Coal pillar destabilization prediction and mining method in residual mining areas under remining disturbances

To address the problem of pillar instability in residual mining areas under repeated mining disturbances, the prediction of pillar instability time and mining technology is studied by using numerical simulations, theoretical examinations, and practical experimentation techniques utilizing the remaining facade of the Shenghua Coal Industry 3103. On the basis of the findings, employing a fully mechanized mining technique exerts greater pressure on the coal pillar, increasing the likelihood of plastic failure. However, adopting the segmental filling mining approach can effectively disperse this pressure, consequently mitigating potential damage to the coal pillar. Initially, the coal pillar undergoes a progressive yielding or collapsing process at its periphery, leading to instability once the area of the projectile core of the coal pillar decreases beyond a specific threshold. The dimensions of both the affected roadway and the coal pillar play a pivotal role in determining the coal pillar's load‐bearing capacity. A prediction model for pillar instability is established. If the area of the projectile core is less than 34% during remining, the mining method should be changed or pillar reinforcement should be implemented. The accurate prediction and adaptation of the model's mining approach ensure the secure extraction of coal pillars within the remaining mining region, with successful validation of the model's performance.


| INTRODUCTION
In China, coal has always occupied a dominant position.][11] Remnant coal, which is a coal resource that has not been mined temporarily for various reasons (e.g., technology and safety) in past coal-mining processes, 12,13 is the focus in terms of how to mine remnant coal efficiently and safely to achieve the maximum utilization of resources. 14][17][18][19][20] Research on the remanufactured coal face revealed that the coal pillar overruns the destabilization problem existing when the remanufactured coal becomes the main source, limiting the remanufacturing efficiency and safety.The remining disturbance can trigger the redistribution of coal pillar stresses, [21][22][23] which may lead to coal pillar overrun destabilization, thus leading to a large-scale collapse of the roof of the empty lane. 16,24,25he working resistance of the support increases sharply, which may even trigger a crushing frame accident. 26,27oal pillars in traditional coal mining are very different from remnant coal pillars, especially in terms of destruction mechanism and stress-related problems: in traditional coal-mining operations, the destruction of coal pillars is usually due to the load exceeding its bearing capacity, causing collapse or compression, and the coal pillars are subjected to relatively stable and predictable stresses.In remnant coal pillars, however, the destruction mechanism is more complicated and is usually affected by a variety of factors, such as the destruction of neighboring coal pillars and changes in the mechanical properties of the rock strata at the top and bottom of the working face. 28,29The stress distribution of the remnant coal pillar may be affected by various complex factors, such as the neighboring mining area and the stress transfer after the yielding of other remnant coal pillars.This may result in the remnant coal pillars being subjected to nonuniform stresses, increasing the unpredictability of their stability and long-term behavior. 30Residual coal mining presents distinct characteristics compared with solid coal mining, rendering the conventional theories of mine pressure and seam control unsuitable for precisely forecasting the destabilization timing of coal pillar groups.Therefore, it is essential to investigate and develop new approaches for forecasting and mining coal pillar destabilization in areas with remining disturbances.The significance of this study lies in its potential to improve the safety and efficiency of coal remining endeavors.
Researchers and engineers have conducted extensive studies on coal pillar group instability. 31,32Wang et al. 33 used a triaxial perturbation creep test rig to study the creep characteristics of deep rocks under the action of perturbations, obtained the change law of creep deformation by changing the perturbation amplitude and frequency, and established a nonlinear perturbation creep model.Yu and Huo 34 constructed an "inverted trapezoidal island overburden structure" and analyzed this structure's mechanical model; Their findings indicated that as the width of the overlying coal pillar increases, and the stress on the underlying coal seam rises accordingly.Then, the coal pillar-rock system is more easily destabilized and more energy is released.Zhang et al. 35 investigated the synchronized distortion characteristics and stability of an exceptionally thick rock-pillar system.They also examined the criterion for destabilization of the hard rock seam-coal pillar, stress distribution patterns, and potential control methods.Liu et al. 36 observed that when the coal-rock sample experienced damage, the rock exhibited strain recovery, resulting in an external loading influence on the axial direction of the coal specimen.Moreover, they found that the compressive strength of the coal in the sample rose with the increasing strength of the rock but declined as the height ratio of the coal body to the rock mass increased.Wu et al. 37 conducted additional research on the stability of the adjacent rock concerning the dimensions and stress distribution of the coal pillar.They put forward a damage criterion and explained the mechanism leading to the destabilization of the coal pillar.Feng et al. 38 evaluated the dynamic stability of remnant group pillars in a pillar mining area during the upward mining of remnant coal pillar.As the advancing distance increases, the pressure transferred to the remnant coal pillar increases, leading to the earliest destabilization of the key pillar.Simultaneously, changes in the location and size of the upward disturbance load affect the transfer and transmission of the load, which consequently affects the stability of the group pillar system.Guo et al. 39 evaluated the applicability of different mining methods for problems such as low recovery rate and risk of spontaneous combustion of remnant coal.Hashiba and Fukui 40 emphasize that the collapse of an individual coal pillar results in the transmission of the load to adjacent pillars, triggering an excess burden.This cumulative overloading process ultimately culminates in collective failure, resulting in a sudden, extensive roof collapse in the working face.Wang et al. 41 research findings indicate that optimizing the layout position of the residual coal pillar return roadway in close coal seam group mining is crucial for ensuring roadway stability.Placing the roadway away from the residual coal pillar reduces stress influence, thus guaranteeing stability.Gong et al. 42 research findings involve the establishment of a numerical model using FLAC3D to analyze the rational size of coal pillars in the goaf on both sides based on the mining geological conditions in the pilot area.The study also delves into the feasibility of coordinated reinforcement with steel pipe columns.
The aforementioned studies were mainly focused on the creep of coal or rock mass, stress distribution law of coal (rock) pillar-roof system, existence form of coal (rock) pillar-roof system, and instability law; however, thorough exploration is necessary to examine the prediction model concerning the instability of the remaining coal pillar and its associated mining strategy.
Accordingly, in this study, numerical simulations, theoretical analyses, and engineering tests are conducted to investigate the remnant coal pillar retrieval process and construct the prediction model of remnant coal pillar destabilization, which can predict whether the remnant coal pillar will be destabilized in the retrieval process.

