Abstract
When dealing with cross-fault mining operations, geological faults pose significant challenges to by unsettling the existing stress conditions, causing lopsided deformation of the surrounding rock, sudden shifts in displacement, and creating hazards such as fault slippage that really hinder the application of the application of Roof Cutting and Retaining by Mining technology. Drawing upon the 11,101 working face at Qipanjing Coal Mine as its foundation, this research brings together theoretical analysis, finite element modelling, and field verification to get to the bottom of the complex dynamics at play. Based on key stratum and voussoir beam theories, a three-stage mechanical model for fault-crossing mining was established. Drawing upon fundamental stratum mechanics and voussoir beam theory, researchers formulated a three-stage mechanical framework for mining across fault lines. This model demonstrates that advanced stress initially plummets as a result of the "fault stress barrier effect" before gradually rebounding once the working face withdraws from the immediate vicinity. Using the double-yield model to simulate caved gangue characteristics, and introducing deviatoric stress to compare three slotting heights, the study found that higher slotting heights reduce surrounding rock distortion energy and compressive deformation intensity, with 10 m achieving optimal economic efficiency (peak deviatoric stress invariant reduced by 24.8 MPa2, overlying strata displacement by 0.11 m, fault displacement by 0.04 m). A “slotting + loose blasting” scheme was proposed. On-site verification showed it cuts hydraulic support force by 15.3% and gangue bulking coefficient by 4.3%, providing technical support for RCRM application in fault-developed areas.
Keywords: Cross-fault mining, Fault stress, RCRM, Strata displacement, Deviatoric stress
Introduction
The Roof Cutting and Retaining by Mining (RCRM) technology utilizes caved gangue for automatic filling to form roadways, and it boasts advantages such as improving resource recovery rate, reducing surrounding rock stress concentration, and lowering roadway maintenance costs1–4. However, with the continuous exploitation of resources, faults, among the most ubiquitous geological features encountered in mining operations, now present a formidable obstacle to both the safety and efficacy of this excavation method5,6.
Faults tend to have natural weak planes, which creates a stark contrast in stress field characteristics between fault zones and typical geological settings. On one hand, fault stress concentration may induce uneven rock deformation and dramatic displacement increases7–10; On the other hand, The contrasting mechanical characteristics observed in the hanging wall and footwall of a geological fault can serve as a catalyst for fault slippage, ultimately unleashing devastating high-energy catastrophes including rock bursts, intense ground pressure events, and widespread structural collapse of underground excavations11,12.
To investigate the disaster-causing mechanism of faults, relevant researchers have conducted extensive studies. Li et al.7 Developed systematic mechanical and numerical frameworks to pinpoint the correlation between fault slip motion, fault inclination, and frictional characteristics of fault planes, and subsequently deduced the essential prerequisites for fault sliding. Shan et al.13 developed a mechanical model for mining at working faces adjacent to fault-related roadways, findings indicated that the gap between the fault and the roadway is the chief determinant of stress concentration and the resulting plastic deformation in the surrounding rock. Zhang et al.14 obtained the occurrence mechanisms of three types of pillar-type impact ground pressure through the establishment of mechanical models. Onur Vardar et al.15 used UDEC (Universal Distinct Element Code) to investigate the influencing factors of rock bursts in fault-proximal roadways; the results showed that as mining depth increases, the total seismic energy and peak value generated near the roadway increase rapidly, raising the probability of rock burst occurrence. Sun et al.16 researched the mechanism by which different fault dip angles affect roadway instability under the condition of 1000 m deep mines, arguing that the uneven stress field formed in the shallow surrounding rock is a crucial factor leading to roadway instability.
Unlike traditional mining approaches, RCRM technology banks on the expansive characteristics of gangue to bear the weight of overlaying rock layers and create the tunnel necessary for the next extraction area. The system’s integrity hinges directly on this gangue, which in turn is affected by the pressure from the strata above. Given the intricate stress conditions typical of fault zones, it’s imperative to implement robust strategies for managing the deformation of surrounding rock. Current research approaches to surrounding rock control in faulted areas can be broadly categorized into three distinct tracks. The first is improving support strength. To tackle the significant deformation of surrounding rock triggered by intense stress disturbances from faults, engineers implement a range of countermeasures. These include beefing up the anchor cables’ support capacity and effectiveness, applying grouting and shotcrete techniques, and installing supplementary temporary supports to stabilize the area13,17–21; The second is reducing the energy generated and transmitted during fault slippage through technologies such as hydraulic pressure and slotting blasting, thereby lessening the interactional pressure exerted on the surrounding rock20; This final approach integrates enhancing rock mass support while mitigating fault displacement effects. For instance, Sun et al.16 implemented the control technique of “roof cutting + CRAC cable”, Fu et al.22 used the combined control scheme of “portal support + roof pre-cracking”, and Liu et al.23 proposed the supporting scheme of “asymmetric reinforcement support + roof directional blasting”. Although existing achievements have provided many useful explorations on surrounding rock control under fault structural conditions, research on the special control technologies for gob-side entry retaining in fault zones remains relatively limited, and the relevant theoretical and technical systems need to be further improved.
Although existing studies have initially explored the impact of faults on stope ground pressure and proposed corresponding control technology schemes, under the cross-fault mining scenario of RCRM technology, the changes in surrounding rock support stress and instability characteristics during the working face crossing the fault remain unclear, and effective control schemes are also lacking. This study aims to analyze the variation characteristics of surrounding rock support pressure and the evolution law of displacement during the working face crossing a fault by combining theoretical analysis, numerical simulation, and on-site experiments. A targeted control scheme is proposed based on the research on the bulking characteristics of gangue. The results of this study provide an effective technical scheme for the safe and efficient mining of RCRM working faces crossing faults.
Project overview
Currently, the main coal seam mined at the East Area of Qipanjing Coal Mine is the 9# coal seam. To enhance the recovery efficiency of resources and decrease the extent of roadway excavation, the RCRM technology is applied in the 11,101 working face, and the transportation roadway is used as the return airway for the next adjacent mining operation. As shown in Fig. 1, the roadway roof is reinforced with a combination of NPR (Negative Poisson’s Ratio) anchor cables and conventional anchor cables. At the same time, blasting holes for cutting cracks are arranged in the direction of the working face at the angle between the coal face support and the roof of the roadway. The average depth of the coal seam from the surface is 510 m, and its thickness ranges from 2.8 to 3.5 m. The direct roof and direct floor are respectively 8 m and 3.2 m thick sandy mudstone.
