A method and system for intelligent dynamic optimization of a coal pillar in a mine

Abstract

The present application belongs to the technical field of mining engineering and mine strata control, and particularly relates to a kind of intelligent dynamic setting optimization method and system of mine protective coal pillar. The method integrates geological and design basic data, dynamically corrects the safety constraint model in real time by collecting monitoring information during mining, and optimizes the solution by taking the model as a hard constraint, minimizing the size of the coal pillar or maximizing resource recovery as the goal, outputs the optimal coal pillar boundary and safety margin, and finally converts the results into executable engineering instructions. When the monitoring data triggers an early warning, the system automatically starts re-optimization, achieving closed-loop dynamic adjustment of the scheme. The corresponding system is realized by the cooperation of modules such as geological database, dynamic perception, constraint modeling, intelligent optimization, result generation and closed-loop control. The present application realizes the transformation of protective coal pillar from static design to dynamic intelligent optimization, significantly improving the resource recovery rate under the premise of safety.

Description

Technical Field

[0001] This application belongs to the technical field of mining engineering and automation control, specifically involving an intelligent dynamic retention optimization method and system for mine protection coal pillars. Background Technology

[0002] In coal mining, protective coal pillars are required to protect surface buildings, water bodies, railways, or important underground roadways from mining damage.

[0003] Existing technologies primarily rely on national standards, regional empirical formulas, and static geological and mining parameters (such as fixed subsidence coefficients and displacement angles) for a one-time design to determine the fixed boundaries of coal pillars. A significant drawback of this method is that:

[0004] First, the rock strata movement parameters used in the design are mostly empirical values ​​or average values ​​of the mining area, failing to fully consider the differences in dynamic rock strata response during the mining process of specific working faces. This leads to conservative designs, with the width of the coal pillar often exceeding the actual requirement, resulting in the permanent covering of valuable coal resources.

[0005] Secondly, existing methods lack the ability to dynamically perceive and assess the mining process in sync with the mining process. Once the design is completed, the coal pillar boundary is fixed, making it impossible to quantitatively assess and proactively adjust the actual protection effect and safety status of the coal pillar based on real-time multi-source monitoring data such as surface subsidence, roadway deformation, and microseismic activity.

[0006] Finally, the entire decision-making process is open-loop. When mining conditions change or abnormal risks are detected, it is difficult to re-optimize the coal pillar size in a timely and scientific manner. The only option is to passively choose between the waste of resources caused by "over-retention" and the safety risks of "adjustment without basis". Summary of the Invention

[0007] To address the shortcomings of existing technologies, this application aims to provide an intelligent dynamic optimization method and system for mine protective coal pillars. This method, while meeting the safety requirements of the protected object, dynamically assesses the impact of mining activities based on real-time monitoring data during the mining process. It then performs closed-loop, quantitative, and auditable optimization and adjustment of the boundary and width of the protective coal pillars. This allows for the maximization of coal resource recovery from the protected coal pillars while controlling safety risks, overcoming the conservative approach and resource waste caused by traditional static, one-time design methods.

[0008] To achieve the above objectives, this application adopts the following technical solution:

[0009] This application provides a method for intelligent dynamic retention and optimization of mine protective coal pillars, including the following steps:

[0010] S1. Obtain and establish a basic design dataset, which includes basic geological and mining data of the target area, permissible critical indicators of the protected objects, and a formalized expression of the initial design boundary of the protective coal pillar and the corresponding safety constraints determined according to current specifications and measured rock movement parameters of the mining area. Among them, the basic geological and mining data includes at least the coal seam occurrence, overlying lithological stratification and key layer location, hydrological and aquitard structure, and geostress background. The measured rock movement parameters include the subsidence coefficient, the tangent of the main influence angle, the inflection point offset distance, and the influence propagation angle.