| ENGINEERING GEOLOGICAL CONDITIONS
After the site investigation, we gathered information regarding the No. 3 variety of coal at the 3103 secondary coal extraction operation at Shenghua Coal's remaining mining front.The No. 3 coal seam displays an average thickness of 3.65 m and is buried at an average depth of 300 m.The remaining coal pillar exhibits a width varying from 3 to 22 m, while the empty roadway measures between 4 and 6 m in width.In addition, the expanded old roadway can extend up to 9-12 m in width.To acquire a comprehensive understanding of the rock mechanics pertaining to the upper and lower layers of Shenghua Coal, two drill holes that penetrated the No. 3 variety of coal layers were strategically placed alongside the remaining coal pillar.These boreholes allowed the collection of pertinent drill cores and coal rock samples, which were subsequently analyzed in the rock mechanics laboratory.The resulting data provided crucial rock mechanics parameters for both, the No. 3 variety of coal and the roof slab, as outlined in Table 1.

| NUMERICAL MODEL AND RESULT ANALYSIS
The selected simulation software was FLAC3D, a finite difference method program, for the numerical calculation of the old mining area thick coal seam lane pillar type, lane release type mining after the residual coal as the base condition, and coal rock layer mechanical parameters using Shenghua coal 3103 remining workings geological parameters.

