Surrounding rock stability and engineering practice of the 52607 working face approaching a fault-protected coal pillar Authors: Zenghui Liu, Minjun Chen, Feiyue Liu Published: 2026 ABSTRACT To clarify the instability behavior of surrounding rock and its control effect when a heading face approaches a fault-protected coal pillar, a combined approach involving in-situ stress measurement, theoretical analysis, three-dimensional numerical simulation, and field monitoring was adopted to investigate the displacement response, plastic-zone evolution, and applicability of the existing support system under different protective coal pillar widths. The results show that the in-situ stress field in the study area is a type-II stress field dominated by horizontal tectonic stress. The coupled effect of fault-induced structural weakening and stress concentration in the protective coal pillar significantly aggravates the deformation and plastic failure of the surrounding rock, and the surrounding rock stability is highly sensitive to the protective coal pillar width. Under fault dip angles ranging from 54° to 70°, representative protective coal pillar widths of 10, 20, 30, and 40 m were selected for theoretical analysis and numerical simulation. The results show that the displacement response and plastic-zone development of the surrounding rock exhibit obvious zonal characteristics with decreasing protective coal pillar width. In particular, when the protective coal pillar width decreases to 10–20 m, the fault-induced disturbance has a more significant influence on roof deformation, rib convergence, and plastic-zone expansion; when the protective coal pillar width increases to 30–40 m, the fault disturbance effect weakens markedly. Numerical simulation results further indicate that when the width of the protective coal pillar is less than 20 m, the plastic zone of the coal pillar expands significantly and its stability decreases markedly. Field monitoring results are generally consistent with the numerical simulation results, indicating that the existing bolt-cable support system can effectively control the deformation of surrounding rock in near-fault roadways. The results can provide a basis for stability analysis and support design of roadways in fault-protected coal pillar zones. Introduction With the continuous increase in mining depth and the growing complexity of geological structures, the stability of surrounding rock during roadway excavation has become an increasingly prominent issue. Kang1 systematically reviewed the development of support technologies for deep and complex roadways in underground coal mines, and pointed out that high in-situ stress and complex structural conditions are the key factors restricting roadway stability control. Xie et al.2 analyzed the instability and support response of surrounding rock in deep coal roadways from the perspective of surrounding-rock control mechanisms, and concluded that the evolution of the bearing structure of the surrounding rock plays a decisive role in roadway stability. Shi et al.3 investigated support technologies for deep roadways based on engineering cases, further demonstrating the difficulty of surrounding-rock control under complex stress environments. These studies indicate that under deep high-stress and complex geological conditions, roadway surrounding rock often exhibits large deformation, strong asymmetric failure, and difficult support control. With respect to fault influence, Wang et al.4 analyzed the instability mechanism and support technology of roadways in fault fracture zones through an engineering case study, and pointed out that the weakening effect of faults can significantly reduce the bearing capacity of surrounding rock. Zhang et al.5 investigated the failure characteristics of surrounding rock near a fault under mining influence by means of physical model tests, revealing the significant effect of faults on surrounding-rock structure and stress transfer paths. Kang et al.6 analyzed the stability of deep coal mine roadways crossing fault zones using numerical simulation, and demonstrated that the extent and mechanical parameters of the fault fracture zone play an important controlling role in surrounding-rock instability. Xie et al.7 further studied the failure mechanism and hierarchical continuous support technology of roadways in unidirectional fault-affected zones, indicating that the plastic failure of surrounding rock and the support response under fault influence exhibit obvious zonal characteristics. These findings suggest that fault fracture zones are generally characterized by broken rock mass, developed fractures, weakened mechanical properties, and non-uniform stress distribution, which can easily lead to local failure of the surrounding-rock bearing structure and induce asymmetric deformation and local damage. Meanwhile, considerable attention has also been paid to the influence of coal pillars, especially residual coal pillars in close-distance coal seams, on roadway stability. Liu et al.8 investigated the instability mechanism and control measures of a roadway under retained coal pillar conditions in close coal seams, and found that stress concentration and plastic failure of the surrounding rock are particularly significant in the strong influence zone of the coal pillar. Wang et al.9 analyzed the layout and mechanism of a lower return airway under the influence of a residual coal pillar, and pointed out that the pillar width and the relative position between the roadway and the coal pillar are key factors affecting roadway stability. Wu et al.10 revealed the evolution law of the stress field and damage in coal and rock masses in the residual pillar area after top-slice mining of a thick coal seam. Wang et al.11 further studied the deformation and failure mechanism of the lower roadway in a close-range residual coal pillar area. Wang et al.12 analyzed the stability control of goaf-driven roadways under interchange remaining coal pillar conditions in close-distance coal seams, showing that complex mining-induced stress can significantly intensify the asymmetric deformation of roadway surrounding rock. Xia et al.13 investigated the stability of roadways with small coal pillars under repeated mining by numerical simulation, and revealed that mining disturbance promotes the expansion of the plastic zone in both the coal pillar and surrounding rock. These studies indicate that coal pillars can significantly alter the stress distribution, plastic-zone range, and deformation pattern of adjacent roadways, and their influence range and action law have become an important topic in the study of roadway surrounding-rock control. In recent years, with the increasing prominence of roadway engineering problems under complex mining and structural conditions, more attention has been paid to surrounding-rock deformation, strong mine pressure control, hard roof weakening, and support optimization under complex stress environments. Fan et al.14 analyzed the mechanical response of a semi-permeable borehole under non-hydrostatic in-situ stresses within transversely isotropic media, indicating that the magnitude, direction, and anisotropy of in-situ stress have important effects on the mechanical response around underground openings. Wang et al.15 studied the deformation and stress evolution of surrounding rock in pre-driven longwall recovery rooms at the end of mining, revealing the stage-dependent characteristics of surrounding-rock response under mining influence. Shang et al.16 analyzed the instability characteristics and control technology of roadway surrounding rock under repeated mining in close-distance coal seams. In terms of mine pressure control and hard-roof weakening, Gao et al.17 investigated the effects of ground fracturing with a horizontal fracture plane on rock breakage characteristics and mine pressure control, and showed that fracture-induced structural weakening can effectively regulate hard-roof behavior and reduce roadway deformation. Xiang et al.18 investigated the failure characteristics and stability control of bolt support in thick-coal-seam roadways. Gao et al.19 further studied ground fracturing of multi-strata for strong ground pressure control in extra-thick coal seams with hard roofs through numerical simulation and engineering practice, demonstrating the importance of controlling hard-roof-induced stress concentration. Rong et al.20 analyzed the surrounding-rock control of roadways crossing collapse columns in strong mine pressure working faces and verified a combined support scheme through theoretical analysis, numerical simulation, and field monitoring. Xie et al.21 investigated the deformation mechanism and control technology of surrounding rock in “three-soft” coal roadways under high horizontal stress. Yao et al.22 analyzed the instability mechanism and surrounding-rock control technology of roadways subjected to short-distance dynamic loading. These studies provide important references for the stability analysis and support design of roadways under complex geological structures, strong mine pressure, and high-stress conditions. Overall, existing studies have mainly focused on roadways crossing fault fracture zones, roadways affected by residual coal pillars, roadways in close-distance coal seam mining, roadways under dynamic pressure conditions, or surrounding-rock control under hard-roof and collapse-column influences. However, relatively few studies have addressed the particular stage in which a heading face gradually approaches a fault-protected coal pillar. In particular, under the coupled action of fault-induced structural weakening and stress concentration in the protective coal