Fig. 1.
Location of the working face fault.
Running parallel to the 11,101 working face’s mining direction, fault ZF44 exhibits a normal displacement of 4400 mm, striking at a 75° relative to the horizontal plane of excavation. Due to the existence of a gap between the faults, there is a slope at the intersection of the 11,101 transportation roadway and the fault; the clear height of the roadway in this section can reach 4.5 m, the temporary support effect is poor, and there is a risk of roadway collapse. The positional relationship between the working face and the fault is shown in Fig. 1.
Mining characteristics
Behavior of the RCRM working face passing through faults
During the fault-crossing mining process, the RCRM technology demonstrates unique advantages. When compared to the traditional longwall mining methods that grapple with stress concentration and rock deformation issues, this new technology cleverly severs the structural tie between the roadway and working face roofs by employing a pre-roof cutting strategy. At the same time, it capitalizes on the load-bearing capacity and bulk properties of gangue in the goaf, which substantially lessens the intensity of stress concentration.
This pioneering method not only reduces the plastic area around the roadway, but also minimizes the cantilever span of the upper structure, thereby enhancing the stability and safety of the roadway. Therefore, this technology enables the implementation of pillarless mining in geologically abnormal areas, improves coal recovery efficiency, and reduces mining costs.
As shown in Fig. 2a, the strike of the ZF44 fault forms an angle of approximately 75° with working face trend. Therefore, the process crossing the fault during mining is simplified as illustrated in Fig. 2b–d. According to the spatial separation from the fault to the operational area, the fault-crossing process of the RCRM working face can be divided into three key stages. Mining stage Ⅰ: The working face is mined on the upper part of the fault, and at the same time, it does not come into contact with the fault plane. Mining stage Ⅱ: The hydraulic support of the working face is located at the fault plane. Mining stage Ⅲ: The working face is mined beneath the fault and gradually moves away from it.
Fig. 2.
Decomposition diagram of the RCRM working face over the fault.
RCRM fault mechanics analysis
In order to analyze the stress transmission characteristics of the working face when it is at different positions relative to the fault under the RCRM technology, cross sections close to the retained roadway side were selected for analysis, as shown in Fig. 2b–d.
Mining stage Ⅰ
As shown in Fig. 3. Based on the equilibrium equation established by the voussoir beam theory, the rock stratum structure on the hanging wall is divided into three parts, namely Key Block E (main roof above the goaf), Key Block F (main roof above the working face), and Key Block G (immediate roof above the working face)24. Half the load from Key Block E is supported by the gangue, while the other half rests on the main roof above the fault coal pillar, with these two components assumed to split the difference equally.
Fig. 3.
Mining stage I.
In addition, through the volume expansion of the gangue itself, the collapsed gangue comes into contact with Key Block F and Key Block G respectively. However, Key Block F and Key Block G have not yet rotated to squeeze the gangue, and the gangue is in a state of slight contact at this time. Thus, it is also considered that the goaf gangue does not provide support force for the roof.
Suppose that the load on the main roof above the fault coal pillar and the coal gangue in the goaf is transferred proportionally by factor K to the lower part of the fault. Therefore, the load borne by Key Block G above the fault coal pillar is
 |
1 |
where 
Load transferred from Key Block E to Key Block F, MN/m; 
: Self-generated load of Key Block F, MN/m; K: Load transfer coefficient of the fault plane, 
;
: Total load borne by the main roof per unit area plus its self-generated load, 
, MN/m2; 
, m.
The load generated by Key Block G is
 |
2 |
Therefore, the total load borne by the hydraulic supports at the working face and the coal pillars in the fault zone is
 |
3 |
where 
, MN/m; 
: Width of the strike cross-section of the working face, m; 
: Width of the fault coal pillar, m; 
, m; 
: Working face mining height, m; 
: Height of the main roof, m; 
: Contact width between Key Block F and gangue, 
, m; 
: Height of the immediate roof, m; 
: Bulk density of the immediate roof stratum, N/m3; 
: Dip angle of the fault, °.
With the ongoing advancement of mining operations, the breadth of the fault coal pillar steadily narrows, yet all other variables stay put. As a result, the cumulative pressure on the working face’s hydraulic supports and the fault coal pillar intensifies. Once this pressure surpasses the bearing capacity, issues like crushed hydraulic supports or the fault coal pillar succumbing to plastic deformation can arise. In the end, this excess pressure shifts onto the surrounding rock on either side of the mining passageways, thereby heightening the stress build-up in the rocks immediately in front and behind the working face.
Mining stage Ⅱ
After the working face finishes mining the coal seam in the hanging wall, the fracture characteristics and connectivity patterns of the rock strata overlying the goaf are presented in Fig. 4. The fault plane transfers the main roof load above the goaf gangue to the fault footwall at a certain proportion K.
Fig. 4.
Mechanical model for completed mining of the fault hanging wall.
Therefore, the load 
transferred from Key Block E and Key Block F to the footwall is
 |
4 |
The vertical component 
of the load acting on the coal seam with a width of 
ahead of the working face is
 |
5 |
The load 
generated by the immediate roof above the working face and the coal mass ahead of it is
 |
6 |
The load 
generated by the main roof itself and the overlying rock strata is
 |
7 |
The total load 
borne by the working face hydraulic supports and the coal mass ahead of them is
 |
8 |
Comparative analysis of Eqs. (3) and. (8) reveals that when the working face has accomplished mining of the fault hanging wall coal seam and the overlying immediate roof length is relatively short, the vertical stress on the hydraulic supports and advanced coal mass mainly comes from the vertical load transferred from the rock strata above the hanging wall gangue accumulation body to the footwall. Thus, an sudden drop in the working face advance stress will happen right after the working face completes mining of the fault hanging wall coal seam.
Mining stage Ⅲ
When the working face gradually advances away from the fault for extraction, the fracture and connection modes of the overlying strata above the goaf are illustrated in Fig. 5.