| Establishment of numerical simulation
During the simulation, modifications were implemented on the mining technique, along with alterations to the size of the prevailing roadway and coal pillar, while maintaining a consistent coal seam thickness and depth.This prompted us to implement an orthogonal test.The digital calculation model was characterized by its size, measuring 150 m in length, 2 m in width, and 60 m in height, employing a mesh size of 0.5 m, and the number of grids is 144,000.The x-direction is the strike of the 31 working faces, the y-direction is the tendency of the working face, and the z-direction is the vertical direction during the model.To mitigate boundary effects, the coal pillar width at the model boundary was set to 25 m.During the model operation, the bottom plate was fixed to move along the y-axis, whereas the sides were fixed to move along the x-axis.To minimize boundary effects, a 25-m-wide boundary coal pillar was incorporated into the model, while the permissible mining distance of the working face was limited to 100 m.The relevant parameters can be found in Table 1.According to the geomechanical test results of the surrounding rock of the working face in the mining area, the stress conditions of the numerical simulation model are determined as follows: the vertical stress is 6.25 MPa, and the horizontal stress is 7.5 MPa.Considering the gravity g = 9.81m/s 2 , the simulation uses the Mohr-Coulomb constitutive calculation.After balancing the initial stress, the working face is remining in the x-direction.The mining method is one-time mining at full height, one-time mining 5 m running 10,000 steps, and the stress and plastic deformation data are recorded.Three measuring points (measurement locations 1-3) were positioned at the center of the three coal pillars to monitor the changes in their vertical stress.In Figure 1, a visual representation of the model is presented.

| Computational simulation approach
In this study, 12 distinct computational models are constructed by manipulating various mining techniques, pre-existing roadway dimensions, and coal pillar dimensions.Throughout the excavation of the working face, data are collected at every 5-m interval.(1) The stress value of each measuring point is extracted, and the stress change curve of each measuring point is drawn.( 2) Record the yield state of the remaining coal pillar.A comprehensive computational simulation of the orthogonal test arrangement is presented in Table 2.