pillar, the evolution of surrounding-rock stability with changes in the protective coal pillar width remains unclear. At the same time, under this specific condition, the control effect and engineering applicability of the existing bolt-cable support scheme still require further evaluation. Therefore, taking the 52607 working face of Chengjiagou Coal Mine as the engineering background, this study investigates the deformation and failure behavior of surrounding rock during the process in which the heading face approaches a fault-protected coal pillar. In-situ stress measurement, theoretical analysis, three-dimensional numerical simulation, and field monitoring were combined to systematically analyze the displacement response, plastic-zone development, and deformation characteristics of the surrounding rock under different protective coal pillar widths. The main influence range of the fault was further identified, and the control effect and applicability of the existing bolt-cable support system were evaluated. The results provide a reference for surrounding-rock stability analysis and support design in fault-protected coal pillar zones under similar engineering conditions. Project overview Engineering background Chengjiagou Coal Mine is located in Yangfangkou Town, Ningwu County, Xinzhou City, Shanxi Province, China, and is a medium-sized mine affiliated with Xuangang Coal and Electricity Co., Ltd. Working face 52607 is located in the southern part of the mine field. The ground elevation in the panel area ranges from 1311 to 1478 m, while the floor elevation of No. 5 coal seam ranges from 1080 to 1300 m, corresponding to a burial depth of 178–230 m. The working face has a strike length of approximately 1800 m and a dip width of about 1000 m, with a designed service life of approximately 8 months. In terms of spatial relationship, it is 309 m away from the western boundary of the mine field, 30 m away from the return airway of working face 52606 on the south side, 180 m away from working face 52609 on the east side, and 201 m away from Panel 512 on the north side. Working face 52607 mainly extracts the No. 5 coal seam, which has an average thickness of 13.93 m and relatively stable occurrence conditions. The immediate roof is a 10.3 m thick fine sandstone interbedded with marl and thin coal seams; the main roof is massive and dense mudstone with poorly developed joints; the immediate floor is mudstone interbedded with No. 6 coal seam and is relatively stable; and the lower floor is medium- to coarse-grained sandstone with a thickness of 2.18–8.04 m and an average thickness of about 4.0 m, representing a hard and stable floor. The comprehensive stratigraphic column near the No. 5 coal seam is shown in Fig. 1. Fig. 1Borehole lithologic column. Borehole lithologic column. Fault characteristics Working face 52607 is mainly affected by the TF-3 fault. The TF-3 fault is located on the northwestern side of the mine field. It starts from Houhegou and extends eastward approximately along the W–E direction, continuing through Yaozigou. Near the valley area of Chengjiagou Coal Mine, the fault strike turns to the northeast and finally disappears into the Huihe River, with a total length of about 3.8 km. The fault is clearly exposed at the surface and has been confirmed by surface stratum correlation as well as drilling data from boreholes Gou-8, Gou-10, and No. 202. Figure 2 shows the relative position of working face 52607 and the fault.The TF-3 fault is a normal fault dipping to the southeast, with a dip angle ranging from 54° to 70°. The throw is about 35–38 m in the western segment and approximately 75 m in the central segment, showing an overall downthrown southern block. Controlled by this fault, the 52607 return airway is located in the hanging wall of the TF-3 fault. As the roadway advances continuously, its driving direction gradually approaches the fault strike, resulting in an increasingly enhanced weakening effect of the fault structure on the surrounding rock. This constitutes the key geological background for the progressively intensified surrounding-rock stability problems during the excavation of working face 52607. Fig. 2Relative position of working face 52607 and the fault. Relative position of working face 52607 and the fault. Theoretical analysis of in-situ stress roof deformation and plastic zone in the fault-protected coal pillar zone In-situ stress measurement and analysis of stress field characteristics To determine the in-situ stress state of the study area, Chengjiagou Coal Mine adopted an in-situ stress testing method based on the acoustic emission Kaiser effect to back-calculate the three-dimensional principal stresses. First, the original rock samples collected in the field were directionally prepared along six orientations, as shown in Fig. 3a. According to the azimuth and dip information marked on the top of the rock sample, the sample was restored to its original in-situ orientation. A nearly horizontal plane was selected, and the true north direction, denoted as N, was determined. Taking N as the X-direction, a three-dimensional coordinate system composed of the X, Y, and Z axes was established. In the XOY, XOZ, and YOZ planes, directions making an angle of 45° with the two coordinate axes were then determined, respectively. On this basis, cylindrical specimens were drilled along the six predetermined directions, with a specimen size of 25 mm × 50 mm. In this test, one measuring point was selected, and three core specimens were prepared in each of the six directions, yielding a total of 18 specimens for subsequent acoustic emission Kaiser effect testing and principal stress calculation. The rock cores used in the experiment are shown in Fig. 3b. The loading process in this experiment was carried out using a rock mechanics testing machine, as shown in Fig. 3c. The acoustic emission signals were collected using an acoustic emission testing system manufactured by PAC (Physical Acoustics Corporation, USA), as shown in Fig. 3c. All acoustic emission signals generated during the entire deformation process of the specimen under loading were recorded. Fig. 3Directional specimen preparation and rock core specimens for the Kaiser effect test. (a) Directional specimen preparation, (b) Rock core specimens used in the experiment, (c) PAC acoustic emission testing system, (left) and TYI-600 computer-controlled electro-hydraulic servo rock mechanics testing machine, (right) Directional specimen preparation and rock core specimens for the Kaiser effect test. (a) Directional specimen preparation, (b) Rock core specimens used in the experiment, (c) PAC acoustic emission testing system, (left) and TYI-600 computer-controlled electro-hydraulic servo rock mechanics testing machine, (right) After specimen preparation, the 18 specimens obtained from the six directions at the same measuring point were subjected to acoustic emission Kaiser effect tests. One representative result from each direction was selected as the final test result for that direction, and all tests were conducted under uniaxial loading conditions. The calculated in-situ stresses and the orientation of the maximum horizontal principal stress at the measuring point located at the end of the track roadway in Panel 526 are listed in Table 1. Table 1Calculated in-situ stress results at the measuring point.Measuring pointBurial depth (m)σV (MPa)σH (MPa)σh (MPa)Orientation of maximum horizontal principal stress526071904.504.633.13E43.2°S Table 1Calculated in-situ stress results at the measuring point.Measuring pointBurial depth (m)σV (MPa)σH (MPa)σh (MPa)Orientation of maximum horizontal principal stress526071904.504.633.13E43.2°S Calculated in-situ stress results at the measuring point. The in-situ stress test results at the measuring point at the end of the track roadway in Panel 52607 of Chengjiagou Coal Mine show that the maximum horizontal principal stress is 4.63 MPa, the minimum horizontal principal stress is 3.13 MPa, and the vertical stress is 4.50 MPa. According to the magnitude relationship among the three principal stresses, the in-situ stress field can be classified into three types: Type I,σH > σh > σV;Type II, σH > σV >σh;and Type III,σV > σH > σh.Since the measured stress relationship at this point is σH > σV >σh,the stress field belongs to Type II, indicating that tectonic stress is greater than gravitational stress. Roof displacement analysis of a near-fault roadway In this study, the protective coal pillar width is defined as the shortest horizontal distance between the fault-side roadway wall and the fault plane, denoted by Wp. Therefore, the terms “protective coal pillar width” and “roadway–fault distance” refer to the same geometric parameter in the following analysis. To further reveal the influence of the fault on the stability of surrounding rock in a near-fault roadway, a theoretical analysis of the loading and displacement characteristics of the roadway roof was carried out based on the in-situ stress measurement results. According to the loading characteristics of the roof in a near-fault roadway, the roof can be simplified as a cantilever beam subjected to the load of the overlying strata, and the corresponding mechanical analysis model is shown in Fig. 4. Fig. 4Mechanical model of the roof in a near-fault roadway. Mechanical model of the roof in a near-fault roadway. The cantilever-beam model was adopted as a simplified theoretical model to describe the first-order effect of protective coal pillar width on roof deflection. It is mainly used to reveal the variation trend of roof deformation with coal pillar width, rather than to reproduce the complete three-dimensional stress redistribution and progressive