Study on failure law and repair technology of weak surrounding rock in adjacent roadway under excavation disturbance Authors: Xiaoping Wang, Longyu Hu, Lin Guo, Xiangqian Wang, Xue Li, Xiang Cheng, Junlong He, Peiluo Wang, Qingheng Gu Published: 2026 ABSTRACT Based on a case study of the track cross-cut in Liuzhuang Coal Mine’s western shaft bottom yard, this research investigates the deformation and failure behavior of weak surrounding rock in adjacent roadways subjected to excavation disturbance. A combined methodology was applied, including field measurement, similarity simulation, and numerical modeling, to analyze failure mechanisms and design a remediation strategy for the affected rock zone. Results indicate that roadway deformation primarily exhibits as cross-section convergence, influenced by surrounding rock strength, support resistance, and excavation disturbance. Stress affected zones from adjacent excavations overlap, expanding the unloading range, with the initially excavated roadway shielding stress redistribution in subsequent ones. Compared to shed support alone, the combination of full cables, steel sets, and an inverted arch reduced displacement growth by 95.31%. Numerical simulation validated that after repair, the maximum horizontal displacement increased by 63 mm while the vertical stress peak dropped by4.05 MPa. The integrated support system—combining active cables with passive shed and inverted arch—effectively stabilized the weak rock by enhancing self-bearing capacity, transferring stress deeper, and optimizing the stress environment. The findings offer practical guidance for repairing adjacent roadways under similar conditions. Introduction The growing demand for coal resources in China has necessitated the development of deeper mines1,2. Under the deep mining environment characterized by “five high and two disturbances” conditions, surrounding rock often degrades, leading to deformation, failure, and repeated repairs of roadways3–6. As critical junctions connecting shaft hoisting with main roadway transport, shaft bottom yard roadways feature large cross-sections, multiple free surfaces, and dense layouts. These characteristics make them especially susceptible to instability and deformation under excavation disturbance7,8. Therefore, understanding the deformation and failure mechanisms of surrounding rock in these areas is essential for ensuring roadway stability and maintaining safe mine operations. A thorough understanding of the deformation and failure behavior of roadways under excavation disturbance is fundamental to ensuring their stability. Recent engineering cases have demonstrated that deep soft rock roadways under high structural stress or adjacent excavation disturbance often experience severe deformation, and combined active–passive support systems have been proven effective in controlling such instability9,10. The principles and rapid advancement of distributed fiber optic sensing (DFOS) provide a powerful tool for this purpose11–13. Numerous scholars have successfully employed fiber-optic monitoring techniques, yielding significant insights through both physical similarity simulations and field measurements14–17. Liu et al.18 found that the disturbance still exists when the roadway is greater than 5 times the diameter of the tunnel. Li et al.19 revealed that tunnel excavation induces stress redistribution and abrupt strain changes in the surrounding rock, emphasizing the critical role of the rock-lining system in ensuring long-term stability. Cheng et al.20 employed distributed fiber optic sensing to reveal that coal mining induces dynamic deformation and crack propagation in overburden strata. Chai et al.21,22 analyzed the application of optical fiber bed technology in the monitoring of surrounding rock deformation and failure, summarized the optical fiber sensing theory and its application characteristics. Sun et al.23 employed distributed optical fiber sensing based on Brillouin scattering to achieve dynamic in-situ monitoring of overburden deformation in a coal mining face. Nowak et al.24 demonstrated the feasibility of deploying a low-cost, high-sensitivity fiber optic sensing network for geomechanical monitoring in large excavation areas using a novel DOFS method. Li et al.25 applied PPP-BOTDA fiber optic sensing in similar model tests to accurately map stope support pressure, demonstrating its high precision and strong anti-interference capability. In summary, while fiber-optic monitoring has been extensively applied and yielded significant results in coal mining research, studies specifically targeting the excavation-induced deformation and failure mechanisms of shaft bottom yard track cross-cuts remain scarce. In particular, one key problem needs to be addressed: the deformation control mechanism and effectiveness of repair technologies under different surrounding rock lithologies and support conditions. Addressing this gap, this study investigates the Liuzhuang Coal Mine’s western shaft bottom yard as a case study. A comprehensive methodology was employed, integrating three-dimensional physical modeling, numerical