[0011] S2. During the mining process, a dynamic sensing network is deployed and operated to continuously collect dynamic variable data related to mining impacts. The dynamic variable data is then fused according to a unified spatiotemporal benchmark to generate a quantifiable snapshot reflecting the current state of mining impacts. The dynamic variable data includes at least the real-time advance and exposed area of ​​the working face, the convergence of the surrounding rock in adjacent roadways, deformation monitoring data of key measuring points on the surface or slope, and hydrological dynamic data under mining impacts as needed.

[0012] S3. Based on the formalized expression of the safety constraints in step S1, construct a machine-readable, parameterized safety constraint inequality model, where the decision variable is the boundary position or width of the protective coal pillar to be optimized; the safety constraint inequality model integrates at least the deformation constraints derived from the allowable deformation index of the protected object, the water-proof constraints derived from the overburden failure height and the waterproof safe coal-rock pillar conditions, and the engineering stability constraints derived from the self-stabilizing conditions of the surrounding rock; the parameters of the safety constraint inequality model can be dynamically corrected based on the latest state snapshot generated in step S2;

[0013] S4. Using the safety constraint inequality model constructed in step S3 and dynamically corrected as hard constraint conditions, with the optimization objective of minimizing the width of the protective coal pillar or maximizing the amount of recoverable coal, solve for the optimal solution that satisfies all constraints, output the optimal width of the protective coal pillar or the boundary adjustment amount under the current mining stage, and simultaneously calculate and output the current safety margin of each constraint as a visual risk indicator.

[0014] S5. The optimization results output in step S4 are converted into executable engineering design instructions, including updated coal pillar boundary coordinates, allowable mining operations limits, and key areas for enhanced monitoring. When the monitoring data collected in step S2 indicates that the safety margin has reached the preset warning threshold, the recalculation and optimization process in steps S3 to S4 is automatically triggered to generate a new version of the retention plan, thereby realizing dynamic, closed-loop, and auditable optimization adjustment of the coal pillar boundary, and retaining complete plan versions and decision records throughout the process.

[0015] As a preferred technical solution, the measured rock movement parameters mentioned in step S1 are a set of parameters obtained through long-term monitoring and data analysis by setting up surface and underground rock strata movement observation stations in the target mining area. This set of parameters includes at least the subsidence coefficient, the tangent of the main influence angle, the horizontal movement coefficient, the inflection point offset distance, the boundary angle, and the movement angle. When making the initial design of the protective coal pillar, this set of measured parameters is used to replace the empirical parameter values ​​in the specifications. In the continuous mining process of step S2, new rock movement and deformation data are collected through the monitoring network, and the set of measured rock movement parameters is periodically corrected and updated. The updated parameters are then fed back into the safety constraint inequality model in step S3 to achieve dynamic calibration of the key calculation coefficients of the model.

[0016] As a preferred technical solution, the deformation constraints derived from the allowable deformation index of the protected object in step S3 are constructed separately for different types of protected objects such as buildings, railways, and water bodies. For the protective coal pillars of buildings and railways, a prediction model based on the probability integral method is used to calculate the maximum deformation value of the protected object affected by mining and ensure that it is less than the critical allowable value specified in the standard, thereby forming a deformation constraint inequality. For the protective coal pillars under water bodies, the maximum height of the fracture zone is calculated according to the empirical formula for the development height of the water-conducting fracture zone in the fully mechanized longwall mining of thick coal seams, and a rock pillar with a thickness greater than that of the waterproof safety coal-rock pillar is retained between it and the bottom of the protected water body to form a water-proof constraint. All constraints are expressed as calculable mathematical inequalities with the width of the coal pillar or the boundary coordinates as variables.

[0017] As a preferred technical solution, the safety constraint inequality model described in step S3 is a multi-objective, multi-constraint optimization mathematical model. Its decision variables are the retention width of each side of the coal pillar or the coordinates of the key control points of the coal pillar boundary. In addition to minimizing the coal pillar width, the objective function can also be set as maximizing the value of coal resource extraction over the entire life cycle or minimizing the impact on mining progress under the premise of satisfying safety constraints. The model is solved using linear programming, nonlinear programming, or intelligent optimization algorithms. Each time optimization is triggered, the algorithm searches for the optimal solution in the feasible region based on the latest dynamic variable data and the corrected constraint parameters, and provides a comparison of schemes under different optimization objective weights.