| Numerical simulation results and analysis
Given the presence of numerous schemes, three representative schemes are chosen for analysis.Among the three schemes, Scheme 1 stands out because of its narrow coal pillar width, which makes it susceptible to damage.Scheme 1 is analyzed in detail further.Among programs 4-6, coal pillar failure is observed in programs 4 and 5. Program 5, characterized by a wide coal pillar, is selected for further examination.Out of Schemes 7-12, none exhibit advanced failure of the coal pillars.Scheme 10, which is distinguished from the broadest | 495 deserted street and narrowest coal support size, is the focus of the analysis.The residual coal pillars and vacant lanes are subsequently allocated in consecutive numbers.Due to the extensive number of simulation steps, four significant stages are selected from each scenario for research purposes.
Figure 2 shows a distribution graph of the failure components in the road and adjacent rock in Scheme 1.According to the depicted graph, when the open roadway and coal pillar both have a width of 5 m, the remaining coal pillar demonstrates substantial plastic zone penetration capacity throughout the mining process.Moreover, the roof above the open roadway does not collapse extensively in advance, enabling safe mining operations.The stability of the roof over the coal pillar and vacant roadway is influenced positively by the coal pillar width and negatively by the width of the vacant roadway.According to Zhang et al., 14 if plan 1 is executed securely, the intact coal pillar will not encounter plastic zone penetration and the overhead structure above the vacant roadway will not undergo premature collapse in plans 2 and 3. Therefore, these plans are excluded from the list.
Figure 3 presents the spatial distribution of the damage units in the roadway and surrounding rock for in Scheme 5 can be linked to the interplay between the roof stability above the coal pillar and the vacant roadway, with a direct correlation to the coal pillar's width and an inverse correlation to the empty roadway's width. 16As a result, the plastic zone penetrates the remaining coal pillar, and the roof experiences advanced failure in the empty roadway in Scheme 5.The failure phenomenon in Scheme 4 is similar to that in Scheme 5 and more serious; therefore, it is not listed separately.
In-depth analysis shows that coal pillar 1 does not appear advanced damage, because it has sufficient width, providing effective support and reducing stress concentration.However, with the completion of coal pillar 1, the plastic zone of coal pillar 2 began to run through, the bearing capacity decreased sharply and the stress concentration was more serious, resulting in a large area of advance collapse of empty roadway 2. It shows that the completion of the recovery of coal pillar 1 and the relative reduction of coal pillar 2 lead to stress migration, which transfers stress to the two roofs of the empty roadway and leads to collapse.When the mining is continued until the completion of coal pillar 2, the roof collapse at the three empty roadways is more serious, because the complete mining of coal pillar 2 causes greater stress migration.
Figure 4 presents the spatial distribution of the damage units in the roadway and surrounding rock for Scheme 6.As depicted in Figure 4, when the size of the empty lane is 10 m and the size of the coal pillar is 20 m, there is no significant decrease in the plastic zone penetration bearing capacity of the coal pillar during the retrieval process, no large overtopping collapse of the roof plate in the empty lane, and the working face achieves safe retrieval.
Figure 5 presents the spatial distribution of the damage units in the roadway and surrounding rock for Scheme 10.When the size of the empty lane is 10 m and the size of the coal pillar is 10 m, the coal pillar does not show a significant decrease in the plastic zone penetration bearing capacity during the retrieval process, the roof plate in the empty lane does not show a large area of overadvance collapse, and the workface achieves safe retrieval.The reasons are the same as Scheme 6; other plans do not show remnant coal pillar plastic zone penetration and empty lane at the roof plate overrun damage; therefore, they are not listed.
Figures 2-5 demonstrate that in Scenarios 4 and 5, the plastic zone of the remnant coal pillar penetrates, resulting in a significant decrease in its bearing capacity and a substantial roof overtopping collapse.The damage to the remnant coal pillar extends from its surface to its interior, and its effective bearing area continues to decrease, which increases its stress concentration degree, leading to a further expansion of the yield range and a corresponding decrease in the elastic core area.The coal pillar will experience destabilization when the decrease in the elastic core area exceeds a certain threshold.Under the integrated mining method, in Scenarios 4 and 5, when the old-lane width was increased to 10 m, the mining pressure was concentrated and intensified on the coal pillar, resulting in plastic zone penetration of the remnant coal pillar and a significant decrease in the bearing capacity, which eventually triggered a large overtopping collapse.Especially when the coal pillar size is small, the elastic core area is relatively small, which cannot effectively buffer and disperse the mining pressure.This makes the coal pillar more vulnerable to plastic damage.In Scenarios 1-3 and 6, due to the smaller width of the old roadway, the coal pillar can better disperse and bear the pressure, and can better bear the pressure, without similar problems.When the fully mechanized mining technology is used for operation, the disturbance caused by the remaining coal pillars is more significant, and the stress diffusion effect generated during the mining process is more significant.This will lead to more severe changes in the stress state of coal pillars and empty roadways, further leading to plastic failure and advanced collapse, which makes coal pillars more likely to suffer plastic failure.whereas with the segmental infill mining method, the mining pressure is effectively dispersed avoiding some possible damage.It can be seen that in fully mechanized mining technology, when the size of the coal pillar is less than a certain threshold and the size of the empty roadway is greater than a certain threshold, the safety of residual coal remining may not be guaranteed.Because small-sized coal pillars are more susceptible to the concentrated influence of internal pressure, they are prone to plastic deformation and failure.Therefore, in this case, the segmented filling mining method may be more appropriate, because it can disperse the mining pressure effectively and reduce the stress concentration of the coal pillar.As displayed in Figure 6, in each scenario, the dynamic changes in stresses show a slow rise first, followed by an accelerated rise and a rapid decline after reaching the peak.This trend illustrates the progression and release of stress within the coal pillar, providing insights into the variations in its bearing capacity during the extraction process.The disparities in old-lane width and coal pillar width indicate that option 1 experiences higher stress levels compared with option 3, option 4 has more significant stress than options 5 and 6, and option 10 exhibits higher stress than option 9.These findings emphasize the critical role of coal pillar width and oldlane width as key factors influencing the coal pillar's bearing capacity.Moreover, in the previous analysis, coal pillar plastic zone penetration and a large overtopping collapse occurred in the integrated mining method, particularly in Scenarios 4 and 5.This indicates that the increased width of the old lane leads to a higher stress concentration, which reduces the coal pillar bearing capacity, as reflected in the stress peak data, which are higher in these scenarios.