failure process near the fault. In the simplified mechanical model, the fault-induced action was represented by an equivalent normal resultant force Fn acting on the fault-influenced boundary of the roof strata. According to the simplified plane-strain model of the fault-protected coal pillar, Fn can be expressed as:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F_n}=\gamma hb/[{W_p}\cos (\theta /2)]$$\end{document} where γ is the unit weight of the roof rock, h is the vertical component of the shortest distance from the roadway corner to the fault plane, b is the roadway width, Wp is the protective coal pillar width, and θ is the angle between the fault plane and the vertical direction. The direction of Fn is normal to the fault plane and points toward the protective coal pillar. Considering the complexity of the actual stress distribution along the fault plane, Fn was treated as an equivalent resultant force, and its line of action was assumed to pass through the centroid of the simplified fault-influenced boundary. Therefore, its vertical component Fy=Fnsinθ, together with the self-weight of the overlying strata, was used to describe the equivalent loading condition of the near-fault roof beam. On this basis, the stress components and deflection equation of the roof beam were derived according to elasticity theory. An Airy stress function was introduced as:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varphi =Ax{y^3}+Bxy$$\end{document} Accordingly, the stress components can be obtained as:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left\{ \begin{gathered} {\sigma _x}=\frac{{{\partial ^2}\varphi }}{{\partial {y^2}}}=6Axy \hfill \\ {\sigma _y}=\frac{{{\partial ^2}\varphi }}{{\partial {x^2}}}=0 \hfill \\ {\tau _{xy}}= - \frac{{{\partial ^2}\varphi }}{{\partial x\partial y}}= - B - 3A{y^2} \hfill \\ \end{gathered} \right.$$\end{document} The undetermined coefficients in the stress function were then solved by applying the mechanical boundary conditions of the simplified roof beam. Subsequently, the geometric equations and physical equations of elasticity were used to establish the relationship among stress, strain, and displacement, from which the roof deflection equation was obtained. Through this deflection equation, the influence of the protective coal pillar width on roadway roof deformation can be analyzed. Theoretical analysis indicates that, with increasing protective coal pillar width, the constraint effect of the relatively intact rock mass is enhanced, the fault-induced disturbance to the roof is weakened, and the roof deflection gradually decreases. The maximum roof deflection occurs at the free end of the cantilever beam, namely at x = 0 and y = 0. Therefore, the maximum deflection v satisfies a functional relationship with the protective coal pillar width Wp, as follows:4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$v(x,0)=\frac{{{a^3}\gamma hb\sin \theta }}{{3\cos \frac{\theta }{2}E\operatorname{I} {W_p}}}=\frac{{4{a^3}b\gamma h{{\cos }^3}\theta }}{{{{({W_p})}^4}E{{\sin }^2}\theta \cos \frac{\theta }{2}}}$$\end{document} where a is the roadway step height, b is the roadway width, γ is the unit weight of the roof rock, h is the vertical component of the shortest distance from the left bottom corner of the roadway floor to the fault plane, E is the elastic modulus of the coal seam, Wp is the protective coal pillar width, and θ is the angle between the fault plane and the vertical direction. According to the actual engineering conditions of Chengjiagou Coal Mine, the dip angle of the TF-3 fault was taken as approximately 65°, the unit weight of the roof rock was taken as 13kN/m3, the roadway section size was 4.2 m × 2.8 m, and the elastic modulus E of the No. 5 coal seam was taken as 2.85 GPa. The parameter h represents the vertical component of the shortest distance from the left bottom corner of the floor to the nearest point on the fault, and its expression can be further determined from the mechanical relationship of the near-fault roadway.The fault dip angle is denoted by α, while θ denotes the angle between the fault plane and the vertical direction. For the TF-3 fault, α is approximately 65°. According to the mechanical model shown in Fig. 5, it can be obtained that, θ = 90° - α :5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\tan \alpha =\tan 65^\circ =h/{W_p}$$\end{document} Fig. 5Mechanical relationship of the near-fault roadway. Mechanical relationship of the near-fault roadway. The protective coal pillar width was set as Wp = 40,30,20 m and 10 m, respectively, to maintain consistency with the numerical simulation cases. By substituting these values into the simplified roof-deflection equation, the variation trend of roof deformation under different protective coal pillar widths was obtained. The theoretical results indicate that the calculated roof deflection increases as the protective coal pillar width decreases, suggesting that the roof becomes more sensitive to fault-induced disturbance under narrow coal pillar conditions. It should be noted that the cantilever-beam model is a simplified theoretical model. The roof strata were idealized as an elastic beam, and the fault-induced action was represented by an equivalent normal resultant force. Therefore, the calculated deflection mainly reflects the relative variation trend of roof deformation with protective coal pillar width, rather than the exact field roof subsidence. Compared with the theoretical model, the numerical simulation further considers the three-dimensional constraint of the surrounding rock, stress redistribution after excavation, and the reinforcement effect of the bolt-cable support system.Therefore, the theoretical deflection results were used only for trend interpretation, whereas the numerical simulation and field monitoring results were used for quantitative evaluation of roadway deformation. As a result, the simulated roof displacement is smaller than the theoretical value, but the two methods show a consistent trend: the roof deformation increases as the protective coal pillar width decreases. These results indicate that, in the near-fault zone, the roadway roof is more strongly disturbed by the fault when the protective coal pillar width is small. As the protective coal pillar width increases, the roof loading condition is alleviated, the calculated deflection decreases, and the surrounding-rock stability improves. Therefore, the theoretical analysis provides a simplified reference for understanding the influence of protective coal pillar width on roof deformation and for subsequent analysis of plastic-zone development in the near-fault coal pillar. Plastic zone analysis of the near-fault coal pillar based on limit equilibrium theory To investigate the development law of the plastic zone in the surrounding rock on the fault-proximal side of the coal pillar, the plastic-zone width of the near-fault coal pillar was theoretically analyzed based on elastoplastic limit equilibrium theory and the Mohr–Coulomb strength criterion. In the theoretical model, the protective coal pillar between the roadway and the fault was simplified as an irregular trapezoidal body under plane-strain conditions. The coal mass adjacent to the roadway was divided into a plastic zone and an elastic bearing zone. The stress state in the plastic zone was assumed to satisfy the Mohr–Coulomb strength criterion, while the stress at the elastic–plastic interface was determined by the lateral confinement provided by the surrounding rock and the support system. According to the simplified plane-strain model, the fault-induced action was represented by the equivalent normal resultant force Fn, and the shear stress along the fault plane was introduced into the limit-equilibrium analysis. The derivation of the plastic-zone width was carried out as follows. First, the stress boundary conditions at the elastic–plastic interface were applied. Second, the differential equilibrium equations of a coal element in the plastic zone were established. Third, the Mohr–Coulomb criterion was used to describe the yield state of the coal mass in the plastic zone. Finally, by integrating the stress distribution from the roadway wall to the elastic–plastic interface, the expression for calculating the plastic-zone width on the fault-proximal side of the coal pillar can be obtained as follows:6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x_m}= - \frac{{\lambda a}}{{2\tan {\varphi _0}}}\ln \left(\frac{{\frac{{{C_0}}}{{\tan {\varphi _0}}}+\frac{{{P_x}}}{a}}}{{\frac{{{C_0}}}{{\tan {\varphi _0}}}+{\sigma _v}+\frac{{{\sigma _v}\cos \theta }}{{2\lambda {W_p}\cos \frac{\theta }{2}}}}}\right)$$\end{document} where a is the roadway step height, Wp is the protective coal pillar width, C₀ and φ₀ are the cohesion and internal friction angle of the coal seam, respectively, Px is the lateral confinement acting on the coal wall, λ is the lateral pressure coefficient, θ is the angle between the fault plane and the vertical direction, and σv is the vertical stress. In the theoretical calculation, Wp was set to 10, 20, 30, and 40 m to correspond to the numerical simulation cases. According to the engineering calculation report and the measured in-situ stress condition, the lateral pressure coefficient λ was taken as 1.0. The lateral confinement Px was taken as 1.23 MPa based on the support/confinement calculation of the roadway side. Considering that the dip angle of the TF-3 fault ranges from 54° to 70°, θ was taken as 20°, 25°, 30°, and 35°, respectively. The cohesion and internal friction angle of the No. 5 coal seam were taken as C₀ = 2.0 MPa and φ₀ = 32° according to the mechanical parameters listed in Table 2, and