simulation, and field trials to analyze the deformation and failure behavior induced by track cross-cut excavation. The findings provide a practical basis for designing effective roadway rehabilitation and stability control strategies. Engineering geological survey Geological conditions of roadway The shaft bottom yard in the west area of Liuzhuang Coal Mine contains three track cross-cuts. No. 1 track cross-cut primarily traverses carbonaceous mudstone, mudstone, sandy mudstone, along with the 17 − 1 and 16 − 2 coal seams. The surrounding rock of No. 2 track cross-cut consists mainly of mudstone, sandy mudstone, and the 17 − 1 and 17 − 2 coal seams. No. 3 track cross-cut is opened in the sandy mudstone interbedded with mudstone of the 17 − 1 coal seam roof, exposing lithologies including sandy mudstone, 17 − 1 coal, and siltstone. The comprehensive geological columnar section of No. 3 track cross-cut, partial roadway failure conditions, layout of the track cross-cuts, and monitoring data of surface displacement and support are summarized in Fig. 1. Fig. 1Comprehensive geological histogram. Comprehensive geological histogram. The No.1 and No.2 track cross-cuts are the first excavation roadways, and the No.3 track cross-cuts are the post-excavation roadways that meet the requirements of ventilation and auxiliary transportation. They are excavated to the corresponding waiting chamber position, and the adjacent No.1 and No.2 track cross-cuts have large deformation, so the No.3 track cross-cuts are stopped and backfilled. Roadway deformation Due to the excavation disturbance from No. 3 track cross-cut, significant deformation occurred in No. 2 and No. 1 track cross-cuts. To monitor roadway stability and ensure mine safety, displacement monitoring points were installed in the shaft bottom yard and related chambers. In No. 1 track cross-cut, eight measuring points (J1 ~ J2 to J15 ~ J16) were set west of the intake shaft. No. 2 track cross-cut was equipped with nine points (F1 ~ F2 to F17 ~ F18) arranged from west to east. The monitoring results are presented in Fig. 2. Fig. 2Monitoring results of roadway horizontal displacement on site. Monitoring results of roadway horizontal displacement on site. Figure 2(a) indicates that the maximum deformation in No. 1 track cross-cut occurred at monitoring points J13 ~ J14, with a convergence of 39 mm. This section, composed mainly of carbonaceous mudstone, mudstone, and the 17 − 1 coal seam, features relatively weak surrounding rock. Its proximity to a chamber and the larger cross-section further contributed to deformation under excavation disturbance from No. 3 track cross-cut.As shown in Fig. 2(b), points F1 ~ F2 exhibited a maximum convergence of 36 mm in sandy mudstone, supported by a combined “inner anchoring–external framing” system with an inverted arch. In contrast, significant sidewall convergence exceeding 80 mm was recorded between F3–F4 and F7–F8, peaking at 100 mm. These sections penetrate mudstone and the 17 − 2 coal seam—regions with poorer rock conditions and lower support intensity using shed-support. Located closer to No. 3 track cross-cut, they experienced stronger excavation-induced disturbance. Meanwhile, points F9–F10 and F13 ~ F14, situated farther from the active excavation, showed considerably less convergence: 27 mm and 9 mm, respectively. Roadway structure detection The geological radar method constructs subsurface profiles by analyzing reflected electromagnetic waves of specific frequencies. The frequency, amplitude, and waveform characteristics of the reflected signals are interpreted to determine the structure, position, lithology, and geometry of the target formation26,27. A radar survey line was arranged along the axis of the track cross-cut in the shaft bottom yard, with results presented in Fig. 3. Taking No. 2 track cross-cut as an example, the roadway is divided from east to west into three zones based on support type: a normally-supported section with internal anchoring and external steel sets, a weakly-supported section with only steel sets, and a strongly-supported section incorporating both steel sets and an inverted arch. Fig. 3Geological radar detection results in different regions. Geological radar detection results in different regions. Figure 3(a) displays the geological radar profile for the normally supported section. Within the upper 1.75 m depth, electromagnetic signals show strong fluctuations, indicating fractured surrounding rock. Below this depth, continuous signals suggest intact rock mass. The minimal disturbance from No. 3 track cross-cut excavation allows the surrounding rock to maintain stability, with a determined loose zone thickness of 1.75 m.In the weakly supported area (Fig. 3(b)), radar measurements reveal fractured surrounding rock extending to 3.0 m depth, with additional cracking observed in deeper regions. The loose zone thickness is determined to be 3.0 m.For the strongly supported section (Fig. 3(c)), the surrounding rock demonstrates better quality, supplemented by an inverted arch support structure. Significant signal fluctuations are confined within 1.6 m of the floor, suggesting