Fig. 5.
Mining on the footwall.
Based on the analysis presented in the preceding section, it can be concluded that the vertical component 
is
 |
9 |
Similar to the scenario where the working face mines in the vicinity of the fault, it is also considered that immediately after the collapse of the goaf gangue, the gangue is in a state of weak contact with both the immediate roof and the main roof. Meanwhile, the immediate roof behind the working face hydraulic supports does not undergo rotational subsidence. Consequently, the goaf gangue fails to provide effective supporting force for either the immediate roof or the main roof of the working face. As a result, the entire load 
transmitted from the fault’s hanging wall to its footwall is borne by the main roof overlying the working face in the fault’s footwall.
The load 
generated by the immediate roof is
 |
10 |
The load 
generated by the main roof and the overlying rock layer is
 |
11 |
The total load 
borne by the hydraulic supports and the coal mass ahead of them is:
 |
12 |
Analysis of Eqs. (8) and (12) reveals that when the working face mines toward the footwall with increasing distance from the fault, the vertical component 
of the load transferred from the hanging wall to the footwall and the width 
of the calculated advanced coal mass remain unchanged. However, the width of the overlying immediate roof expands, and the generated load 
continues to increase. Thus, as the working face moves away from the fault, the advance stress rises continuously until it returns to the normal geological stress state.
Numerical simulation of mining across fault
Numerical simulation scheme
Gangue characteristics
After gangue collapses, accumulates, and fills the goaf, it is gradually compressed under the pressure of the goaf rock strata. In laboratory experiments, this process can be approximately reproduced using the confined uniaxial compression test of crushed stone. According to the results of previous experiments, compaction behavior of collapsed gangue follows the equation below25
 |
13 |
where 
represents the pressure borne by the caving gangue during the compression process; 
represents the prime Young’s modulus of the gangue accumulation mass; 
stands for the strain generated in the compression process; and 
is the bulking coefficient of the goaf gangue.
Then the maximum strain exhibited by the gangue can be expressed as
 |
14 |
YAVUZ26 through a large number of laboratory experiments, a obtained the expression for the initial elastic modulus of the gangue accumulation body.
 |
15 |
where 
indicates the unconfined compressive strength of the caving gangue in the state corresponding to standard rock samples.
By combining Eqs. (13)–(15), we can obtain
 |
16 |
It can be seen from Eq. (16) that the larger the bulking coefficient, the lower the bearing strength of the gangue. This is because a larger bulking coefficient leads to a higher porosity, which provides more space for the gangue to deform, thus resulting in a lower deformation resistance of the gangue accumulation body. Combined with Fig. 6, it can be observed that the supporting strength 
of the gangue accumulation body is negatively correlated with the bulking coefficient 
, and positively correlated with the strength 
of the gangue.
Fig. 6.
Influencing factors of bearing capacity of goaf gangue accumulation body.
Based on the characteristics of RCRM technology, it uses gangue as the gangue wall and incorporates it as an important component of the retained roadway. Therefore, the supporting strength of the collapsed gangue directly affects the final deformation of the roadway’s surrounding rock. Meanwhile, combined with the comprehensive analysis results of Eqs. (13)–(16), the goal of suppressing surrounding rock deformation in the fault zone can be attained by increasing the height of the caving zone, narrowing the void spaces in the gangue accumulation body, and improving the load-bearing strength of the gangue accumulation body.
Simulation scheme
-
Construction of the numerical model
Based on the actual engineering geological conditions of the 11,101 working face in the Eastern District of Qipanjing Coal Mine, numerical models with a slot height of 8 m (bulking coefficient = 1.4), 10 m (bulking coefficient = 1.32), and 12 m (bulking coefficient = 1.27) were established respectively. It is assumed that under the above slot height conditions, the roof can be fully bulked after caving and completely fill the goaf. To reduce the calculation load of the numerical model, the model dimensions are specified as 400 m × 40 m × 120 m (length × width × height), as shown in Fig. 7. The inclination of the model’s fault is 75° and the drop is 4400 mm. The fault plane is simulated using an interface, with a cohesion of 0.3 MPa, an internal friction angle of 15°, a normal stiffness of 5.0 GPa, and a shear stiffness of 2.0 GPa for the fault plane. The model is restrained from moving horizontally along its front, rear, left, and right faces, while the bottom surface is fixed against vertical movement. The model is progressively dug from the hanging wall toward the footwall of the fault, advancing in 10 m increments that proceed only after stability has been reached. A 10 m wide protective coal pillar is left in place prior to both the initial and final excavation phases. Drawing upon findings from in-situ stress measurements and stress gradient computations, an 8.6 MPa vertical stress is imposed downward onto the model’s upper surface to replicate the overlying strata’s influence, while lateral boundaries are subjected to horizontal stress maintaining a 0.5 lateral pressure coefficient.
Table 1 presents the physical and mechanical properties for each rock layer in our simulation model. To capture the unique behavior of broken rock in the mined-out area that provides support when compressed, we employed the Double-Yield model to effectively represent the effects of gob material that fills the void after collapse and increases in volume.
-
Simulation of gangue compression characteristics
It can be concluded from the above analysis results that the gangue in the goaf exhibits the characteristic of gradually increasing strength with the increase in strain. Regarding how to realize this strain hardening characteristic in numerical simulations, the prevailing consensus among researchers indicates that the Double-Yield (D-Y) model implemented in FLAC3D effectively captures the stress restoration behavior of waste material in goaf, making it suitable for application in both backfilling excavated voids and collapsed zones27,28.