| Forecasting model for the stress of coal pillars during mining disturbances
To understand the intrinsic connection among various mining factors, including the passageway width, coal pillar breadth, distance from the cutting opening, and vertical stress, Figure 6 demonstrates that, overall, the stress level is higher at measurement point 3 compared with points 2 and 1. Origin software is used to fit the regression equation for the data at measuring points 1 and 3 in the simulation.Specifically, the two mining methods (fully mechanized mining and segmental filling mining) and two old roadway widths (5 and 10 m) are divided into four prediction models.During the analysis, independent variables encompassed the coal pillar's width and the distance from the measuring point to the cutting hole.The corresponding stress at the measuring point was taken as the dependent variable.Figure 7 displays the construction curves representing the residual coal pillar's width, the distance between the measuring point and cutting hole, and the coal pillar's maximum stress throughout the mining process.
Let x represent the width of the remaining coal pillar (in meters), y represent the distance from the measurement point to the opening eye (in meters), and z denote the maximum stress recorded at the measurement point during mining (in MPa).
The predictive model can precisely calculate the maximum stress value encountered by the coal pillar throughout the extraction process of the remaining coal pillar.Specifically, Equation (1) represents the maximum stress value under fully mechanized mining with a 5-mwide empty roadway, while Equation ( 2) denotes the maximum stress value for the same scenario but with a 10-m-wide empty roadway.For segmental filling mining, Equation (3) yields the maximum stress value when the empty roadway is 5 m wide, and finally, Equation ( 4) corresponds to the maximum stress value under segmental filling mining with a 10-m-wide empty roadway.

| Validation of the model
The aforementioned prediction model was verified using data from measurement point 2 in the simulation.Table 3 shows a comparison of the maximum stress value of each coal pillar with the predicted value.The analysis shows that the error of the fitted data is <3% and its accuracy is suitable for engineering applications.

| Coal pillar stability analysis
As the elastic nucleation zone of the coal pillar decreases, instability occurs, which can be considered a nonlinear process far from the equilibrium state. 43Therefore, the nonlinear theory can be used to explore the destabilization mechanism of remnant coal pillars.Using mathematical tools such as topology and singularity theory, the discontinuous properties near various types of critical points are discussed, especially when parameters are changed under certain conditions to produce abrupt changes in system performance.][46][47] When retrieval is performed, the boundary of the coal pillar first appears as the ruptured band structure.The width of the band structure is Y, the length of the empty lane is a, the width of the coal pillar is w, and the elastic core area of the remnant coal pillar is (w − 2Y), as seen in Figure 8.
The constitutive relationship of the elastic core region of the coal pillar is linear, while the softening property of the yield region of the coal pillar is nonlinear.where the relationship between the remnant coal pillar stress σ, strain ε and damage variable D can be expressed by the following formula: where D = 1 − e ε ε − / 0 , ε 0 is a constant, and E is the modulus of elasticity.When the height of the coal pillar is h, Equation ( 5) can be used to describe the relationship between load p s and deformation u: where u 0 indicates the deformation value corresponding to the peak load.
The area of the coal pillar nucleation zone is w Y ( − 2 ) 2 , and the corresponding load is Therefore, strain energy V s of the remnant coal pillar in the yield zone and elastic potential energy V e in the elastic core region are as follows: The coal pillar is compressed and the self-weight potential energy of the overlying rock is From the total potential energy function of the system shown in Figure 9, we obtain Calculating the first-order derivative of Equation ( 10) and setting this derivative to zero yield the formula for the equilibrium surface as follows: Equation (11) represents the equilibrium state of the mechanical model.Considering the second-order derivative of Equation (11) and setting it to zero, obtaining a meaningful solution of u u u = = (3 − 3 ) 1 0 , expanding the formula, and taking the third term and simplifying it yield We use x as the state variable and p and q as control variables, that is, Coal pillar elastic core and plastic partition.
Curve of area share of coal pillar bullet core area.
Combining Equations ( 13)-( 15) yields the equilibrium formula for the standard form of the mutation model with state variable x and control variables p and q as follows: Calculating the first-order derivative of Equation ( 16) yields the singularity value formula: Combining Equations ( 16) and ( 17) yields Simplifying Equation ( 14) yields Substituting p and q from Equation (20) into p q Δ = 4 + 27 = 0 3 2 and simplifying yield When Δ = 0, the coal pillar will reach the critical equilibrium state; when Δ < 0, the coal pillar will exceed this critical state and plastic damage will occur.Therefore, Δ < 0 is a necessary condition for instability to occur in the remnant coal pillar and solving this inequality for a meaningful solution yields: That is, when the unilateral yield zone width Y < 0.33w (when the area of the elastic core zone is larger than 34% of the coal pillar cross-segmental), the coal pillar will not be destabilized.