the vertical stress σv = 4.50 MPa was obtained from the in-situ stress measurement results listed in Table 1. These parameter values were used to calculate the plastic-zone width under different protective coal pillar widths and fault dip-angle conditions. The final calculation results are shown in Fig. 6. Fig. 6Plastic-zone range of the surrounding rock on the fault-proximal side under different fault dip angles and protective coal pillar widths. Plastic-zone range of the surrounding rock on the fault-proximal side under different fault dip angles and protective coal pillar widths. According to Eq. (6) and Fig. 6, the plastic-zone range of the surrounding rock on the fault-proximal side is jointly controlled by the roadway height, in-situ stress, protective coal pillar width, lateral confinement, and fault dip angle. The expression used in Eq. (6) was established based on the classical limit-equilibrium theory for roadway-side coal pillars and the Mohr–Coulomb strength criterion, with the fault-induced boundary effect further incorporated into the stress analysis. Therefore, it is suitable for evaluating the relative variation trend of the plastic-zone width under different near-fault coal pillar conditions. When the other parameters remain constant, the plastic-zone range gradually decreases with increasing protective coal pillar width. In particular, the plastic zone develops more significantly when the protective coal pillar width decreases to 10–20 m, whereas its variation becomes relatively gradual when the protective coal pillar width increases to 30–40 m. This indicates that the protective coal pillar width is the dominant factor controlling the plastic-zone development of the near-fault coal pillar, while the fault dip angle has a secondary influence within the investigated range. Numerical simulation analysis of surrounding rock stability in the fault-protected coal pillar zone Establishment of the numerical calculation model During roadway excavation, excavation-induced disturbance breaks the original equilibrium state of the in-situ stress field, resulting in stress redistribution in the surrounding rock and obvious local stress concentration. Under the complex stress environment near a fault, such stress redistribution is an important factor inducing surrounding-rock deformation and failure. Therefore, it is necessary to systematically analyze the displacement response and plastic-zone evolution of the surrounding rock during roadway excavation by means of numerical simulation. In this study, a three-dimensional numerical model was established using FLAC3D, and the Mohr–Coulomb yield criterion was adopted to determine rock mass failure. The Mohr–Coulomb model was selected because it can reasonably describe the shear failure characteristics of coal-measure rock masses and is widely used in roadway stability analysis. The main objective of the numerical simulation was to compare the relative deformation response and plastic-zone evolution of the surrounding rock under different protective coal pillar widths, rather than to reproduce the complete post-peak failure, strain-softening, and dilation processes of the fractured rock mass. Therefore, plastic flow and post-peak softening were not explicitly considered. This simplification may underestimate local deformation and the progressive failure range in the fault fracture zone, and this limitation is further discussed in Discussion. During model construction, a three-dimensional geometric model was first established, and an equivalent load was applied to the top of the model to represent the effect of the overlying strata that were not explicitly modeled. After the initial stress equilibrium was reached, step-by-step excavation of the 52607 return airway was carried out, while the installation of the support structure was simulated simultaneously. The calculation results were extracted after the model reached equilibrium again. Figure 7a shows the established numerical model, and Fig. 7b shows the relative position of the roadway and the fault. According to the engineering geological conditions of working face 52607, the model dimensions were set to 300 m × 105 m × 750 m. The roadway cross-section was 4.2 m × 2.8 m, the fault zone width was 7 m, and the fault dip angle was 65°. To improve the calculation accuracy in the key study area, the mesh around the roadway was refined to 0.5 m × 0.5 m, while a coarser rectangular mesh was used in the other regions. Both bolts and cables were simulated using built-in cable elements, and the support structure adopted in Chengjiagou Coal Mine is shown in Fig. 7c. The boundary conditions were set as follows: horizontal displacement was constrained on the lateral boundaries, displacement in all three directions was constrained at the bottom boundary, and an equivalent stress of the overlying strata was applied at the top boundary. To simulate the actual excavation process, the roadway was excavated in increments of 10 m, and the calculation was continued to convergence after each excavation step. Fig. 7Numerical simulation model. (a) Overall model, (b) Positional relationship between the roadway and the fault, (c) Support structure. Numerical simulation model. (a) Overall model, (b) Positional relationship between the roadway and the fault, (c) Support structure. Based on the above model and parameter settings, typical working conditions with protective coal pillar widths of 10, 20, 30, and 40 m were selected for numerical simulation. These four widths were determined according to the variation range of the roadway–fault distance during roadway advancement and the actual layout of the 52607 return airway. Specifically, Wp = 10 m represents an unfavorable narrow-pillar condition close to the fault, under which the roadway surrounding rock is expected to be strongly affected by fault-induced disturbance. Wp = 20 m corresponds to the monitored near-fault section and also represents the recommended minimum stable protective coal pillar width obtained from the preceding theoretical analysis. In contrast, Wp = 30–40 m represents the transition to a relatively weakly affected zone away from the fault, where the influence of the fault on roadway deformation gradually decreases. Therefore, the selected simulation cases can reflect the deformation response of the surrounding rock from the strongly affected near-fault zone to the relatively stable zone away from the fault. These four cases were consistent with the theoretical calculation conditions and were used to analyze the influence of protective coal pillar width on the deformation and plastic-zone evolution of the surrounding rock. The displacement variation law and plastic-zone development characteristics of the surrounding rock under different protective coal pillar widths were then analyzed, and the displacement responses under different fault dip angles were further examined to reveal the asymmetric deformation mechanism of the surrounding rock in near-fault roadways. The mechanical parameters used in the numerical model were determined based on laboratory tests of rock samples, the geological report of Chengjiagou Coal Mine, and the existing support-design data for the No. 5 coal seam. Considering that the numerical model represents the engineering-scale rock mass rather than intact laboratory specimens, the laboratory mechanical parameters were appropriately reduced according to the integrity of the rock mass, joint development, and engineering experience in similar coal roadways. The final parameters used in the numerical simulation are listed in Table 2. Table 2Rock mechanical parameters for numerical simulation.Rock LayerThickness(m)Density(kg/m3)Cohesion(MPa)Internal friction angle(°)Tensile strength(MPa)Poisson’s ratioElastic modulus(GPa)Compressive strength (MPa)Mudstone1026906.5401.50.393.216.3Fine sandstone10.3242018.437.317.70.243.1107.8No. 5 coal seam13.9313502.0321.60.380.765.6Mudstone1.6726906.5401.50.373.115.3No. 6 coal seam1.3513502.2371.70.350.765.8Mudstone826906.5401.50.393.216.1Medium- to coarse-grained sandstone424006.0544.72.50.204.262.4 Table 2Rock mechanical parameters for numerical simulation.Rock LayerThickness(m)Density(kg/m3)Cohesion(MPa)Internal friction angle(°)Tensile strength(MPa)Poisson’s ratioElastic modulus(GPa)Compressive strength (MPa)Mudstone1026906.5401.50.393.216.3Fine sandstone10.3242018.437.317.70.243.1107.8No. 5 coal seam13.9313502.0321.60.380.765.6Mudstone1.6726906.5401.50.373.115.3No. 6 coal seam1.3513502.2371.70.350.765.8Mudstone826906.5401.50.393.216.1Medium- to coarse-grained sandstone424006.0544.72.50.204.262.4 Rock mechanical parameters for numerical simulation. Displacement variation of surrounding rock under different protective coal pillar widths To analyze the influence of protective coal pillar width on the deformation characteristics of roadway surrounding rock, typical cases with protective coal pillar widths of 10, 20, 30, and 40 m were selected for FLAC3D numerical simulation. Based on the simulation results for each case, the displacement contours of the surrounding rock were extracted, and the variation laws of roof subsidence, rib convergence, and floor displacement were comparatively analyzed, so as to reveal the evolution characteristics of surrounding-rock deformation under near-fault conditions. Figure 8 presents the displacement contours under different protective coal pillar widths.The numerical simulation results show that, under different working conditions, the displacement of the surrounding rock is mainly concentrated around the roadway profile. Different degrees of deformation occur in the roof, ribs, and floor, among which roof subsidence and rib convergence are more