The uniaxial compressive strength
of the caved gangue is 51.2 MPa. This value is substituted into Eq. (16) to calculate the stress–strain relationships (Capping pressure) of the gangue under bulking coefficients of 1.4, 1.32, and 1.27 respectively. Table 2 displays the computed values.In the Double-Yield constitutive model, several commonly used parameters related to the stress–strain characteristics of materials are typically the maximum elastic bulk modulus
, maximum elastic shear modulus 
, density 
, dilation angle 
, internal friction angle 
, and Capping pressure 
. To ensure that the material properties exhibited by the numerical model align as closely as possible with the stress–strain characteristics calculated in Table 2, it is necessary to optimize the maximum elastic bulk modulus 
, maximum elastic shear modulus 
, dilation angle 
, and internal friction angle 
, so that the calculation results of the model can better reproduce the on-site conditions.To fine-tune the parameters within the D-Y model, a grid of 1 m × 1 m × 1 m was crafted using FLAC3D for a uniaxial compression simulation. A meager force of 1 × 10⁻5 m/s was exerted at the cube’s peak to mimic the loading, while the base and all adjacent sides were capped with zero-velocity boundary constraints. The capping pressure and material density were treated as fixed values; by continuously adjusting the remaining parameters, the stress–strain curve calculated by the numerical model was fitted to the theoretically calculated results in Table 2 to the greatest extent possible. Figure 8 shows the theoretical and numerically calculated stress–strain curves under the three bulking coefficient conditions, and the corresponding mechanical parameters in the D-Y model are presented in Table 3.
-
Excavation and monitoring plan
As shown in Fig. 9, the first mining step starts from the fault coal pillar, with a mining width of 10 m for each step. The excavation width is 2.6 m when approaching the fault. Then, the calculated Double-Yield model material is input into the model according to the corresponding filling height for equilibrium calculation. After the equilibrium is achieved, the next mining step is carried out.
In order to monitor the displacement of the roof of the goaf under different slit heights, displacement monitoring points were arranged in the model as illustrated in Fig. 9. The displacement monitoring points for the upper rock layer are located 0.1 m above the backfill, numbered m1–m17. Two vertical displacement monitoring lines were set 2 m away from both sides of the junction between the hanging wall and footwall. The first monitoring points of these two lines are h1 and f1 respectively, which are located 0.1 m above the backfill on the fault’s hanging wall. The interval along the fault interface upward is 10 m (vertical height), with 6 monitoring points arranged on each side of the fault.
Fig. 7.
Numerical calculation grid and boundary conditions.
Table 1.
Rock layer mechanical properties in simulation model.
| Lithology | ρ (kg/m3) | E (GPa) | φ (°) | c (MPa) | Te (MPa) |
|---|---|---|---|---|---|
| Fine sandstone | 2621 | 32.4 | 32 | 2.5 | 4.84 |
| Sandy mudstone | 2580 | 21.8 | 31 | 2.1 | 3.60 |
| Sandy mudstone-2 | 2562 | 20.7 | 31 | 2.2 | 3.61 |
| Coal | 1710 | 4.6 | 22 | 0.92 | 0.43 |
| Coarse sandstone | 2536 | 37.3 | 35 | 3.2 | 4.32 |
| Medium sandstone | 2600 | 32.8 | 33 | 3.3 | 4.1 |
Table 2.
Capping pressure in the D-Y model.
| Strain ε | Stress (MPa) | Strain ε | Stress (MPa) | ||||
|---|---|---|---|---|---|---|---|
| Kbmax = 1.4 | Kbmax = 1.32 | Kbmax = 1.27 | Kbmax = 1.4 | Kbmax = 1.32 | Kbmax = 1.27 | ||
| 0.00 | 0 | 0 | 0 | 0.14 | 12.91 | 24.52 | 40.85 |
| 0.01 | 0.49 | 0.77 | 1.05 | 0.15 | 14.86 | 29.12 | 50.75 |
| 0.02 | 1.01 | 1.61 | 2.2 | 0.16 | 17.11 | 34.83 | 64.43 |
| 0.03 | 1.58 | 2.53 | 3.48 | 0.17 | 19.75 | 42.11 | 84.53 |
| 0.04 | 2.19 | 3.55 | 4.91 | 0.18 | 22.89 | 51.73 | 116.96 |
| 0.05 | 2.85 | 4.66 | 6.51 | 0.19 | 26.68 | 65.02 | 178.08 |
| 0.06 | 3.57 | 5.9 | 8.33 | 0.20 | 31.36 | 84.58 | 336.25 |
| 0.07 | 4.36 | 7.28 | 10.4 | 0.21 | 37.28 | 116.19 | |
| 0.08 | 5.23 | 8.84 | 12.78 | 0.22 | 45 | 176.01 | |
| 0.09 | 6.18 | 10.59 | 15.55 | 0.23 | 55.49 | 332.12 | |
| 0.10 | 7.24 | 12.6 | 18.81 | 0.24 | 70.56 | ||
| 0.11 | 8.41 | 14.9 | 22.71 | 0.25 | 94.09 | ||
| 0.12 | 9.73 | 17.59 | 27.45 | 0.26 | 135.9 | ||
| 0.13 | 11.22 | 20.75 | 33.34 | 0.27 | 230.94 | ||
Fig. 8.
Numerical calculations and theoretical calculations fitted results.
Table 3.
Parameters in the D-Y model.
| Slit height (m) | K bmax |
 (GPa) |
 (GPa) |
 (kg/m3) |
 (°) |
 (°) |
|---|---|---|---|---|---|---|
| 8 | 1.40 | 3.36 | 4.7 | 1843 | 13.6 | 3 |
| 10 | 1.32 | 8.36 | 6.7 | 1955 | 10 | 20 |
| 12 | 1.27 | 7.36 | 6.7 | 2037 | 10.4 | 22 |
Fig. 9.
Displacement measurement point layout.
Surrounding rock evaluation indicators
In underground engineering, if only the vertical stress distribution is focused on and the geotechnical mass is simplified as a uniaxially loaded material, the actual situation where the geotechnical mass is subjected to multi-directional stresses will be ignored. It is difficult to accurately determine the elastic deformation degree, strain state, and stability of the rock based on the vertical stress distribution29. In accordance with elastoplastic mechanics theory, the deformation degree and state of a material mainly depend on the stress deviator, whose magnitude can reflect the elastoplastic state of the engineering rock mass30. The surrounding rock stress results from combined in-situ and deviatoric stress components. The plastic failure exhibited by rock masses is fundamentally governed by deviatoric stress, which primarily dictates the resulting plastic deformation patterns. Consequently, a thorough investigation of the deviatoric stress tensor proves crucial for comprehending the evolution and expansion of plastic zones within various geotechnical formations.