localized fracturing while maintaining overall stability. The loose zone is limited to 1.6 m in this region. Based on comprehensive field measurements, the deformation and failure pattern of the track cross-cut is characterized by sidewall convergence and floor heave. The failure mechanism involves continuous rheological behavior of the weak surrounding rock under high deep-ground stress and excavation-induced disturbances. The deformation is attributed to poor rock conditions, insufficient support resistance, and significant disturbance from adjacent cross-cut excavation. Similar simulation test Model construction and monitoring scheme Similar simulation can use the results of the model to infer the mechanical phenomena that may occur in the prototype and the law of rock mass pressure distribution, so as to solve the practical problems in rock mass engineering production28,29. In this paper, a three-dimensional plane similarity simulation test platform is selected, and a similarity model is established based on the test background of the track gate of the shaft bottom yard in the western area of Liuzhuang Coal Mine. The similarity simulation test determines the model similarity constant under the conditions of geometric similarity, boundary condition similarity and motion similarity, as shown in Table 1. Table 1Model similarity constants.model scale /cm×cm×cm200 × 100 × 80similitude parametergeometry75poisson ratio1volumetric weight1.5stress112.5load112.5time8.6 Table 1Model similarity constants.model scale /cm×cm×cm200 × 100 × 80similitude parametergeometry75poisson ratio1volumetric weight1.5stress112.5load112.5time8.6 Model similarity constants. In order to make the similar model more in line with the actual situation, the bolt and anchor cable support are added to the simulated roadway, and the prefabricated roadway and bolt ( cable ) method are used to simulate the roadway support. The roadway is prefabricated with high-strength foam, and the prefabricated bolts ( cables ) are arranged on the foam prefabricated into the roadway according to the calculated row spacing30,31. According to the size of the similar model, the total weight of the required similar materials, the weight of each stratum, and the weight of river sand, lime, and gypsum are determined. The similar materials that are well matched according to the experimental scheme are slowly laid in the model pouring area, and the similar materials are knocked to the specified height to ensure that the similar materials are dense. The disturbance influence range and stress distribution of roadway excavation are analyzed. Taking the shaft bottom yard of Liuzhuang Coal Mine as the background, the model establishment is shown in Fig. 4 ( a ). The embedded optical fiber in the model is a high transmission tight sheath strain sensing cable ( NES-DSS-C07 type ). The diameter of the optical fiber is 2 mm, and the length of the embedded optical fiber is greater than the length of the model to prevent the reflection effect at the end of the optical fiber. Connect the optical fiber demodulator before the experiment, debug the parameters and calibrate the optical fiber test range. In the experiment, three horizontal optical fibers H1, H2 and H3 are arranged, which are 25,50 and 75 cm from the front of the model respectively. Five vertical optical fibers are arranged along the cross section of y = 45 cm, V1, V2, V3, V4 and V5, as shown in Fig. 4( b ). After the model is built, it is maintained for 1–2 weeks, waiting for the model to dry. According to the calculated pressure, the counterweight is added at the top of the model, waiting for the model to stabilize and excavate the roadway. Fig. 4Similar simulation test scheme. Similar simulation test scheme. Similar simulation monitoring system is mainly composed of distributed optical fiber strain sensor and digital camera. Digital speckle monitoring first whitens the surface of the roadway, and arranges speckle measuring points32. The high-definition camera system is used to photograph the speckle changes during the excavation of the roadway. Distributed optical fiber monitoring uses a distributed optical fiber strain sensor based on optical frequency domain scattering technology. The sensing accuracy reaches ± 1.0µε, which is widely used in high-resolution and high-precision measurement fields7. Analysis of monitoring results Analysis of digital speckle results The high-definition camera is used to record the changes of speckles under different excavation stages. The digital speckle software system established by MATLAB 2018a (The MathWorks, Inc., Natick, MA, USA. https://www.mathworks.com/products/matlab.html.) is used to analyze the displacement changes of speckle images. The horizontal and vertical displacements are shown in Fig. 5, taking the No.1 track stone gate as an example. Fig. 5Displacement cloud of similar model. Displacement cloud of similar model. Roadway excavation-induced stress leads to deformation and cross-sectional convergence, as illustrated in Fig. 5(a). The speckle pattern on the surface of No. 1 track cross-cut shows corresponding displacement