pronounced, and the overall deformation exhibits a certain degree of asymmetry. As the protective coal pillar width increases, the overall displacement level of the surrounding rock gradually decreases, and the range of displacement concentration correspondingly shrinks, indicating that the fault has a significant controlling effect on roadway deformation, which is more pronounced in the near-fault zone. Fig. 8Displacement contours under different protective coal pillar widths. Displacement contours under different protective coal pillar widths. On this basis, the roof subsidence and floor displacement under each simulation case were further extracted, and the displacements of the fault-proximal rib and the coal-seam-side rib were comparatively analyzed, as shown in Fig. 9. The results indicate that, as the protective coal pillar width increases, the roof subsidence, floor displacement, and rib displacements all show an overall decreasing trend, suggesting that the fault proximity effect significantly intensifies the deformation of roadway surrounding rock. In terms of deformation characteristics in different parts of the roadway, the roof and ribs are more sensitive to changes in protective coal pillar width. In particular, the displacement of the fault-proximal rib is generally greater than that of the coal-seam-side rib, indicating that the surrounding rock on the fault side is more strongly disturbed and exhibits more obvious asymmetric deformation. By contrast, although the floor displacement also varies with protective coal pillar width, its overall variation amplitude is relatively small. Comprehensive analysis of the simulation results shows that, as the roadway gradually approaches the fault, surrounding-rock deformation becomes significantly aggravated; whereas when the roadway moves away from the fault, the displacement gradually decreases and tends to stabilize. This indicates that the protective coal pillar width is an important controlling factor governing the deformation characteristics of surrounding rock in near-fault roadways. Fig. 9Displacement curves obtained from the numerical simulation. (a) Roof and floor, (b) Two ribs. Displacement curves obtained from the numerical simulation. (a) Roof and floor, (b) Two ribs. Evolution of the plastic zone under different protective coal pillar widths The development characteristics of the coal pillar plastic zone under different protective coal pillar widths show that, as the roadway gradually approaches the fault, the plastic zones on both sides of the coal pillar expand progressively, and the plastic zone on the fault-proximal side develops more significantly, Fig. 10 shows the plastic-zone contours under different protective coal pillar widths. When the coal pillar width is 40 m and 30 m, respectively, although the coal pillar is already affected by the fault, the overall plastic-zone range remains relatively small. In this case, the maximum development depths of the plastic zone on the coal-seam side are 1.1 m and 1.2 m, respectively, while those on the fault-proximal side are 1.4 m and 1.5 m, respectively. Because a relatively wide elastic core is still preserved inside the coal pillar, the overall deformation of the coal pillar is small, the surrounding rock remains relatively stable, and the coal pillar as a whole is in a stable state. When the coal pillar width decreases to 20 m, the plastic zones on both sides of the coal pillar expand further. The maximum development depth of the plastic zone on the coal-seam side increases to 1.7 m, while that on the fault-proximal side increases to 2.5 m. At this stage, the elastic core inside the coal pillar is obviously reduced, and the bearing capacity of the surrounding rock declines to some extent; however, the coal pillar as a whole can still remain basically stable. When the coal pillar width is further reduced to 10 m, the plastic-zone range of the coal pillar expands significantly. The maximum development range of the plastic zone on the fault-proximal side increases to about 3.0 m, while that on the coal-seam side is about 2.5 m. The plastic zones on both sides of the coal pillar then occupy a relatively large range, and the elastic core is further reduced, indicating that the bearing structure of the coal pillar is significantly weakened, accompanied by a certain degree of deformation and a marked reduction in stability. Comprehensive analysis indicates that, with decreasing coal pillar width, the plastic zones on both sides of the coal pillar continue to expand, and the plastic zone on the fault-proximal side is always larger than that on the coal-seam side, indicating that the weakening effect of the fault on coal pillar stability is clearly asymmetric. When the coal pillar width is relatively large, the interior of the coal pillar still retains strong bearing capacity. However, when the coal pillar width decreases to a certain extent, the plastic zone expands rapidly, the elastic core is significantly reduced, and the stability of the coal pillar correspondingly declines. These numerical simulation results are basically consistent with the previous theoretical analysis, further verifying that the plastic zone of the coal pillar in the near-fault area exhibits obvious asymmetric expansion. Fig. 10Plastic-zone contours under different protective coal pillar widths. Plastic-zone contours under different protective coal pillar widths. Displacement response of surrounding rock under different fault dip angles To investigate the influence of fault dip angle on the deformation characteristics of roadway surrounding rock, it is first necessary to clarify why Wp = 20 m was selected as the representative protective coal pillar width. According to the preceding theoretical analysis and numerical simulation results, when the protective coal pillar width decreases to 10 m, the plastic zone expands significantly and the bearing capacity of the coal pillar is markedly weakened. When the protective coal pillar width increases to 30–40 m, the fault-induced disturbance is obviously reduced, and the roadway deformation tends to be relatively stable. In contrast, Wp = 20 m represents a critical near-fault condition under which the coal pillar can still remain basically stable while the influence of the fault is still relatively significant. Therefore, Wp = 20 m was selected as the recommended minimum stable protective coal pillar width and used as a representative engineering condition for analyzing the influence of fault dip angle. It should be emphasized that the selection of Wp = 20 m does not mean that it is a universally optimal coal pillar width under all geological and mining conditions. The purpose of the fault dip-angle analysis is to evaluate the sensitivity of roadway deformation to fault dip angle under this critical representative condition. Therefore, the conclusions obtained in this section are mainly applicable to the Wp = 20 m condition of the 52607 working face. For other protective coal pillar widths, the influence of fault dip angle may vary, and further comparative simulations are still required before broader generalization. Based on this representative condition, FLAC3D numerical simulations were conducted for four typical fault dip angles of 55°, 60°, 65°, and 70°, while keeping the other parameters unchanged. The displacement distributions of the surrounding rock under different dip-angle conditions were comparatively analyzed, and the corresponding displacement contour plots are presented in Fig. 11. The numerical simulation results indicate that, under different fault dip angles, the displacement of the surrounding rock is mainly concentrated around the roadway profile. Roof subsidence and rib convergence are relatively pronounced, and the overall deformation exhibits a certain degree of asymmetry. As the fault dip angle increases from 55° to 70°, the roof displacement decreases from 11.5 cm to 9.0 cm, the displacement of the fault-proximal rib decreases from 10.4 cm to 9.4 cm, and the displacement of the coal-seam-side rib decreases from 7.4 cm to 6.1 cm. Overall, the displacements of all parts of the roadway surrounding rock show a decreasing trend with increasing fault dip angle. Among the different deformation components, roof displacement shows the most obvious variation, decreasing by 21.7% from 55° to 70°. The displacement of the fault-proximal rib decreases by 9.6%, while that of the coal-seam-side rib decreases by 17.6%. These results indicate that the roof and ribs are affected by changes in fault dip angle to different degrees. However, under all dip-angle conditions, the displacement of the fault-proximal rib remains greater than that of the coal-seam-side rib. This suggests that the surrounding rock on the fault side is more strongly affected by structural disturbance, and the roadway deformation still exhibits obvious asymmetric characteristics. Comprehensive analysis shows that, under the representative condition of Wp = 20 m, variations in fault dip angle have a certain influence on the displacement distribution of the surrounding rock. However, compared with the deformation differences caused by changes in protective coal pillar width, the overall influence of fault dip angle is relatively limited. Therefore, for the 52607 working face and within the investigated dip-angle range, the protective coal pillar width should be regarded as the primary factor controlling surrounding-rock deformation, while the fault dip angle can be considered a secondary influencing factor under the Wp = 20 m condition. This conclusion should not be directly extended to other protective coal pillar widths without further verification. Fig. 11Displacement contours under different