[0018] As a preferred technical solution, the data collected by the dynamic sensing network in step S2 includes the three-dimensional coordinates and energy of rock mass fracture events obtained by the microseismic monitoring system, the surrounding rock separation and deformation rate monitored by the roadway roof delamination instrument and multi-point displacement gauge, the surface movement time series data obtained by the high-precision level and GNSS monitoring station, and the hydrological dynamic data obtained by the underground borehole seepage monitoring equipment. After preprocessing and spatiotemporal registration, these data are not only used to update the state snapshot, but also used for inversion analysis of the stress distribution evolution in front of the working face and inside the coal pillar, the fracture state of the key overburden layer, and the development height of water-conducting fractures, thereby providing a more refined and forward-looking correction basis for the right-hand side of the constraint conditions in step S3.

[0019] As a preferred technical solution, the early warning threshold mentioned in step S5 is set in stages according to the safety margin of the constraint. The system sets a green safety zone, a yellow early warning zone, and a red alarm zone. When the monitoring data shows that the safety margin of a certain constraint drops to the yellow early warning zone, the system automatically marks the risk and prompts the engineering technicians to pay attention. It can also trigger a low-priority optimization recalculation to assess potential adjustment directions. When the safety margin enters the red alarm zone, the system immediately triggers a high-priority optimization recalculation and automatically generates emergency plan suggestions for temporary reinforcement of coal pillars, control of mining speed, or local increase of coal pillars. At the same time, the alarm and plan are pushed to relevant decision-makers for emergency approval and execution.

[0020] As a preferred technical solution, this method is suitable for optimizing the protection coal pillars of roadways affected by the lateral support pressure of adjacent goaf areas. In this scenario, the engineering stability constraints in step S3 need to focus on the long-term strength stability and creep characteristics of the coal pillar. By combining the roadway deformation rate and coal pillar stress data monitored in step S2, the damage and plastic zone range of the coal pillar are dynamically evaluated, and the calculation model of the effective bearing width of the coal pillar is corrected accordingly. The optimization goal is to minimize the width of the section coal pillar as much as possible while ensuring that the cross-section meets the requirements during the roadway's use. The optimization results guide the dynamic adjustment of the working face stop line and roadway support scheme.

[0021] As a preferred technical solution, the executable design results output in step S5 are issued in the form of a digital mining engineering plan, a three-dimensional geological model update layer, and a construction parameter form. The system automatically compares the coal pillar boundaries of the old and new versions and calculates the increase or decrease of coal volume in the changed area and evaluates the safety and economic benefits brought about by the adjustment. All optimization decisions, triggering conditions, input data, output results, and approval processes are completely recorded in the blockchain or a database with anti-tampering function, forming a traceable and auditable complete decision chain.

[0022] As a preferred technical solution, the method is implemented through an integrated intelligent optimization system platform. This intelligent optimization system platform includes geological data management, real-time monitoring access, model calculation engine, optimization solver, visualization display and workflow driving module. The intelligent optimization system platform provides application programming interfaces to enable data integration and command interaction with the mine's geological spatial management information system, safety monitoring and control system and production execution system, so as to realize automated and intelligent closed-loop management from data perception to design optimization to production command issuance.

[0023] This application also provides an intelligent dynamic retention and optimization system for mine protective coal pillars, used to implement the above-mentioned intelligent dynamic retention and optimization method for mine protective coal pillars, the system comprising:

[0024] The geological and design foundation database module is used to store and manage coal seam geological information, rock movement parameters, protected object information, and initial coal pillar design schemes.

[0025] The dynamic sensing and data fusion module is used to access and process real-time and historical data from various sensors such as surface subsidence monitoring, underground surrounding rock deformation monitoring, microseismic monitoring, and hydrological monitoring, in order to generate a spatiotemporally unified state snapshot;

[0026] The safety constraint modeling engine is used to formalize deformation constraints, watertightness constraints, and stability constraints into parameterized machine-readable inequality models based on the type of protected object and safety standards, and can receive state snapshot data to dynamically correct the model parameters.