responses to the excavation of four adjacent roadways, with sidewall convergence values measuring 5 mm, 10 mm, 12 mm, and 15 mm, respectively. Horizontal displacement analysis indicates that the excavation of No. 1 and No. 2 track cross-cuts represents the dominant contributing factor to the deformation, while the influence of the connecting roadway located farther from the surface was minimal. Nevertheless, the excavation of No. 3 track cross-cut still exerted a discernible disturbance effect.As shown in Fig. 5(b), the recorded roof-to-floor convergence values in No. 1 track cross-cut during the four excavation stages were 10 mm, 16 mm, 20 mm, and 25 mm, respectively. The vertical displacement trend resembles the horizontal pattern, though with greater magnitude, suggesting that roof and floor regions respond more significantly to excavation-induced disturbances. Analysis of vertical optical fiber monitoring results According to the monitoring data of 5 vertical optical fibers, the coordinates of the corresponding measuring points are converted into matrices, and the contour cloud map at y = 45 cm is drawn. The results are shown in Fig. 6. Fig. 6Contour cloud image in vertical direction. Contour cloud image in vertical direction. Figure 6 shows a cross-section along the y-direction of the similarity model. As excavation progresses, the range of excavation-induced disturbance expands. Following the excavation of No. 1 track cross-cut, the roof strain shifts abruptly from tensile to compressive, displaying an approximately butterfly-shaped distribution around the roadway, with a tensile stress zone formed between No. 1 and No. 2 track cross-cuts.After the excavation of No. 2 track cross-cut, a compressive stress zone emerges above it, while the tensile stress zone between the two roadways enlarges, significantly reducing the undisturbed area. Subsequent excavation of the connecting roadway doubles the compressive strain in No. 1 track cross-cut, further extending the disturbed zone and markedly increasing tensile stress regions in the model.Upon completion of No. 3 track cross-cut excavation, compressive strain dominates the roof and tensile strain the floor. Between adjacent roadways, the lower strata experience compressive stress, while the upper strata remain under tensile stress. Analysis of horizontal optical fiber results According to three horizontal optical fibers, the contour cloud map of the model z = 28 section is drawn, and the results are shown in Fig. 7 . Fig. 7Contour cloud map in vertical direction. Contour cloud map in vertical direction. Following the excavation of No. 1 track cross-cut, tensile deformation developed along both sidewalls. A similar deformation pattern was observed in No. 2 track cross-cut, with sidewalls also exhibiting tensile deformation. The surrounding rock between the two cross-cuts experienced an expanded tensile stress zone due to superimposed excavation disturbances. After the connecting roadway was excavated, the tensile stress zones on both sides of the joint roadway became interconnected, further influencing the rock between No. 1 and No. 2 track cross-cuts. The subsequent excavation of No. 3 track cross-cut intensified the disturbance in this region, while the rock between No. 2 and No. 3 track cross-cuts displayed a similar deformation response. During adjacent roadway excavation, stress disturbance zones interact and superimpose. The surrounding rock immediately adjacent to the tunnel surface enters a plastic state under unloading. As excavation progresses, particularly near roadway intersections, the combined effect of stress redistribution from neighboring tunnels enlarges the unloading range and intensifies the disturbance. The overlapping stress zones may eventually connect, potentially initiating progressive failure. Additionally, the initially excavated roadway exerts a shielding effect on subsequent excavations, reducing stress concentration in later-driven roadways due to the pre-existing opening. Simulation of deformation and failure law of track cross-cut Establishment of numerical model Based on the deformation and failure characteristics observed in the deep shaft bottom yard track cross-cut under various influencing factors, this study accounts for the significant variation in surrounding rock conditions and supporting forms along different sections of the same roadway. The investigation focuses on the deformation behavior of surrounding rock under excavation disturbance, considering lithological diversity and support configurations. Displacement and stress evolution following roadway excavation are analyzed to inform subsequent repair strategies and stability control measures33,34. The simulation schemes are designed as follows: Scheme 1: The model considers three track cross-cuts at a burial depth of 740 m. Focusing on the section between the intersection of No. 2 track cross-cut and the waiting chamber—where the lithology is poorest—variations in stress and displacement are analyzed under different surrounding rock types: mudstone, interbedded sandy mudstone, sandy mudstone, and siltstone, by adjusting their