With coordinate axes aligned to principal directions, the three stress deviator invariants become expressible as30–32:
 |
17 |
In the equation, 
, 
and 
represent the three principal stresses, 
, 
and 
denote the expressions of the first, second, and third invariants of the deviatoric stress tensor in terms of the principal stresses, respectively.
Equation (17) clearly demonstrates that the first invariant of the stress deviator consistently equals zero, rendering it from a research perspective. In contrast, the second invariant of the stress deviator not only goes hand in hand with the distortional energy experienced by the material under stress but also shows a direct parallel with shear stress levels29. The stress and elastoplastic behavior of the geotechnical mass can be analyzed by examining the second invariant of the stress deviator. Similarly, in the field of mining engineering, rocks typically exhibit the characteristic of “being prone to compression but not tension”. When J3 is greater than zero, the element behaves as being in a tensile state; When J3 is equal to zero, the unit exhibits a planar state; and When J3 is less than zero, the unit exhibits a compressive state. The stress and elastoplastic behavior of the geotechnical mass can be analyzed by examining the second invariant of the stress deviator.
To get a handle on how the J2 and J3 are distributed backfilled goaf rock and in the advance stress influence zone under various bulking coefficients, while also examining the shear stress levels, distortional energy density, and strain patterns in the surrounding rock at different points as the hanging wall working face progresses, researchers must set up a stress monitoring line straight down the middle of the coal seam along the mining direction. This allows them to keep tabs on the maximum, minimum, and intermediate principal stresses affecting the surrounding rock. Once they’ve gathered these readings, they can plug the principal stress values into Eq. (17) to work out the corresponding J2 and J3.
Simulation results
Deviatoric stress characteristics of surrounding rock
-
Distribution features of J2
As shown in Fig. 10, J2 before and after the working face exhibits a “positive triangular” distribution pattern.
The J2 value of the surrounding rock in the backfilled part of the goaf and the non-mining part relatively far from the working face is relatively low, while stress concentration occurs in the rock mass closer to the working face, and the J2 in this area is relatively high. This indicates that high deformation energy is stored in the coal wall.
As illustrated in Figs. 10 and 11, with the working face and the fault becoming increasingly close to each other, the peak value of the J2 in the coal wall increases progressively. This indicates that, affected by fault structures and surrounding rock movement, the distortional energy in the coal pillar ahead of the working face is increasing. In the actual production process, this phenomenon may lead to severe deformation of the coal wall ahead of the working face or the occurrence of coal burst, thereby affecting normal mining activities.
As illustrated in Figs. 10 and 11, with the increase in slot height, the apex value of the J2 in the coal body ahead of the hydraulic support decreases correspondingly. This demonstrates that increasing the slot height reduces the porosity of gangue in the goaf, thereby enhancing its load-bearing capacity. This phenomenon becomes increasingly obvious as the mining distance increases. When the mining distance of the working face is 40 m, the corresponding apex values are 43 MPa2, 42.1 MPa2, and 40 MPa2 respectively when the slot heights are 8 m, 10 m, and 12 m. The differences in apex values between the slot height of 8 m and those of 10 m and 12 m are 0.9 MPa2 and 3 MPa2 respectively. However, when the mining length is 160 m, these differences increase to 24.8 MPa2 and 48.8 MPa2 respectively. It’s clear to see that as the working face approaches the fault, the differences in deviatoric stress apex values between different slot heights expand significantly. The data clearly show that when surrounding rock stress is higher, boosting the slot height has a more pronounced impact on improving the support capacity of gangue. Meanwhile, as the slot height increases, the apex value of the second invariant in the coal wall decreases, while the average value of the backfill in the goaf increases. This is because increasing the slot height improves the load-bearing capacity of the gangue in the goaf, enabling it to bear more stope pressure and thus reducing the distortional energy in the coal mass ahead of the working face.
-
Distribution features of J3
As shown in Figs. 12 and 13, after the coal seam is excavated, the J3 exhibits an “inverted triangular” distribution pattern. The J3 is relatively large in the gangue-backfilled area after mining and the area far from the working face, while a “valley bottom” (low-value region) appears in the area near the coal wall ahead of the working face, and most of this region is in a compressive deformation state.
As shown in Figs. 12 and 13, with the working face approaches the fault, the “valley bottom” (low-value region) of the J3 in the coal body ahead of the hydraulic support becomes deeper. This indicates that, affected by fault structures and surrounding rock movement, the coal body near the hydraulic support is in a high-intensity compressive deformation state. Meanwhile, with the increase in slot height, the “valley bottom” of the J3 in the coal body ahead of the hydraulic support becomes increasingly shallow, and the J3 of the backfill increases progressively. This shows that a smaller slot height leads to a larger porosity between gangue particles and poorer support capacity of the gangue, making it difficult to provide effective support during roof weighting. Consequently, raising the slot height boosts the load-bearing capacity of the gangue while lessening the compression stress exerted on the coal body. This phenomenon becomes increasingly obvious as the mining distance increases; like the J2, higher surrounding rock stress amplifies the impact of gangue porosity reduction in increasing support strength.
Fig. 10.
Characterization of J2.
Fig. 11.
Maximum value of the J2 in front of the working surface.
Fig. 12.
Characterization of the third invariant vacriation of the bias stress.
Fig. 13.
Maximum value of the J3 in front of the working face.
Distribution characteristics of surrounding rock displacement
-
Goaf roof displacement
Take monitoring points m2 to m3 as an example for analysis. Figure 14 illustrates strata displacement patterns in the goaf as slot heights (bulking coefficients) change.
Figure 14a–b illustrates the roof displacement of the overlying strata is zero in the initial state, which is consistent with the actual displacement law of surrounding rock. With the continuous increase in the mining length, the roof displacement of the goaf under different slot heights all demonstrates initial gradual growth, accelerated expansion, moderate advancement, and eventual stabilization.
Furthermore, as the working face progresses to the m16 monitoring station, the roof displacement demonstrates considerably more dramatic fluctuations compared to earlier observations. This phenomenon can be attributed to the fact that during mining operations in close proximity to the fault, the fault coal pillar has been reduced to its narrowest dimension while the stress within the coal mass has peaked. Under these conditions, substantial compressive deformation is expected to take place, aligning perfectly with the conclusions drawn from the analysis of surrounding rock deviatoric stress invariants detailed in the preceding section.