[0027] The intelligent optimization solution module incorporates various mathematical programming and intelligent optimization algorithms to solve for the optimal coal pillar boundary scheme and calculate the real-time safety margin of each constraint within the feasible region defined by the safety constraint model, with the goal of minimizing the coal pillar size or maximizing resource recovery.

[0028] The design results generation and publishing module is used to automatically convert optimization results into updated engineering drawings, control coordinates and construction guidance documents, and drive them for review and publication through a workflow.

[0029] The risk warning and closed-loop control module is used to continuously monitor the safety margin and automatically trigger process re-optimization and manage emergency adjustment plans when the warning threshold is reached.

[0030] Each module is connected and collaborates with the data service via a system bus.

[0031] Compared with the prior art, the beneficial effects of this application are as follows:

[0032] This application changes the static mode of protective coal pillar retention by constructing an intelligent method that integrates dynamic perception, modeling optimization and closed-loop control.

[0033] First, this method continuously integrates multi-source real-time monitoring data during the mining process and dynamically updates key parameters characterizing the movement of rock strata, enabling the design basis for coal pillar placement to be calibrated according to the actual mining response. This significantly improves the pertinence and accuracy of the design and provides a scientific basis for reducing the size of coal pillars.

[0034] Secondly, by formalizing the safety requirements of the protected object into a computable safety constraint inequality model, and performing online optimization with the goal of minimizing coal pillar size or maximizing resource recovery, a lean design of coal pillar size is achieved under the premise of strictly meeting the safety threshold, thereby enabling the direct and quantitative release of coal resources covered by conservative design.

[0035] Finally, this method establishes a complete closed loop of "monitoring-evaluation-optimization-early warning-re-optimization". When real-time monitoring data indicates that the safety margin deviates from the preset range, the system can automatically trigger the re-evaluation and re-optimization process and generate corresponding early warnings or adjustment instructions. This transforms the setting of protective coal pillars from a one-time decision in the early stage into an intelligent management process that runs through the entire mining cycle, is dynamically adjustable, and has controllable risks, thereby maximizing the resource recovery rate while ensuring safety. Attached Figure Description

[0036] Figure 1 The flowchart illustrates the steps of the intelligent dynamic setting optimization method for mine protection coal pillars in this application. Detailed Implementation

[0037] To enable those skilled in the art to better understand the present application, the technical solutions in specific embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0038] like Figure 1 As shown, this application provides an intelligent dynamic retention optimization method for mine protective coal pillars, including the following steps:

[0039] S1. Obtain and establish the design basis dataset. The dataset includes the geological and mining basic data of the target area, the allowable critical index of the protected object, and the formal expression of the initial design boundary of the protective coal pillar and the corresponding safety constraints determined according to the current specifications and the measured rock movement parameters of the mining area.

[0040] The basic geological and mining data should include at least the coal seam occurrence, overlying lithological stratification and key layer location, hydrological and aquitard structure, and geostress background. Measured rock movement parameters include subsidence coefficient, tangent of the main influence angle, inflection point offset distance, and influence propagation angle.

[0041] The basic geological and mining data are derived from geological exploration reports, mine geological descriptions, and 3D seismic exploration data of the mining area. These data are used to construct a geological model that includes coal seams, various rock strata, faults, collapse columns, etc. The permissible critical indicators for protected objects are determined according to national standards, such as deformation values ​​corresponding to building damage levels. Measured rock movement parameters are obtained by establishing surface movement observation stations at working faces with similar conditions in the mining area. After monitoring the entire process at at least one working face, the parameters are inverted using methods such as curve fitting. These values ​​are more consistent with local realities than standard empirical values.

[0042] For example, the subsidence coefficient of this mining area can be obtained through monitoring. It is 0.75, mainly affecting the tangent angle. The inflection point offset is 2.0. for ( (For mining depth). These measured parameters and design specifications are combined, and traditional methods such as probability integration are used to calculate the initial, conservative protective coal pillar boundary. Simultaneously, the protection requirements are transformed into preliminary mathematical constraints, such as surface subsidence values. It must be less than the critical value .