| Coal pillar instability prediction model
Let the maximum stress value (see Figure 6) to which the coal pillar is subjected during the retrieval process be x, and the area of the elastic core zone (without plastic damage) of the coal pillar in the numerical simulation be y.Four prediction models are constructed for two retrieval methods (integrated mining and segmental infill mining) and two old-lane widths, respectively.Figure 9 showcases graphical representations depicting the peak stress magnitude within the coal pillar and the ratio of the elastic core zone area of the coal pillar (see Figure 10 for the distribution grid indicating the share of damaged cells).
Figure 9 presents the correlation curve illustrating the relationship between the coal pillar's mining width and the safety factor of coal pillar stability during the instability of the remaining coal pillar.The mining method's full mechanization with an open roadway size of 5 m results in maximum stress values of 21.95, 18.78, and 17.77 MPa for the remaining coal pillar, accompanied by pillar projectile core areas of 44%, 53%, and 61%, respectively.When the open roadway size is 10 m, the maximum stress values for the remaining coal pillar are 26.39,24.55, and 21.14 MPa, while the corresponding pillar projectile core areas are 23%, 30%, and 35%, respectively.Alternatively, using the segmental filling mining method with a 5 m open roadway size yields maximum stress values of 19.81, 17.87, and 16.66 MPa, along with pillar projectile core areas of 53%, 59%, and 66%, respectively.Finally, when the open roadway size is 10 m, the segmental filling mining method results in maximum stress values of 21.44, 20.64 WEN ET AL.
| 503 respectively, and the area of the coal pillar nucleation area is 38%, 49%, and 57%, respectively.Equations ( 23)-( 26) are derived by data fitting of the regular curve.The predictive model exhibits the ability to calculate the proportion of the projectile core area within the coal pillar when provided with a specific stress value, and it maintains a fitting data error of less than 3%.If, during the remining phase, the core area constitutes more than 34% of the total area, the extraction of the remaining coal pillar can be considered safe.However, if the core area is less than 34% during the recovery of abandoned coal, it is necessary to modify the mining methods or reinforce the remaining coal pillars before extraction.

| Mining method at the operational site
To validate the practicality of the forecasting model, practical tests were conducted at the Shenghua Coal Industry's 3103 working face.The site spans a total length of 84 m with a mining length of 420 m.The coal seam No. 3, having an average thickness of 5.65 m, was supported using a four-pillar hydraulic support model ZF3800/15/23 for the purpose of low-top coal caving.