Considering the m2 monitoring point as a case study, when the slot heights measure 8 m, 10 m, and 12 m, the resulting maximum displacements clock in at 0.42 m, 0.31 m, and 0.27 m respectively. Compared to the 8 m slot height scenario, the maximum roof displacements drop by 0.11 m and 0.15 m. This clearly indicates that, as the slot height increases, the gangue’s ability to bear load in the goaf gets a significant boost, leading to reduced subsidence in the overlying strata. This discovery points to a promising avenue for controlling strata subsidence in goaf during mining operations.
-
Variation law of fault displacement
Figure 15a–c shows the distribution characteristics of fault displacement under different slot heights. The data clearly shows that when the working face remains a considerable distance from the fault, the fault displacement remains negligible, which suggests that the excavation process has minimal impact on the fault under these conditions. As the mining face advances continuously, a marked uptick in fault displacement becomes evident upon completion of the tenth excavation, though the rate remains gradual at this stage—with the fault situated 72.6 m away. Once the mining face closes the gap to just 32.6 m following the fourteenth excavation, the fault displacement enters a period of accelerated growth, and the fault starts to exhibit clear sliding behavior under the mining activities’ influence. Following the conclusion of the sixteenth excavation, the fault displacement hits its peak rate of change. This phenomenon can be attributed to the coal pillar between the fault planes being unable to bear the weight of the overlying rock strata, leading to compressive deformation—a finding that corroborates the conclusions drawn in Section “Deviatoric stress characteristics of surrounding rock”.
Figure 15d shows the maximum fault displacements at 6 monitoring points under the three slot heights. When the slot height is 8 m, the fault displacement is the largest, while the fault displacements are relatively smaller when the slot heights are 10 m and 12 m. Taking the h6f6 monitoring point as an example, the fault displacement is 0.182 m when the slot height is 8 m, while it is only 0.143 m and 0.142 m respectively when the slot heights are 10 m and 12 m. The findings indicate that increasing the slot height significantly improves the gangue’s load-bearing capacity while effectively reducing the fault’s sliding displacement. This discovery proves particularly valuable for controlling the surrounding rock as the mining face progresses through the fault zone.
Fig. 14.
Displacement evolution law of overlying rock layers.
Fig. 15.
Evolution law of fault displacement.
Rock control scheme
Control mechanism
When the working face sits in close proximity to a fault, the support force applied to the fault’s hanging wall is a complex interplay of several elements, with the primary contributors being the gangue, the coal body within the hanging wall, and the rock mass in the footwall. The combined effect and load-bearing capacity of these three components work in concert to keep the fault’s hanging wall stable. As depicted in Fig. 16a, with a slot height of 8 m and the roof capable of filling the goaf upon collapse, the gangue’s bulking factor is substantial, indicating wide spaces between individual particles and minimal compaction in the accumulated material. Under these circumstances, the gangue pile’s ability to bear the weight of the overlying strata is considerably diminished. Consequently, as illustrated in Fig. 17, when the overlying strata come under pressure, a greater share of stress gets transferred to the coal mass in the fault’s hanging wall and the fault plane itself, leading to heightened stress concentrations in the surrounding rock on both flanks of the working face. As shown in Fig. 16b, to improve this situation, increasing the collapse height of the gangue can enhance the compactness of the gangue accumulation and reduce the porosity. At the same time, reducing the particle size of the gangue can further improve the compactness of the accumulation. These measures can effectively reduce the porosity of the gangue accumulation and enhance its load-bearing capacity. When the load-bearing capacity of the gangue accumulation is improved, the stress transmitted from the overlying strata to the surrounding rock ahead of the working face decreases. At the same time, the stress transfer from the fault’s hanging wall to the footwall formation and the disruption to the roadway passage are mitigated, which in turn lessens the extent of stress buildup and deformation in the surrounding rock of the underground passage.
Fig. 16.
Distribution of abutment pressure in fault hanging wall rock mass.
Fig. 17.
Stress map of fault surface.
Control scheme
Numerical simulation findings indicate that fault displacements measure 0.182 m, 0.143 m, and 0.142 m for slot heights of 8 m, 10 m, and 12 m respectively. The data clearly shows that increasing the slot height from 8 to 10 m yields the most substantial reduction in fault sliding displacement. Considering the factors of fault control effect and slotting cost comprehensively, the slot height in the fault structural zone is optimized to 10 m (
). Combined with the previous similar engineering geological conditions, the slot angle is finally determined to be 15°. Based on the angular calculations, a slot height of 10 m results in a slot hole length of 10.35 m. To streamline construction operations, the depth of the slot hole has been standardized at 10 m.
Based on the results of theoretical analysis and previous laboratory experiments33, under normal circumstances, the smaller the particle size of gangue formed after crushing, the smaller its bulking coefficient, and the higher its load-bearing capacity. To ensure the full collapse of gangue within the slot height range and reduce the particle size of the collapsed gangue (thereby decreasing the bulking coefficient of the collapsed gangue), loose blasting should be carried out within the slot height range. As shown in Fig. 18, on the basis of slot blasting, three loose blasting holes are arranged downward along the roadway shoulder angle. The vertical distances from the roadway shoulder angle are 1000 mm, 1500 mm, and 2000 mm respectively; the angles with the vertical line are 35°, 45°, and 55° respectively; and the lengths are 13 me, 16 m, and 21 m respectively. Every three loose holes form a cross-section, and the distance between adjacent cross-sections is 1000 mm.
Fig. 18.
Blasting scheme and effect.
Engineering application
It’s clear from Fig. 19 that as the gap between the working face and the fault narrows, the hydraulic supports at the working face begin to provide more support. After the working face hydraulic supports cross the ZF44 fault, their supporting force decreases rapidly, showing an obvious stress barrier effect, which is consistent with the analysis results mentioned earlier. As the mining face advances into the coal deposit within the hanging wall, the barrier effect causes the vertical stress on both the supports at the working face and the coal ahead of these supports to be primarily derived from the vertical load transferred downward from the rock strata above the accumulated waste material in the hanging wall to the footwall. Therefore, the supporting force of the working face hydraulic supports decreases to a certain extent. During the entire fault-crossing period, the average supporting force of the 140# hydraulic support on the side where no slotting holes or loosening holes are applied is 30.1 MPa, and the average supporting force of the 5# hydraulic support on the side close to the retained roadway is 25.5 MPa. It can be seen that after adopting optimization and adjustment measures such as slotting and loosening holes, the supporting force of the working face hydraulic supports has decreased by 15.3%.