[0043] The benefit of this step is that it establishes a quantifiable initial benchmark and constraint framework supported by measured data for subsequent dynamic optimization.

[0044] S2 involves deploying and operating a dynamic sensing network during the mining process to continuously collect dynamic variable data related to mining impacts. This dynamic variable data is then fused according to a unified spatiotemporal benchmark to generate a quantifiable snapshot reflecting the current state of mining impacts. The dynamic variable data includes at least the real-time advance and exposed area of ​​the working face, the convergence of surrounding rock in adjacent roadways, deformation monitoring data from key surface or slope measuring points, and, as needed, hydrological dynamic data under the mining impacts.

[0045] Data integration is carried out based on existing monitoring systems such as the hydraulic support pressure monitoring system at the working face, the roadway roof delamination meter, the multi-point displacement meter, the roadway surface convergence monitoring station, the ground GNSS monitoring station, and the microseismic monitoring system.

[0046] For example, monitoring points are densely deployed on the ground above the protective coal pillar and in the affected roadways. All sensor data is transmitted to a unified data platform via an industrial ring network, where each data point is marked with a precise timestamp and spatial coordinates. The data fusion engine generates a status snapshot at a set time frequency (e.g., hourly), which integrates information such as the working face position up to that moment, the latest displacement of each measuring point, and the cumulative distribution of microseismic events.

[0047] Its working principle is to transform discrete, static monitoring into continuous, spatially correlated dynamic sensing through Internet of Things (IoT) technology, capturing the rock strata response caused by mining activities in real time. The beneficial effect is the realization of all-weather, multi-dimensional monitoring of the protective coal pillar and surrounding rock conditions, providing a real-time, objective data stream for dynamic assessment, and overcoming the shortcomings of traditional methods such as information lag and bias.

[0048] S3, based on the formalized expression of the safety constraints in step S1, construct a machine-readable, parameterized safety constraint inequality model, where the decision variable is the boundary position or width of the protective coal pillar to be optimized. The safety constraint inequality model integrates at least the deformation constraints derived from the allowable deformation index of the protected object, the water-tight constraints derived from the overburden failure height and the waterproof safety coal-rock pillar conditions, and the engineering stability constraints derived from the self-stabilizing conditions of the surrounding rock. The parameters of the safety constraint inequality model can be dynamically corrected based on the latest state snapshot generated in step S2.

[0049] The core of the model is the constraint set. Deformation constraints are typically based on probabilistic integral models. Taking the protective coal pillar under a building as an example, its constraint can be expressed as: the expected maximum subsidence value. The predicted function is... It concerns the coal pillar boundary coordinates (decision variables) and rock movement parameters ( , Complex functions (e.g., water-conducting fractures). The water-tight constraint can be expressed as: the expected height of the water-conducting fracture zone. ,in It is the height of the bottom interface of the water body. This refers to the thickness of the waterproof and safe rock pillar. Engineering stability constraints may be expressed as: the aspect ratio of the coal pillar must be greater than a certain critical value. Or the load borne by the coal pillar is less than its long-term strength.

[0050] Model parameters (e.g.) , , (etc.) are no longer fixed values, but dynamic variables associated with the state snapshot in step S2. For example, temporary subsidence coefficients can be derived from monitoring data of the current mining stage. Substitute it into the constraint model.

[0051] Its working principle involves encoding various safety regulations and engineering experience into a set of inequalities with coal pillar size as the variable, and incorporating the influence of real-time monitoring data into these inequalities through dynamic parameters. The beneficial effect is the creation of a flexible, accurate, and computable "digital sandbox," making it possible to assess and adjust the theoretical effects of coal pillar boundaries at any given mining time.