fault dip angles. Displacement contours under different fault dip angles. Analysis of surrounding rock control measures and evaluation of engineering practice effect Support parameters and support form of the test section To carry out field verification of the surrounding-rock control effect in the fault-protected coal pillar zone, the section located 50–70 m ahead of measuring point Y12 was selected as the test section. The test section adopted a combined bolt-cable support system. For roof support, Φ20 mm × 2500 mm left-handed rebar bolts without longitudinal ribs were used, together with Z2360 and K2335 resin cartridges for extended anchorage. The bearing plate size was 110 mm × 110 mm × 10 mm, and the bolt spacing was 1000 mm × 1000 mm. The roof cables were Φ17.8 mm × 7300 mm steel strands, anchored with two Z2360 resin cartridges and two K2335 resin cartridges for extended anchorage. The cable plate size was 300 mm × 300 mm × 15 mm, and the cable spacing was 1450 mm × 2000 mm. A W-shaped steel strip with dimensions of 4180 mm × 220 mm × 2.75 mm was installed on the roof, and No. 10 diamond-shaped metal mesh was used for surface support; in roof-fall sections, one additional layer of square mesh and one layer of No. 10 diamond-shaped metal mesh were installed. For rib support, Φ20 mm × 2500 mm left-handed rebar bolts without longitudinal ribs were used. The anchorage method and plate specifications were the same as those of the roof bolts, and the bolt spacing was 1000 mm × 1000 mm. A W-shaped steel strip with dimensions of 400 mm × 220 mm × 2.75 mm was installed on the ribs. When the ribs encountered broken surrounding rock, fault zones, or weak zones, additional W-shaped steel strips were installed for reinforced support. No. 10 diamond-shaped metal mesh was also used for rib support. In roof-fall sections, square mesh and diamond-shaped metal mesh were additionally installed and fixed with tie wires in a “three twists and one fastening” pattern. The material consumption per meter for the support system of the working face is listed in Table 3. In this way, a support system dominated by bolts and cables, and coordinated with W-shaped steel strips and metal mesh, was formed, providing the support basis for subsequent monitoring of surrounding-rock deformation and field evaluation of the support effect. Table 3Support material consumption per meter in working face 52607.MaterialSpecificationUnitQuantityRemarksRebar boltΦ20 × 2500 mm left-handed rebarSet11Resin cartridgeK2335Cartridge17Resin cartridgeZ2360Cartridge17Steel strandφ17.8 × 7300 mmPiece3Row spacing: 2 mPrestressed bolt plate110 × 110 × 10 mmPiece11Prestressed cable plate300 × 300 × 15 mmPiece3W-shaped steel strip4180 × 220 × 2.75 mmPiece1 Table 3Support material consumption per meter in working face 52607.MaterialSpecificationUnitQuantityRemarksRebar boltΦ20 × 2500 mm left-handed rebarSet11Resin cartridgeK2335Cartridge17Resin cartridgeZ2360Cartridge17Steel strandφ17.8 × 7300 mmPiece3Row spacing: 2 mPrestressed bolt plate110 × 110 × 10 mmPiece11Prestressed cable plate300 × 300 × 15 mmPiece3W-shaped steel strip4180 × 220 × 2.75 mmPiece1 Support material consumption per meter in working face 52607. Field monitoring scheme and monitoring point layout Combined with the previous theoretical analysis and numerical simulation results, it can be concluded that under fault influence, a reduction in the width of the protective coal pillar will significantly aggravate surrounding-rock deformation and the expansion of the plastic zone. Comprehensive analysis indicates that the width of the protective coal pillar should not be less than 20 m. To further verify the control effect of the existing support conditions on the roadway surrounding rock, field monitoring of surrounding-rock deformation was carried out during roadway reinforcement construction in Chengjiagou Coal Mine. Figure 12a shows a field photograph of the monitoring layout. The field monitoring adopted the cross-measurement method. One surrounding-rock deformation monitoring station was arranged every 10 m along the roadway strike, and each monitoring group contained three monitoring points. The monitoring points were arranged near the roof and ribs of the roadway to record the deformation of the surrounding rock. Two measuring stations were arranged in each monitoring group, and the spacing between the two stations did not exceed the support distance of two rows of bolts. The deformation of the surrounding rock was dynamically monitored by measuring the change in horizontal distance between the left and right ribs and the change in vertical distance between the roof and floor. The layout of the monitoring points and the measurement directions are shown in Fig. 12b. Specifically, rib convergence was characterized by the reduction in horizontal spacing between the monitoring points on the two ribs, namely the change in horizontal distance between points C and D. Roof subsidence was characterized by the reduction in vertical spacing between the roof monitoring point and the floor reference point. In this study, the field evaluation focused on roof subsidence and rib convergence, and the monitoring results were expressed in terms of convergence and subsidence rather than total displacement. With the above monitoring arrangement, key deformation indices of the roadway surrounding rock, such as roof subsidence and rib convergence during excavation and support, could be continuously monitored, thus providing a field basis for the subsequent analysis of surrounding-rock deformation behavior and the evaluation of the control effect of the existing support scheme. Fig. 12Field monitoring arrangement. (a) Field layout, (b) Cross-measurement method. Field monitoring arrangement. (a) Field layout, (b) Cross-measurement method. To establish the correspondence between the field monitoring section and the numerical simulation cases, the spatial relationship between the monitoring cross-sections and the fault plane was further examined based on the roadway layout and geological survey data. In this study, the horizontal distance from the fault-proximal roadway wall to the fault plane was consistently defined as the protective coal pillar width, Wp. According to the surveyed position of the TF-3 fault and the layout of the test section, the three monitored cross-sections were all located in the near-fault area, where the protective coal pillar width was approximately 20–25 m. Although the protective coal pillar width at the monitoring cross-sections was not exactly identical to the idealized numerical simulation case, it was close to the representative condition of Wp = 20 m. In particular, Wp = 20 m represents the recommended minimum stable protective coal pillar width obtained from the preceding theoretical analysis and numerical simulation. Therefore, using the numerical simulation result for Wp = 20 m as the reference for comparison is reasonable and relatively conservative. It should be noted that the field monitoring results and numerical simulation results do not represent a strict point-to-point comparison at exactly the same roadway–fault distance. Instead, the comparison is intended to verify whether the numerical model can reasonably reflect the deformation magnitude and asymmetric deformation trend of the surrounding rock in the near-fault section with a protective coal pillar width of approximately 20–25 m. This treatment helps establish a clearer engineering correspondence between the field monitoring results and the numerical simulation analysis. Comparative analysis of field monitoring results and numerical simulation data According to the measured results, the maximum rib convergence at Cross-Sect. 1 was 16.1 cm, including 9.3 cm on the fault-proximal side and 6.8 cm on the coal-seam side, while the maximum roof subsidence was 8.2 cm. At Cross-Sect. 2, the maximum rib convergence was 14.7 cm, including 8.7 cm on the fault-proximal side and 6.0 cm on the coal-seam side, and the maximum roof subsidence was 9.0 cm. At Cross-Sect. 3, the maximum rib convergence was 15.8 cm, including 8.9 cm on the fault-proximal side and 6.9 cm on the coal-seam side, while the maximum roof subsidence was 9.5 cm. Comprehensive analysis indicates that the overall surrounding-rock deformation at all monitored cross-sections was relatively small, and the roadway as a whole remained in a stable state. Meanwhile, the displacement on the fault-proximal side was consistently greater than that on the coal-seam side, indicating that the fault exerted a certain asymmetric influence on roadway surrounding-rock deformation, which is basically consistent with the previous theoretical analysis and numerical simulation results. Among the three monitored cross-sections, Cross-Sect. 1 showed the largest rib convergence and was therefore selected as a representative section to illustrate the displacement variations of the roof, floor, and two ribs, as shown in Fig. 13. Fig. 13Measured 24-h displacement variation at Cross-Sect. 1. Measured 24-h displacement variation at Cross-Sect. 1. Table 4Statistical comparison of field monitoring results and numerical simulation results.Cross-sectionMax convergence of Ribs (cm)Max. fault-proximal side (cm)Max. coal-seam side(cm)Roof subsidence (cm)Cross-Sect. 116.19.36.88.2Cross-Sect. 214.78.76.09.0Cross-Sect. 315.88.96.99.5Numerical simulation(Wp = 20 m)13.57.567.1 Table 4Statistical comparison of field monitoring results and numerical simulation results.Cross-sectionMax convergence of Ribs (cm)Max. fault-proximal side (cm)Max. coal-seam side(cm)Roof subsidence (cm)Cross-Sect. 116.19.36.88.2Cross-Sect. 214.78.76.09.0Cross-Sect. 315.88.96.99.5Numerical simulation(Wp = 20 m)13.57.567.1 Statistical comparison of field monitoring results and numerical simulation results. Table 4 was prepared based on a comparison between the field monitoring results and the numerical simulation results. A comparison between the field monitoring results and the numerical simulation results was conducted to verify the rationality of the established numerical model and to evaluate the control effect of the existing support scheme. As shown in Table 4, the maximum roof subsidence at Cross-Sects. 