Fig. 19.
Supporting force of the working face hydraulic support.
As illustrated in Fig. 20, the gangue piled up on the fractured gangue wall side experiences progressive mutual compression under overlying strata pressure as the roadway–working face distance increases. Compressive failure initially occurs in the skeletal supporting sections of large gangue fragments, resulting in a reduction in the void ratio of the gangue accumulation and a corresponding gradual decrease in the tested bulking coefficient. As overlying strata pressure keeps increasing, the voids between gangue blocks are gradually compressed; consequently, the compactness of the gangue accumulation increases, and its load-bearing capacity is enhanced step by step. Upon the roadway at the fault zone being 200 m away from the working face, the impact of mining activities on the roadway diminishes progressively. Correspondingly, the load-bearing capacity of the broken rock rib stops varying and attains a steady state eventually, while the tested bulking coefficient exhibits no more fluctuations. Composed of the fragmented gangue wall, coal body, roof, and floor, the roadway exhibits stable and reliable performance, meeting the safety operation requirements of the mine. The roadway after stabilization is shown in Fig. 21. The bulking coefficient of the fault structural zone under conventional parameters is 1.39; after optimizing the blasting parameters, the bulking coefficient of the fault structural zone decreases to 1.33, a reduction of 4.3% compared with that before optimization. As demonstrated, increasing the slot length and supplementing loosening holes can diminish the particle size of collapsed gangue, elevate the height of the caving zone, and improve the density of the gangue accumulation. Compared with the expected target of reducing the bulking coefficient from 1.4 to 1.32, the final effect reaches 75% of the expected value, which indicates that the adopted method is practically effective.
Fig. 20.
Law of variation in bulking coefficient.
Fig. 21.
Final effect of roadway.
Conclusions
Aiming at the surrounding rock control problem of Roof Cutting and Retaining by Mining (RCRM) technology in fault structural zones, this study combines theoretical analysis, numerical simulation and field verification to conduct in-depth research on the strata deformation law and control technology of RCRM working face in cross-fault mining, and the key conclusions with theoretical innovation and engineering application value are as follows:
Based on the key stratum theory and voussoir beam theory, a three-stage mechanical model for RCRM cross-fault mining was innovatively established, which for the first time clarifies the dynamic evolution law of advance stress under the action of the fault stress barrier effect. The model reveals that the advance stress drops sharply after the working face mines out the coal seam in the fault hanging wall, and gradually rebounds to the normal geological stress state with the increase of the immediate roof load when the working face advances to the fault footwall and moves away from the fault.
FLAC3D and the Double-Yield model were adopted to simulate the compaction characteristics of caved gangue, and deviatoric stress was introduced as the evaluation index of surrounding rock stability. A comparative study of three slotting heights (8 m, 10 m, 12 m) showed that the larger the slotting height, the smaller the surrounding rock distortion energy and compressive deformation intensity; 10 m is the economically optimal slotting height. Compared with the 8 m slotting height, the peak deviatoric stress invariant of the coal mass in front of the working face is reduced by a maximum of 24.8 MPa2, the overlying strata displacement is reduced by 0.11 m, and the fault displacement is reduced by 0.04 m, while the control effect of the 12 m slotting height is limitedly improved.
A targeted surrounding rock control scheme of “10 m slotting + loose blasting” was proposed for the fault structural zone. Field verification results show that this scheme effectively improves the bearing capacity of gangue in the goaf, reduces the stress concentration of surrounding rock and the slip displacement of fault plane, and achieves remarkable control effects: the average support force of hydraulic supports is reduced by 15.3%, the gangue bulking coefficient is reduced by 4.3%, and the actual control effect reaches 75% of the expected value, which provides reliable technical support for the safe and efficient application of RCRM technology in fault-developed areas.
Author contributions
Yang Dongshan wrote the main manuscript and derived the mechanical models.Liang Bo guided the on-site engineering application, collected on-site monitoring data, and evaluated the safety of the surrounding rock. Kuai Xiaohui was mainly responsible for data analysis. Wang Hao Hao was in charge of numerical simulation. Yin Songyang was responsible for complex data plotting. Zhang Shukun was responsible for building the framework of the paper and enhancing its content.
Funding
This study was funded by the National Natural Science Foundation of China [Grant No. 52574168], Shandong Provincial Natural Science Foundation [Grant No. ZR2024ME257].