mechanical parameters. Scheme 2: Using the same model as Scheme 1, this scheme examines the influence of support systems in the weakly supported section between the intersection of No. 2 track cross-cut and the waiting chamber. Stress and displacement responses are evaluated under four support configurations: steel set support only; combined bolt-cable and steel set; bolt-cable with floor anchor cables and steel set; and full cable support with an inverted arch and steel set. In order to simulate the disturbance and failure of the surrounding rock of the No.3 track and the No.2 track, the field deformation area is selected for simulation. The model size is 150 m×200 m×80 m. The three-dimensional cavern group model established in Rhino will be generated into a grid, and the grid will be transformed into a numerical model through Griddle2.0 (Itasca Consulting Group, Inc. https://www.itascacg.com/software/griddle.). The numerical model grid is imported into FLAC3D 6.0 (Itasca Consulting Group, Inc., Minneapolis, MN, USA. The official website is: https://www.itascacg.com/software/flac3d.)numerical simulation software for simulation operation. The measuring points are arranged along the X direction and the Y direction at the model Z = 37 m. According to the data of the measuring points, the three-dimensional displacement cloud diagram of the horizontal displacement of the model at Z=37 m is drawn, and then X = 70.4 m,69.4 m,66.4 m is selected to arrange the measuring lines. The measuring lines are 1 m, 2 m, 5 m and 10 m away from the left side of the No.2 track, and Y = 40 m, 120 m and 150 m are selected to arrange the measuring lines, which are located in the strong support area, the weak support area and the normal support area respectively. By analyzing the variation characteristics of horizontal displacement of different survey lines before and after the excavation of No.3 track cross-cut, the influence of track cross-cut excavation on the excavation disturbance of adjacent roadway is studied. The model and survey line layout scheme are shown in Fig. 8. Fig. 8Model and survey line layout scheme. Model and survey line layout scheme. The model employs displacement boundary conditions, with the lateral boundaries constrained in the x-direction and the front-rear boundaries constrained in the y-direction. The bottom boundary is fixed in all directions, while a vertical stress of 15.75 MPa—equivalent to the overburden pressure at 700 m depth—is applied to the top surface. Displacement values were reset following initial stress application. The Mohr–Coulomb constitutive model was adopted for the simulation. Rock mass mechanical parameters, assigned based on field data and laboratory testing, are summarized in Table 2. Table 2Physical and mechanical parameters of surrounding rock.Lithologic charactersDensity/kg×m− 3Young’s modulus/GPaPoisson ratioInternal friction angle/°Cohesion/MPaTensile strength/MPamudstone26003.60.3230.02.801.68sand-mud interbed26005.50.2640.84.041.98sandy mudstone258013.60.2342.45.914.07siltstone260015.50.1348.27.835.57coal14000.60.3330.41.911.07 Table 2Physical and mechanical parameters of surrounding rock.Lithologic charactersDensity/kg×m− 3Young’s modulus/GPaPoisson ratioInternal friction angle/°Cohesion/MPaTensile strength/MPamudstone26003.60.3230.02.801.68sand-mud interbed26005.50.2640.84.041.98sandy mudstone258013.60.2342.45.914.07siltstone260015.50.1348.27.835.57coal14000.60.3330.41.911.07 Physical and mechanical parameters of surrounding rock. Deformation law of roadway under different surrounding rock conditions In the simulation process, the simulated buried depth is 740 m, and the surrounding rocks from the three forks of the No.2 track to the waiting chamber are mudstone, sand-mudstone interbedded, sandy mudstone and siltstone, respectively. The horizontal displacement of the area after the excavation of the No.3 track is analyzed.as shown in Fig. 9. Fig. 9Horizontal displacement contourprofile of No.3 track cross-cut excavation under different surrounding rock conditions. Horizontal displacement contourprofile of No.3 track cross-cut excavation under different surrounding rock conditions. Under different surrounding rock conditions, the maximum horizontal displacement at 1 m into the left side of the No. 2 track cross-cut increased as follows: in mudstone, from 256 mm to 320 mm; in sandstone-mudstone interbeds, from 105 mm to 132 mm; in sandy mudstone, from 70 mm to 85 mm; and in siltstone, from 52 mm to 60 mm. Analysis of horizontal displacement trends reveals that both the initial displacement before the excavation of No. 3 track cross-cut and the displacement increment induced by its excavation decreased progressively with increasing rock strength. In the section between the intersection and the waiting chamber of No. 2 track cross-cut, the distribution of excavation-induced horizontal displacement varied significantly with lithology. In mudstone, displacements near the middle of the section were comparable to those close to the intersection. For sandstone-mudstone interbeds, mid-section displacements were slightly lower, while in sandy mudstone