[0052] S4. Using the safety constraint inequality model constructed in step S3 and dynamically corrected as a hard constraint condition, with the optimization objective of minimizing the width of the protective coal pillar or maximizing the amount of recoverable coal, solve for the optimal solution that satisfies all constraints, output the optimal protective coal pillar width or boundary adjustment amount under the current mining stage, and simultaneously calculate and output the current safety margin of each constraint as a visual risk indicator.

[0053] The optimization objective function can be ,in It is the reserved width on each side of the coal pillar, or it can be... The recoverable volume of coal. The constraints are the system of inequalities constructed in S3. This is a typical constrained nonlinear programming problem.

[0054] In specific software implementations, functions such as MATLAB's `fmincon` can be called, or heuristic algorithms such as genetic algorithms can be used for solving the problem. The solver searches the space of decision variables (coal pillar boundaries) and finds the point that optimizes the objective function while satisfying all dynamic constraints. After solving, it outputs not only the optimal coal pillar boundary but also... It also calculates the "slack" of each constraint under the current optimal solution, i.e., the safety margin.

[0055] For example, for deformation constraints, calculate the margin. This margin directly reflects the security reserves of this protection dimension.

[0056] Its working principle is to use operations research methods to automatically find the design scheme with the highest resource recovery efficiency within the "feasible region" defined by the safety law. The beneficial effect is that it realizes the leap from "manual empirical calculation" to "automatic machine optimization", which can quickly find the theoretically optimal solution under complex constraints, and reveal the "safety buffer space" of the scheme with quantitative safety margin, providing a clear basis for decision-making.

[0057] S5 transforms the optimization results output from step S4 into executable engineering design instructions, including updated coal pillar boundary coordinates, permitted mining clearances, and key areas for enhanced monitoring. When subsequent monitoring data collected in step S2 indicates that the safety margin reaches a preset warning threshold, the recalculation and optimization process in steps S3 to S4 is automatically triggered to generate a new version of the retention plan. This achieves dynamic, closed-loop, and auditable optimization adjustments to protect the coal pillar boundary, while retaining complete plan versions and decision records throughout the process.

[0058] The optimization results are automatically generated through the system interface, design change orders are generated, coal pillar boundary graphics in the geodetic spatial database are updated, and then sent to the production command system.

[0059] The system is configured with early warning rules. For example, when a safety margin M falls below 30% of its initial value, a yellow alert is triggered, and the system automatically initiates a non-urgent re-optimization calculation to assess whether the next stage of mining can be optimized in advance. When the margin M falls below 50% of its initial value, a red alert is triggered, and the system immediately initiates emergency optimization, potentially providing emergency solutions such as "stop advancing towards the coal pillar" or "reinforce the coal pillar," while simultaneously sending an alarm to the responsible person. All raw data, optimization calculation parameters, results, early warnings, and approval records are stored in the database, forming an immutable audit log.

[0060] Its working principle is to establish a rule-based automated decision support and execution feedback system. The beneficial effect is that optimization is transformed from isolated calculations into a routine management process, realizing "dynamic health checks" and "dynamic prescriptions" for coal pillar protection, ensuring real-time controllability of safety risks throughout the entire lifecycle, and ensuring that all decisions are traceable, meeting the requirements of modern mine lean management and compliance.

[0061] As a preferred technical solution, the measured rock movement parameters in step S1 are a set of parameters obtained through long-term monitoring and data analysis by setting up surface and underground rock strata movement observation stations in the target mining area. This set of parameters includes at least the subsidence coefficient, the tangent of the main influence angle, the horizontal movement coefficient, the inflection point offset distance, the boundary angle, and the movement angle. When making the initial design of the protective coal pillar, this set of measured parameters is used to replace the empirical parameter values ​​in the specifications. In the continuous mining process in step S2, new rock movement and deformation data are collected through the monitoring network, and the set of measured rock movement parameters is periodically corrected and updated. The updated parameters are then fed back into the safety constraint inequality model in step S3 to achieve dynamic calibration of the key calculation coefficients of the model.

[0062] In the initial mining face or working faces with similar geological conditions, surface mobile monitoring stations are deployed along the strike and dip of the main cross-section, with the spacing between monitoring points meeting accuracy requirements. Simultaneously, roof and floor approach and sidewall convergence monitoring stations are deployed in key underground roadways. After the entire mining impact period, complete data on surface subsidence basins and underground deformation are obtained.