Data availability
The data that support the findings of this study are available from the corresponding author, [ZHANG Shukun], upon reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.He, M., Zhu, G. & Guo, Z. Longwall mining “cutting cantilever beam theory” and 110 mining method in China—The third mining science innovation. J. Rock Mech. Geotech. Eng.7, 483–492 (2015). [Google Scholar]
- 2.Wang, Q. et al. Research on an automatic roadway formation method in deep mining areas by roof cutting with high-strength bolt-grouting. Int. J. Rock Mech. Min. Sci.128, 104264 (2020). [Google Scholar]
- 3.Yang, D., Wang, H., Kuai, X. & Zhang, S. Deformation mechanism and control technology of retained roadway in qipanjing coal mine under the influence of fault structure. Min. Metall. Explor.10.1007/s42461-025-01387-x (2025). [Google Scholar]
- 4.Gao, J., Guo, Z. & Zhu, J. Model test on the ultimate failure mechanism of automatically formed roadway in weakly cemented soft rock. Tunn. Undergr. Space Technol.166, 106968 (2025). [Google Scholar]
- 5.Wang, Q. et al. Experimental study on the mechanism of pressure releasing control in deep coal mine roadways located in faulted zone. Geomech. Geophys. Geo-Energy Geo-Resour.8, 50 (2022). [Google Scholar]
- 6.Yang, D. et al. Geomechanical model test investigation of roof cutting and roadway maintenance technology under fault structures. Min. Metall. Explor.41, 889–908 (2024). [Google Scholar]
- 7.Li, Z. et al. Study on the influence of the fault dip angle on the stress evolution and slip risk of normal faults in mining. Bull. Eng. Geol. Environ.80, 3537–3551 (2021). [Google Scholar]
- 8.Ma, K., Yuan, F., Zhuang, D., Li, Q. & Wang, Z. Study on rules of fault stress variation based on microseismic monitoring and numerical simulation at the working face in the Dongjiahe coal mine. Shock. Vib.2019, 7042934 (2019). [Google Scholar]
- 9.Wang, H. et al. Mechanical model for the calculation of stress distribution on fault surface during the underground coal seam mining. Int. J. Rock Mech. Min. Sci.144, 104765 (2021). [Google Scholar]
- 10.Wang, H. et al. Investigation of fault displacement evolution during extraction in longwall panel in an underground coal mine. Rock Mech. Rock Eng.53, 1809–1826 (2020). [Google Scholar]
- 11.Zhang, Q. et al. Experimental studies and failure mechanisms of strain and fault-slip rockburst: A review. J. Cent. South Univ.31, 3741–3781 (2024). [Google Scholar]
- 12.Wei, C, Zhang, C. & Canbulat, I. Numerical analysis of fault-slip behaviour in longwall mining using linear slip weakening law. Tunn. Undergr. Space Technol.104, 103541 (2020). [Google Scholar]
- 13.Shan, R. et al. Research on the mechanism of asymmetric deformation and stability control of near-fault roadway under the influence of mining. Eng. Fail. Anal.127, 105492 (2021). [Google Scholar]
- 14.Zhang, X. & Chen, Y. Research on the type and occurrence mechanism and prevention ofcoal pillar rockbursts. Coal Sci. Technol.51, 1–11 (2023). [Google Scholar]
- 15.Vardar, O., Wei, C., Zhang, C. & Canbulat, I. Numerical investigation of impacts of geological faults on coal burst proneness during roadway excavation. Bull. Eng. Geol. Environ.81, 1–12 (2022). [Google Scholar]
- 16.Sun, X. et al. Instability mechanism of mining roadway passing through fault at different angles in kilometre-deep mine and control measures of roof cutting and NPR cables. J. Mt. Sci.21, 236–251 (2024). [Google Scholar]
- 17.Xie, F., Zhang, Z., He, W. & Chen, D. Research on ground directional grouting reinforcement technology and coupling control of fault fracture zone through rock roadway. Appl. Sci.10.3390/app15105647 (2025). [Google Scholar]
- 18.Huang, Y., Zhao, A., Zhang, T. & Guo, W. Plastic failure zone characteristics and stability control technology of roadway in the fault area under non-uniformly high geostress: A case study from Yuandian Coal Mine in Northern Anhui Province, China. Open Geosci.12, 406–424 (2020). [Google Scholar]
- 19.Zeng, C., Zhou, Y., Zhang, L., Mao, D. & Bai, K. Study on overburden failure law and surrounding rock deformation control technology of mining through fault. PLoS ONE17, e0262243 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pan, W., Li, H., Hua, X., Liu, B. & Li, C. Research on grouting reinforcement technology of fault crossing roadway in fully mechanized mining face with large dip angle. Bull. Eng. Geol. Environ.83, 216 (2024). [Google Scholar]
- 21.Zhang, S., Wang, L., Yang, H. & Lu, L. Construction and numerical simulation research of functional supporting in deep roadways. J. Min. Strata Control Eng.5, 013012–013012 (2023). [Google Scholar]
- 22.Fu, Q. et al. Mechanism and surrounding rock control of gob-side entry formation passing through normal fault: A case study. Environ. Earth Sci.82, 603 (2023). [Google Scholar]
- 23.Liu, X. et al. Dynamic instability mechanism and stress staged evolution of gob-side entry during cross-fault mining for thick coal seam. Energy Sci. Eng.12, 5602–5616 (2024). [Google Scholar]
- 24.Wang, J., Xu, J., Yang, S. & Wang, Z. Development of strata movement and its control in underground mining: In memory of 40 years of Voussoir Beam Theory proposed by Academician Minggao Qian. Coal Sci. Technol.51, 80–94 (2023). [Google Scholar]
- 25.Gao, Y., Wang, Y., Yang, J., Zhang, X. & He, M. Meso-and macroeffects of roof split blasting on the stability of gateroad surroundings in an innovative nonpillar mining method. Tunn. Undergr. Space Technol.90, 99–118 (2019). [Google Scholar]
- 26.Yavuz, H. An estimation method for cover pressure re-establishment distance and pressure distribution in the goaf of longwall coal mines. Int. J. Rock Mech. Min. Sci.41, 193–205 (2004). [Google Scholar]
- 27.Li, W., Bai, J., Peng, S., Wang, X. & Xu, Y. Numerical modeling for yield pillar design: A case study. Rock Mech. Rock Eng.48, 305–318 (2015). [Google Scholar]
- 28.Wang, M., Bai, J., Li, W., Wang, X. & Cao, S. Failure mechanism and control of deep gob-side entry. Arab. J. Geosci.8, 9117–9131 (2015). [Google Scholar]
- 29.Li, Z. Study on the Stress Deviator Invariant Distribution of Gob Floor and Its Application (China University of Mining and Technology (Beijing), 2022). [Google Scholar]
- 30.Shan, R. et al. Study on the distribution characteristics of stress deviator in the surrounding rock when mining closely spaced coal seams. Environ. Earth Sci.80, 602 (2021). [Google Scholar]
- 31.Zhao, H. et al. Distribution characteristics of deviatoric stress field and failure law of roadway surrounding rock under non hydrostatic pressure. J. China Coal Soc.46, 370–381 (2021). [Google Scholar]
- 32.Hou, G. et al. Influence of deviatoric stress on the deformation and damage evolution of surrounding rock under unloading conditions. Shock. Vib.2020, 3158209 (2020). [Google Scholar]
- 33.Yu, T. et al. Compressive deformations and energy dissipation characteristics of broken rock masses under lateral confinement. Coal Geol. Explor.53, 59 (2025). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, [ZHANG Shukun], upon reasonable request.