and siltstone, they were substantially lower—indicating negligible excavation disturbance in the central zone under these conditions. The area from the third fork point to the waiting chamber of the No.2 track stone gate simulates the vertical stress changes before and after the excavation of the No.3 track stone gate under different surrounding rock conditions as shown in Fig. 10. Fig. 10Stress distribution surface diagram of No.3 track cross-cut excavation under different surrounding rock conditions. Stress distribution surface diagram of No.3 track cross-cut excavation under different surrounding rock conditions. Figure 10 illustrates the vertical stress distribution at the Z = 37 m plane before and after excavating No. 3 track cross-cut under different surrounding rock conditions. The maximum vertical stress initially increases with rock strength up to sandy mudstone, beyond which it declines with further strength improvement, with stress concentration most pronounced near the roadway intersection. Under mudstone conditions, the maximum vertical stress increased from 41.6 MPa to 42.7 MPa, with the stress concentration factor rising from 2.64 to 2.71; for sandstone-mudstone interbeds, these values changed from 44.5 MPa to 44.68 MPa and 2.82 to 2.84; in sandy mudstone, from 51.42 MPa to 51.57 MPa and 3.26 to 3.27; while in siltstone, from 47.06 MPa to 47.13 MPa and 2.987 to 2.99.The reason why the stress concentration factor in sandy mudstone is higher than that in siltstone lies in the interaction between rock strength and stress redistribution capacity. Although both rocks are relatively strong, sandy mudstone has a moderate elastic modulus of 13.6 GPa, which limits lateral stress transmission and causes stress to concentrate more intensely near the roadway. Siltstone, with a higher modulus of 15.5 GPa and a greater tensile strength of 5.57 MPa, allows stress to dissipate over a wider range and transfer to deeper zones, thereby reducing the local peak stress and the corresponding concentration factor. Therefore, a higher rock strength does not always lead to a higher stress concentration factor; the efficiency of stress propagation plays a key role. Deformation law of roadway under different supporting conditions The simulated buried depth is 740 m, and the surrounding rock from the three forks of the No.2 track to the waiting chamber is mudstone. By changing the support in this area, the horizontal displacement and vertical stress of the area after the excavation of the No.3 track under different support methods are analyzed. The first support scheme is shed support; the second scheme is shed + anchor cable ; the third scheme is shed + full anchor cable ; scheme 4 is shed + full anchor cable + inverted bottom arch. Fig. 11Horizontal displacement contour profile before and after excavation of No.3 track cross-cut under different support schemes. Horizontal displacement contour profile before and after excavation of No.3 track cross-cut under different support schemes. Under different support schemes, the maximum horizontal displacement at 1 m into the left side of No. 2 track cross-cut exhibited distinct responses, as shown in Fig. 11. Under Support Scheme 1, displacement increased from 256 mm to 320 mm; Scheme 2 showed an increase from 132 mm to 171 mm; Scheme 3 from 108 mm to 132 mm; and Scheme 4 demonstrated the most controlled behavior, with displacement rising only from 29 mm to 32 mm. Analysis reveals that both the initial horizontal displacement before No. 3 track cross-cut excavation and the subsequent displacement increment induced by its excavation decreased progressively with enhanced support intensity. In the section between the intersection and waiting chamber of No. 2 track cross-cut, the distribution of excavation-induced displacement varied significantly with support conditions. Under basic steel set support, mid-section displacements approached those near the intersection. In contrast, under reinforced support schemes, mid-section displacements were substantially lower, indicating effective containment of excavation disturbance. The performance comparison demonstrates that roadway supported solely by steel sets experienced significant displacement both during initial excavation and under subsequent adjacent excavation. The incorporation of bolt-cable systems substantially improved support resistance, reducing sensitivity to both inherent and external excavation disturbances. The addition of an inverted arch further enhanced stability, resulting in lower displacements in reinforced sections compared to roadway ends. These findings confirm that enhanced support intensity not only improves inherent roadway stability but also reduces susceptibility to adjacent excavation effects. Although the bypass section of No. 2 track cross-cut initially exhibited greater displacement than the ends due to lower support intensity, this pattern reversed under reinforced support conditions. This demonstrates that horizontal deformation is primarily governed by support configuration, while remaining subject to influencing factors such as surrounding rock conditions. Fig. 