1–3 was 8.2 cm, 9.0 cm, and 9.5 cm, respectively, while the maximum rib convergence was 16.1 cm, 14.7 cm, and 15.8 cm, respectively. In comparison, the numerical simulation results for Wp = 20 m show that the maximum roof subsidence was 7.1 cm, and the maximum displacements of the fault-proximal rib and the coal-seam-side rib were 7.5 cm and 6.0 cm, respectively, corresponding to a maximum rib convergence of 13.5 cm. From the perspective of deformation magnitude, the monitored values were generally larger than the simulated values. The maximum monitored roof subsidence was approximately 15.5%–33.8% larger than the simulated value, and the maximum monitored rib convergence was approximately 8.9%–19.3% larger than the simulated value. Specifically, the largest roof subsidence occurred at Cross-Sect. 3, reaching 9.5 cm, which was about 33.8% higher than the simulated value of 7.1 cm. The largest rib convergence occurred at Cross-Sect. 1, reaching 16.1 cm, which was about 19.3% higher than the simulated value of 13.5 cm. These differences indicate that the numerical simulation slightly underestimated the actual deformation of the surrounding rock, but the deviation remained within an acceptable engineering range. The 24-h field monitoring curve further shows that the deformation of the roadway surrounding rock had obvious stage-dependent characteristics. In the early monitoring stage, roof subsidence and rib convergence increased relatively rapidly, which was mainly related to excavation disturbance, stress redistribution around the roadway, and the gradual mobilization of the support system. With the continuous action of the bolt-cable support, the growth rate of deformation gradually decreased, and the monitoring curves tended to become stable in the later stage. This indicates that the surrounding rock experienced a process from rapid deformation adjustment to gradual stabilization after excavation and support installation. Therefore, the field monitoring results not only provide the final deformation values, but also reflect the time-dependent stabilization process of the supported roadway. In terms of spatial distribution, both the field monitoring and numerical simulation results show that the deformation on the fault-proximal side was greater than that on the coal-seam side. For example, at Cross-Sect. 1, the maximum displacement of the fault-proximal rib was 9.3 cm, while that of the coal-seam-side rib was 6.8 cm. Similar asymmetric deformation characteristics were also observed at Cross-Sects. 2 and 3. This spatial deformation pattern is consistent with the numerical simulation results, indicating that the fault-proximal surrounding rock was more strongly affected by structural weakening and stress disturbance. Therefore, although there are quantitative differences between the monitored and simulated values, the two methods show good consistency in terms of deformation trend and spatial distribution. The difference between the field monitoring results and the numerical simulation results can be attributed to several factors. First, the numerical model assumes relatively idealized geological conditions, whereas the actual surrounding rock contains joints, fractures, local weak zones, and construction-induced disturbances, especially near the fault-protected coal pillar. Second, the numerical simulation reflects the overall deformation response after excavation and support installation under equilibrium conditions, while the field monitoring records the actual deformation process affected by excavation sequence, support installation timing, support–rock contact conditions, and local construction quality. Third, the Mohr–Coulomb model used in the simulation does not explicitly consider post-peak strain softening, plastic flow, or dilation of the fractured rock mass, which may lead to an underestimation of local deformation in the fault-affected zone. In addition, monitoring errors, reading conditions, and the limited number of monitored cross-sections may also contribute to the difference between the simulated and measured values. To further clarify the relationship among the theoretical calculation, numerical simulation, and field monitoring results, it should be emphasized that the roof-deflection values obtained from the simplified cantilever-beam model are generally larger than those obtained from numerical simulation and field monitoring. This difference is mainly caused by the simplifying assumptions of the theoretical model. In the theoretical analysis, the roof strata were idealized as an elastic cantilever beam, and the influence of the fault was represented by an equivalent load. However, the model does not fully consider the three-dimensional confinement of the surrounding rock, stress redistribution after roadway excavation, the continuity of the roof strata, or the reinforcement effect of the bolt-cable support system. Therefore, the theoretical calculation should not be regarded as a direct prediction of the actual roof subsidence in the field. Instead, it provides a conservative trend-based interpretation, showing that roof deformation tends to increase as the protective coal pillar width decreases. Overall, although certain quantitative differences exist among the theoretical calculation, numerical simulation, and field monitoring results, the three methods show a consistent deformation trend. As the roadway approaches the fault and the protective coal pillar width decreases, the deformation of the surrounding rock and the asymmetric response on the fault-proximal side become more pronounced. Under the existing bolt-cable support conditions, the monitored roof subsidence and rib convergence remained within a controllable range, and the field deformation curves tended to stabilize within the monitoring period. Therefore, the existing support system had a satisfactory control effect on the surrounding rock in the fault-protected coal pillar zone. These results verify the applicability of the numerical model and further demonstrate the effectiveness of the existing support scheme under the present engineering conditions. Discussion The results of this study are generally consistent with previous studies on roadway stability near faults and coal pillars. Existing research has shown that fault fracture zones can weaken the mechanical properties of surrounding rock, change the stress transfer path, and induce asymmetric roadway deformation. Studies on retained or residual coal pillars have also indicated that coal pillar width has an important influence on stress concentration, plastic-zone development, and roadway stability. In this study, the numerical simulation and field monitoring results further show that the deformation on the fault-proximal side is greater than that on the coal-seam side, indicating that the TF-3 fault has an obvious asymmetric weakening effect on the surrounding rock. Different from studies that consider fault influence or coal pillar influence separately, this study focuses on the coupled condition in which the heading face approaches a fault-protected coal pillar. Under this condition, roadway deformation is controlled by both fault-induced structural weakening and stress concentration in the protective coal pillar. The results show that when the protective coal pillar width decreases to 10–20 m, roof subsidence, rib convergence, and plastic-zone expansion become more pronounced; when the width increases to 30–40 m, the fault disturbance effect is significantly weakened. Therefore, the protective coal pillar width should be regarded as a key factor in the stability evaluation of near-fault roadways. The coupling effect of fault dip angle and protective coal pillar width can be further understood from the simulation results. The protective coal pillar width controls the distance between the roadway and the fault-disturbed zone, and therefore has a direct influence on the deformation magnitude and plastic-zone development. By contrast, the fault dip angle mainly changes the direction and distribution of the fault-induced disturbance. Under the representative condition of Wp = 20 m, increasing the fault dip angle from 55° to 70° reduced the roof displacement from 11.5 cm to 9.0 cm, but the deformation on the fault-proximal side remained greater than that on the coal-seam side. This indicates that protective coal pillar width is the primary controlling factor, while fault dip angle further modifies the asymmetric deformation pattern under the same coal pillar width. The findings also have practical implications for roadway layout and support design. For the geological and mining conditions of the 52607 working face, a protective coal pillar width of not less than 20 m is suggested. This value should be understood as a recommended minimum stable width under the present engineering conditions, rather than a universally optimal value. In addition, because the fault-proximal rib shows larger deformation than the coal-seam-side rib, local reinforcement on the fault-proximal side and roof should be considered when the roadway approaches the fault, such as increasing bolt density, strengthening cable support, or enhancing surface control with steel strips and