| Experimental program
The coal pillar's ultimate strength was determined through the application of Bieniawski's formula, 48,49 yielding the following calculation: Let σ p represent the ultimate strength of the coal pillar (in MPa), while σ m denotes the uniaxial compressive strength of the coal pillar (in MPa).The width of the coal pillar is denoted by w (in meters), and the height of the coal pillar is represented by h (in meters).When the width-to-height ratio (w/h) exceeds 5, the exponent value (n) is set to 1.4.Conversely, when the width-to-height ratio (w/h) is less than 5, the exponent value (n) is set to 1.In this case, as the coal pillar width-to-height ratio is typically less than 5, we consider n = 1.
The ultimate strengths of coal pillars 1-3 shown in Figure 10 were calculated using Equation (27).The uniaxial compressive strength of coal was 19 MPa, and the results are listed in Table 4.
As is evident from the ultimate strength data provided in Table 4, coal pillar No. 2 will most likely experience yielding failure.The coal pillar had an approximate width of 8 m, whereas the surrounding empty roadway had a width of approximately 4.5 m.According to calculations based on Equation ( 1), the coal pillar may experience a maximum stress of 24.55 MPa during the mining period.Furthermore, using Equation (23), it is anticipated that the area of the projectile core at this time will be 30%, which falls below the safety threshold of 34%.As a result, the coal pillar will undergo deformation and failure while being mined, resulting in a reduction in its load-bearing capacity.This, in turn, may render the neighboring coal pillar unstable owing to an excessive load, resulting in a cascading effect similar to that of dominos and leading to an extensive collapse.This scenario poses a significant threat to the safety of the mine.
Therefore, the method of mining was changed to segmental infill mining for recovery, which can effectively reduce the stresses on the coal pillar during the recovery process.Using Equation (3), we predicted that the maximum stress value of coal pillar No. 2 would be reduced to 20.24 MPa, which is much lower than the previously predicted value.Later, we used Equation (25)  to predict that the area of the coal pillar's elastic core zone would increase to 47% at this time, far exceeding the safety threshold of 34%.
At the same time, five bracket pressure recorders were arranged along the 3103 working face, which were  7.The measurement holes in the recorders were connected to the lower cavities in the front and rear pillars of the brackets.The monitoring of the hydraulic support load facilitates the determination of roof pressure changes with the advance of the working face and the analysis of the prediction model against the industrial tests.

| Assessment of distribution of resistance during operation of the hydraulic bracket
According to Figure 11, the front pillar experiences an average stress ranging from 392.2 to 493.3 kN, constituting approximately 44%-54% of the rated resistance.On the other hand, the rear pillar bears an average stress ranging from 79.2 to 152.1 kN, accounting for approximately 8%-16% of the rated resistance.The support system can adequately withstand the load exerted by the working face, thereby ensuring safe mining operations.These findings suggest that the adoption of the segmental filling technique effectively mitigates the stress on the coal pillar during mining, thereby improving the overall stability of the mine.The stress data collected from both the hydraulic support system's front and rear pillars offer corroborative evidence for the proposed method.The prediction model aligns with the results of the actual industrial tests, thereby confirming the model's feasibility and accuracy.

| Industrial test conclusions
A prediction model was used to predict the possible yield failure of No. 2 coal pillar.After the method was changed to segmental filling mining, safe mining could be performed on the working face.The results obtained from the prediction model align well with the findings of the industrial experiments, further validating the feasibility and accuracy of the model.This predictive tool can offer valuable guidance for ensuring the safe remining of abandoned coal mines under similar mining conditions.