12Stress distribution surface diagram of No.3 track cross-cut excavation under different support schemes. Stress distribution surface diagram of No.3 track cross-cut excavation under different support schemes. Figure 12 illustrates the vertical stress distribution before and after excavation of No. 3 track cross-cut under varying support conditions. As support intensity increases, the maximum vertical stress exhibits a progressive decline, with peak stress concentrations consistently localized around the roadway intersection zone. Under Support Scheme 1, the maximum vertical stress rose from 41.6 MPa to 42.7 MPa, accompanied by an increase in the stress concentration factor from 2.64 to 2.71. Scheme 2 recorded an increase from 39.82 MPa to 41.30 MPa, with the concentration factor advancing from 2.53 to 2.62. For Scheme 3, stresses increased from 37.78 MPa to 40.07 MPa while the concentration factor grew from 2.40 to 2.54. Notably, Scheme 4 demonstrated minimal variation, with stress values remaining nearly constant between 31.42 MPa and 31.49 MPa and maintaining a stable concentration factor of 1.99. The analysis indicates that reinforced support structures effectively absorb excavation-induced energy, thereby mitigating dynamic instability risks. Enhanced support configurations reduce both the stress magnitude within disturbed zones and the peak vertical stress values, while simultaneously lowering stress concentration factors. Following excavation of No. 3 track cross-cut, robust support systems demonstrated significant resistance to disturbance stresses. The incorporation of an inverted arch structure proved particularly effective, virtually eliminating excavation-induced impacts in the study area. These reinforced support mechanisms substantially improve the load-bearing capacity of surrounding rock and facilitate stress transfer to deeper geological formations. Research on stability control technology of rail cross-cut Track cross-cut optical fiber monitoring In order to monitor the long-term stability of the surrounding rock of the roadway, a metal-based cable-shaped strain sensing cable with good toughness was implanted into the optical fiber monitoring hole by pouring cement mortar. After the installation of the optical fiber is completed, the grouting is carried out to make it fully coupled with the surrounding rock, and the optical fiber test is carried out regularly35,36. The monitoring results are shown in Fig. 13. Fig. 13On-site optical fiber test. On-site optical fiber test. Representative monitoring points were selected for detailed analysis of the fiber-optic strain measurements. The resulting strain profiles plot borehole depth along the horizontal axis against both measured strain and strain variation on the vertical axis. These strain distributions directly reflect mine pressure manifestations in the surrounding rock at each fiber location, as shown in Fig. 14. Fig. 14Fiber optic strain curves of different supporting areas. Fiber optic strain curves of different supporting areas. Fiber-optic monitoring results for the normally supported section in Fig. 14(a) reveal minimal strain variation along the roadway sidewalls. Consecutive measurements showed strain changes below 200 microstrain, indicating both instrument stability and consistent deformation behavior in the surrounding rock. The absence of significant mine pressure manifestations suggests that excavation disturbances from No. 3 track cross-cut did not substantially affect this area. The weakly supported section, shown in Fig. 14(b), experienced considerable excavation impact. During the initial 20-day monitoring period, significant strain developed in two distinct zones: within 2.9 m of the borehole collar and at depths between 4 and 7 m. While shallow strains may reflect thermal influences, the deep-seated variations indicate ongoing rheological behavior in the surrounding rock. Long-term monitoring confirmed stabilized shallow strains but revealed expanding tensile strain at 6–7 m depth, demonstrating continued rheological activity in this stress-sensitive zone. In the strongly supported section presented in Fig. 14 (c), minimal strain variation with consistent patterns across monitoring cycles reflects stable conditions without significant mine pressure. This performance results from favorable geological conditions combined with enhanced support resistance, including the additional reinforcement provided by an inverted arch structure in the floor area. Track cross‑cut repair technology According to the test results of surrounding rock structure, combined with the field distributed optical fiber stability test of surrounding rock. In Sect. 3.3 of the article, different support schemes of the track cross-cut were simulated. The study found that the area to be repaired can remain stable under strong support conditions. Therefore, the full anchor cable is selected to replace the bolt for active support, the metal support is used for passive support, and the support scheme of grouting and backfill is adopted. As shown in Table 3; Fig. 15. Table 3Key parameters of the repair scheme.ParameterCable