm/day. The roof exhibited a slow subsidence trend under backfill support, tending toward stability and effectively maintaining structural integrity.

Fig. 19
figure 19

Observation of roof subsidence values.

Full size image
  1. (2)

    Analysis of load-bearing structure stress monitoring

Stress data from different nodes at monitoring station 2 were retrieved at 3-day intervals, as shown in Fig. 20. The stress in the solid coal increased from 4.38 to 4.69 MPa, while stress in the backfill rose from 0.44 to 2.03 MPa. This suggests that the solid coal remains the primary load-bearing structure, with normal stress distribution and a stable supporting structure, confirming the validity of the simulation results.

Fig. 20
figure 20

Observation of stress in the load-bearing structure.

Full size image
  1. (3)

    Integrity of surface and load-bearing structure

Figure 21a and b show the surface conditions for different mining methods. After applying the SRBM technique, the surface remained intact, with no cracks or collapses. Field monitoring revealed maximum values of surface tilt, curvature, and horizontal deformation of 0.17 mm/m, 0.0036 mm/m2, and 0.22 mm/m, respectively, all well below the limit values in Table 8. These results confirm that the SRBM method effectively mitigates surface deformation and protects the ecological environment. Figure 21c and d show the surrounding rock on both sides of crosscut No. 62, highlighting the integrity of the backfill and coal wall.

Fig. 21
figure 21

Real shots of surface and load-bearing structures. (a) Subsidence method. (b) SRBM. (c) Backfill. (d) Solid coal

Full size image

Conclusion

  1. 1.

    This paper presents a short-wall roadway backfill method (SRBM) and a corresponding design protocol to address issues such as gangue and fly ash accumulation and ecological damage in mining areas. The basic principles of strata control in SRBM are analyzed, and a mechanical model for determining the failure deformation of the basic roof beam, based on the elastic foundation beam theory, is established.

  2. 2.

    The effects of geological conditions, mining height, stope length, and backfill mechanical properties on roof failure in SRBM workings were examined. The relationship between basic roof tensile stress and controlling factors was clarified: under specific geological conditions, tensile stress increased with mining height and stope length, but decreased with the elastic modulus of the backfill.

  3. 3.

    A prediction model for basic roof tensile stress was developed using multivariate nonlinear regression, based on the relationship between controlling factors and stress. Sensitivity analysis indicated that the elastic modulus of the backfill had the greatest impact, followed by stope length and mining height. Different backfill material ratios were designed for varying factors, and corresponding parameters were obtained through flowability and mechanical property testing.

  4. 4.

    Based on the designed geological conditions, mining parameters were determined: a stope length of 855 m, a mining height of 5 m, and a backfill mixture of X2. Three-dimensional numerical simulations revealed that with the progression of mining and backfilling, both the maximum vertical stress in the load-bearing structure and surface movement deformation increased. The application was validated through field monitoring.