| CONCLUSION
The study focuses on investigating the internal loads experienced by the coal pillar group and roof during the remining process under disturbed conditions.On the basis of the analysis, the following conclusions are drawn: (1) In the fully mechanized mining method, the remaining coal pillar experiences greater pressure, resulting in a greater susceptibility to plastic failure.In this case, the coal pillar first gradually yields or collapses from the edge, then spreading from the surface of the coal pillar to the interior, and the nuclear area inside the coal pillar gradually shrinks.Instability failure is observed when the nuclear area of the coal pillar decreases beyond a specific threshold.Segmental filling mining can effectively disperse the mining pressure and reduce damage to the remaining coal pillar.The dimensions of both the pre-existing roadway and coal pillar play a crucial role in determining the coal pillar's load-bearing capacity.When the width of the old roadway increases, particularly in fully mechanized mining, the stress concentration is high, which may lead to the premature caving of a large-area roof.
(2) The model has the capability to forecast the maximum stress value of the coal pillar and the proportion of the elastic core area under a specified stress, with a fitting data error below 3%.In cases where the projectile core area is below 34% during the remining of abandoned coal, it is recommended to modify the mining operation by reinforcing the remaining coal pillar before proceeding with the mining activities.(3) Upon using the prediction model, it was observed that the mining process for No. 2 coal pillar could potentially lead to failure.In response, the mining method was promptly modified to adopt the segmented filling mining approach.Additionally, a ZF3800/15/23 four-pillar hydraulic support system for low-top coal lowering was employed to protect the roof.As a result, successful mining of the remaining coal pillar group was achieved, providing Average working resistance distribution of the front and rear pillars of the support.

F
Scheme 5.As depicted, there is no initial failure observed during the mining of coal pillar 1, and the open roadway does not experience significant roof caving.However, upon reaching a mining distance of 42.5 m (the completion of mining pillar 1), the plastic zone's penetration capacity of pillar 2 declines considerably, leading to extensive roof caving in open roadway 2. When mining to 47.5 m (7.5 m mining in pillar 2), the penetration bearing capacity of the plastic zone of pillars 2 and 3 decreases significantly, and the roofs of open roadways 2 and 3 show large-area advance caving.Upon reaching a mining distance of 65 m (the completion of mining pillar 2), the roof caving in the third section of the empty roadway intensifies.The collapse phenomena observed T A B L E 2 Numerical simulation orthogonal test scheme.I G U R E 2 Distribution of the roadway and surrounding rock damage unit of Scheme 1: Mining distance (A) 20 m, (B) 35 m, (C) 45 m, and (D) 50 m.

F
I G U R E 3 Scheme 5 roadway and surrounding rock damage unit distribution map: Mining distance (A) 20 m, (B) 42.5 m, (C) 47.5 m, and (D) 65 m.

F
I G U R E 4 Scheme 6 roadway and surrounding rock damage unit distribution map: Mining distance (A) 20 m, (B) 42.5 m, (C) 47.5 m, and (D) 65 m.F I G U R E 5 Scheme 10 roadway and surrounding rock damage unit distribution map: Mining distance (A) 20 m, (B) 35 m, (C) 45 m, and (D) 50 m.

Figure 6
Figure6shows the graphs of stresses at measurement points 1-3 in Schemes 1 and 3, Schemes 4-6, and Schemes 9 and 10 with the change in retrieval distance; the recording time is from the start of retrieval to the completion of coal pillar mining at the measurement point.

F
I G U R E 7 Stress variation curve of each measurement point with retrieval distance: (A) Integrated mining-old-lane width 5 m, (B) integrated mining-old-lane width 10 m, (C) segmental filling mining-old tunnel width 5 m, and (D) segmental filling mining-old tunnel width 10 m. q 18 − 0.002 − 144.24.

2 4 ( 26 )
Let x represent the maximum stress value of the remaining coal pillar during the mining process (in MPa), and y indicate the proportion of the coal pillar's projectile nuclear area.
Physical and mechanical parameters of the top and bottom slabs of coal seams.
Comparison of simulated and predicted maximum stress values of coal pillars.
T A B L E 3 , and 19.66 MPa, F I G U E 10 Distribution of hydraulic bracket load measurement points, empty lane, and leftover coal pillar in 3103 working face.
Ultimate strength of coal pillar.