mesh. Several model assumptions may influence the quantitative results. First, the numerical simulation adopted the Mohr–Coulomb model and did not explicitly consider post-peak strain softening, plastic flow, or dilation of fractured rock masses. This may lead to an underestimation of local deformation and plastic-zone development in the fault-affected zone. This tendency is consistent with the comparison between simulation and monitoring results, where the monitored roof subsidence was approximately 15.5%–33.8% larger than the simulated value, and the monitored rib convergence was approximately 8.9%–19.3% larger than the simulated value. Second, the model assumed that the rock layers were relatively homogeneous and isotropic, while the actual surrounding rock near the fault may contain joints, weak interlayers, and local fractured zones. This simplification may weaken the simulated deformation localization and asymmetric failure characteristics. In addition, the TF-3 fault was simplified as a single fault plane with a fixed dip angle and fault-zone width, whereas the actual fault geometry and fracture intensity may vary along the roadway strike. The support system and excavation process were also idealized in the numerical model, which may further explain why the simulated deformation values are smaller than the field monitoring values. Therefore, the numerical results in this study should be interpreted mainly as a comparative evaluation of deformation trends rather than an exact prediction of absolute field deformation. Future studies should consider strain-softening or ubiquitous-joint models, heterogeneous fault damage zones, and longer-term field monitoring to further improve the accuracy of near-fault roadway stability evaluation. Conclusions This study took the 52607 return airway of Chengjiagou Coal Mine as the engineering background and employed in-situ stress measurement, theoretical analysis, numerical simulation, and field monitoring to investigate the surrounding-rock stability and support performance as the heading face approached a fault-protected coal pillar. The main conclusions are as follows: The in-situ stress field in the study area is characterized by relatively strong horizontal tectonic stress. At the measuring point at the end of the track roadway in Panel 526, the maximum horizontal principal stress, vertical stress, and minimum horizontal principal stress were 4.63 MPa, 4.50 MPa, and 3.13 MPa, respectively, showing a stress relationship of σH > σV > σh, which corresponds to a Type-II stress field. This stress condition suggests that horizontal tectonic stress may play an important role in roadway deformation in the study area.The protective coal pillar width is the key factor controlling the deformation and plastic-zone development of the near-fault roadway. When Wp decreases to 10–20 m, roof deformation, rib convergence, and plastic-zone expansion become more pronounced, indicating that this range is the main fault-affected zone. When Wp increases to 30–40 m, the displacement concentration and plastic-zone development are significantly weakened. The plastic-zone depth on the fault-proximal side increases from about 1.4–1.5 m under Wp = 30–40 m to 2.5 m under Wp = 20 m and approximately 3.0 m under Wp = 10 m. Therefore, for the 52607 working face, a protective coal pillar width of not less than 20 m is recommended.Field monitoring verified the deformation trend obtained from the numerical simulation. The maximum rib convergence at the three monitored cross-sections was 16.1 cm, 14.7 cm, and 15.8 cm, respectively, and the maximum roof subsidence was 8.2 cm, 9.0 cm, and 9.5 cm, respectively. The deformation on the fault-proximal side was consistently larger than that on the coal-seam side, showing obvious asymmetric deformation. Compared with the numerical simulation result for Wp = 20 m, the monitored roof subsidence was approximately 15.5%–33.8% larger, and the monitored rib convergence was approximately 8.9%–19.3% larger. Nevertheless, the deformation curves tended to stabilize within the monitoring period, indicating that the existing bolt-cable support system can effectively control the surrounding-rock deformation under the present engineering conditions.Under the representative condition of Wp = 20 m, increasing the fault dip angle from 55° to 70° reduced the roof displacement from 11.5 cm to 9.0 cm, while the fault-proximal rib displacement remained greater than the coal-seam-side rib displacement. This indicates that fault dip angle affects the deformation distribution, but its influence is weaker than that of protective coal pillar width within the investigated range. The in-situ stress field in the study area is characterized by relatively strong horizontal tectonic stress. At the measuring point at the end of the track roadway in Panel 526, the maximum horizontal principal stress, vertical stress, and minimum horizontal principal stress were 4.63 MPa, 4.50 MPa, and 3.13 MPa, respectively, showing a stress relationship of σH > σV > σh, which corresponds to a Type-II stress field. This stress condition suggests that horizontal tectonic stress may play an important role in roadway deformation in the study area. The in-situ stress field in the study area is characterized by relatively strong horizontal tectonic stress. At the measuring point at the end of the track roadway in Panel 526, the maximum horizontal principal stress, vertical stress, and minimum horizontal principal stress were 4.63 MPa, 4.50 MPa, and 3.13 MPa, respectively, showing a stress relationship of σH > σV > σh, which corresponds to a Type-II stress field. This stress condition suggests that horizontal tectonic stress may play an important role in roadway deformation in the study area. The protective coal pillar width is the key factor controlling the deformation and plastic-zone development of the near-fault roadway. When Wp decreases to 10–20 m, roof deformation, rib convergence, and plastic-zone expansion become more pronounced, indicating that this range is the main fault-affected zone. When Wp increases to 30–40 m, the displacement concentration and plastic-zone development are significantly weakened. The plastic-zone depth on the fault-proximal side increases from about 1.4–1.5 m under Wp = 30–40 m to 2.5 m under Wp = 20 m and approximately 3.0 m under Wp = 10 m. Therefore, for the 52607 working face, a protective coal pillar width of not less than 20 m is recommended. The protective coal pillar width is the key factor controlling the deformation and plastic-zone development of the near-fault roadway. When Wp decreases to 10–20 m, roof deformation, rib convergence, and plastic-zone expansion become more pronounced, indicating that this range is the main fault-affected zone. When Wp increases to 30–40 m, the displacement concentration and plastic-zone development are significantly weakened. The plastic-zone depth on the fault-proximal side increases from about 1.4–1.5 m under Wp = 30–40 m to 2.5 m under Wp = 20 m and approximately 3.0 m under Wp = 10 m. Therefore, for the 52607 working face, a protective coal pillar width of not less than 20 m is recommended. Field monitoring verified the deformation trend obtained from the numerical simulation. The maximum rib convergence at the three monitored cross-sections was 16.1 cm, 14.7 cm, and 15.8 cm, respectively, and the maximum roof subsidence was 8.2 cm, 9.0 cm, and 9.5 cm, respectively. The deformation on the fault-proximal side was consistently larger than that on the coal-seam side, showing obvious asymmetric deformation. Compared with the numerical simulation result for Wp = 20 m, the monitored roof subsidence was approximately 15.5%–33.8% larger, and the monitored rib convergence was approximately 8.9%–19.3% larger. Nevertheless, the deformation curves tended to stabilize within the monitoring period, indicating that the existing bolt-cable support system can effectively control the surrounding-rock deformation under the present engineering conditions. Field monitoring verified the deformation trend obtained from the numerical simulation. The maximum rib convergence at the three monitored cross-sections was 16.1 cm, 14.7 cm, and 15.8 cm, respectively, and the maximum roof subsidence was 8.2 cm, 9.0 cm, and 9.5 cm, respectively. The deformation on the fault-proximal side was consistently larger than that on the coal-seam side, showing obvious asymmetric deformation. Compared with the numerical simulation result for Wp = 20 m, the monitored roof subsidence was approximately 15.5%–33.8% larger, and the monitored rib convergence was approximately 8.9%–19.3% larger. Nevertheless, the deformation curves tended to stabilize within the monitoring period, indicating that the existing bolt-cable support system can effectively control the surrounding-rock deformation under the present engineering conditions. Under the representative condition of Wp = 20 m, increasing the fault dip angle from 55° to 70° reduced the roof displacement from 11.5 cm to 9.0 cm, while the fault-proximal rib displacement remained greater than the coal-seam-side rib displacement. This indicates that fault dip angle affects the deformation distribution, but its influence is weaker than that of protective coal pillar width within the investigated range. Under the representative condition of Wp = 20 m, increasing the fault dip angle from 55° to 70° reduced the roof displacement from 11.5 cm to 9.0 cm, while the fault-proximal rib displacement remained greater than the coal-seam-side rib displacement. This indicates that fault dip angle affects the deformation distribution, but its influence is weaker than that of protective coal pillar width within the investigated range.