length/mmCable diameter/mmCable spacing/mmU‑steel typeU‑steel spacing/mmInverted arch thickness /mmValueSidewall:6200Roof:920021.8800 × 60036U6001500 Table 3Key parameters of the repair scheme.ParameterCable length/mmCable diameter/mmCable spacing/mmU‑steel typeU‑steel spacing/mmInverted arch thickness /mmValueSidewall:6200Roof:920021.8800 × 60036U6001500 Key parameters of the repair scheme. Sidewall:6200 Roof:9200 Fig. 15Schematic diagram of repair scheme. Schematic diagram of repair scheme. The specific construction scheme and process are adopted in the area to be repaired : deep hole grouting consolidation surrounding rock → brush expansion roadway surface forming → anchor net cable grouting support → floor pre-excavation, frame anti-bottom arch → frame U-shaped shed → post-frame filling. Simulation of track cross‑cut repair scheme The repair scheme of full anchor cable + metal support + anti-bottom arch support is adopted. The area to be repaired of the No.2 track crosscut is brushed and expanded, and supported. Under the condition that the thickness of the brush is 1 m, the displacement and stress evolution around the roadway are shown in Fig. 16. The layout of the survey line is shown in Fig. 8. Fig. 16The change of displacement stress in the repair of cross‑cut of No.2 track. The change of displacement stress in the repair of cross‑cut of No.2 track. Figure 16 illustrates the brush excavation repair conducted in the significantly deformed section of No. 2 track cross-cut. The rehabilitation employed a combined support system incorporating full-length anchor cables, steel sets, and an inverted arch. Following repair, the maximum horizontal displacement at 1 m from the roadway increased from 320 mm to 383 mm, representing a 63 mm increment, while the vertical stress peak decreased from 42.7 MPa to 38.65 MPa, a reduction of 4.05 MPa. The brush excavation enlarged the roadway cross-section, triggering stress redistribution around the opening. This expansion caused the original stress concentration zone to extend outward, increasing the unloading range within the surrounding rock and leading to continued deformation. Although such sectional enlargement could potentially raise stress peaks, the implemented support system—integrating active reinforcement through full-length anchor cables with passive support from steel sets and an inverted arch—enhanced the surrounding rock’s self-bearing capacity. This composite system successfully transferred stresses deeper into the rock mass, reducing the vertical stress peak by 9.4%. The results demonstrate that this rehabilitation support scheme effectively controls roadway deformation, optimizes the stress environment, and ensures surrounding rock stability. Limitations of the study (1) This study is based on a single mine case. Although the geological conditions and support schemes are representative of many deep soft rock roadways, the generalizability of the results may be limited. Multi-site validation is needed to confirm the broader applicability of the proposed repair scheme. (2) The numerical simulations employed the Mohr-Coulomb constitutive model, which is suitable for describing the instantaneous strength and failure of rock masses. However, this model does not capture long-term creep or time-dependent deformation, which may be significant for weak surrounding rock over extended periods. Future work should adopt creep models to assess long-term stability. (3) The on-site monitoring duration after repair was 14 months. While this period has been sufficient to observe stabilization of deformation rates, it may not cover the full service life of the repaired roadway. Continued monitoring is recommended to verify long-term performance. Conclusions (ⅰ) The deformation and failure of weak surrounding rock in adjacent roadways under excavation disturbance are primarily controlled by three factors: surrounding rock lithology, support structure, and disturbance intensity. The weakly supported section with only shed support experienced severe cross-section convergence, while the strongly supported section with an inverted arch remained stable with a loose zone of only 1.6 m. (ⅱ) By applying OFDR and DIC monitoring techniques in physical similarity simulation experiments, it was found that adjacent roadway excavation enlarges the unloading zone through stress superposition, and the initially excavated roadway exerts a “shielding effect” that reduces stress concentration in later-driven roadways. (ⅲ) Compared to shed support alone, the combined support system of full cables, steel sets, and an inverted arch reduced displacement growth by 95.31%. After repair, the maximum horizontal displacement increased by 63 mm as controlled deformation, while the peak vertical stress decreased by 4.05 MPa, reflecting the yield-resistant support concept. (ⅳ) For the stability control of surrounding rock disturbed by excavation in similar deep soft rock roadways, a combined active–passive support system should be adopted. For weakly supported sections, full cable support with an additional inverted arch can be applied to ensure long-term stability.