[0063] By employing optimization algorithms such as the least squares method, the probability integral method prediction model is fitted with measured data to derive the optimal set of rock movement parameters. In the subsequent dynamic optimization of the protective coal pillar, the new data collected in step S2 (which may come from newly added monitoring points near the protective coal pillar) is regarded as a new and incomplete sample. Online learning algorithms such as recursive least squares are used to optimize the parameter set. Fine-tuning was performed to obtain... Then Substitute into the constraint model of S3 middle.

[0064] Its working principle utilizes the concept of "mining as experimentation," treating the entire mining process as a continuous learning process for rock movement parameters, allowing the prediction model to continuously approach reality. The beneficial effect is a significant improvement in the accuracy of rock movement prediction, making the constraint model upon which optimization relies increasingly "intelligent." This provides a solid scientific foundation for further narrowing the width of coal pillars and releasing stagnant coal reserves, while ensuring safety—a fundamental prerequisite for achieving dynamic optimization.

[0065] As a preferred technical solution, in step S3, the deformation constraints derived from the allowable deformation index of the protected object are constructed separately for different types of protected objects such as buildings, railways, and water bodies. For the protective coal pillars of buildings and railways, the maximum deformation value of the protected object affected by mining is calculated using a prediction model based on the probability integral method, and it is ensured that it is less than the critical allowable value specified in the standard, thereby forming a deformation constraint inequality. For the protective coal pillars under water bodies, the maximum height of the fracture zone is calculated based on the empirical formula for the development height of the water-conducting fracture zone in the fully mechanized longwall mining of thick coal seams, and a rock pillar with a thickness greater than that of the waterproof safety coal-rock pillar is retained between it and the bottom of the protected water body, thereby forming a water-proof constraint. All constraints are expressed as calculable mathematical inequalities with the width of the coal pillar or the boundary coordinates as variables.

[0066] For buildings (structures), their constraint systems typically include several inequalities:

[0067] Let the boundary coordinates of the coal pillar be the decision variable vector. The expected model function is The allowable value is The constraint can then be written as: , , , These functions Based on the probability integral method, it is and rock movement parameters It has a complex integral expression, but it can be implemented through numerical computation.

[0068] For coal pillars submerged in water, the key to water-conducting confinement is calculating the height of the water-conducting fracture zone. Statistical formulas based on actual measurements in the mining area are used, for example... ,in To accumulate thickness, , , , These are the fitting coefficients for this mining area. The constraints are: ,in It is the thickness of the waterproof and safe rock column, which is determined based on the lithology of the waterproof layer and the water head pressure.

[0069] Its working principle is to transform textual safety procedures for different protected objects into a series of functional inequalities with design variables as independent variables, enabling computers to understand and process them. The beneficial effect is the formalization and computability of safety rules, which serves as a bridge to introducing artificial intelligence and optimization algorithms into the traditional mining design field, making automated compliance review and scheme optimization possible.

[0070] As a preferred technical solution, the safety constraint inequality model in step S3 is a multi-objective and multi-constraint optimization mathematical model. Its decision variables are the retention width of each side of the coal pillar or the coordinates of the key control points of the coal pillar boundary. In addition to minimizing the coal pillar width, the objective function can also be set as maximizing the value of coal resource extraction over the entire life cycle or minimizing the impact on mining progress under the premise of satisfying safety constraints. The model is solved using linear programming, nonlinear programming or intelligent optimization algorithms. Each time optimization is triggered, the algorithm searches for the optimal solution in the feasible region based on the latest dynamic variable data and the corrected constraint parameters, and provides a comparison of schemes under different optimization objective weights.

[0071] In practice, objectives are often conflicting, such as minimizing the width of the coal pillar versus minimizing interference with the mining plan.

[0072] A multi-objective optimization model can be constructed.

[0073] Let the objective function be: ,

[0074] .

[0075] The linear weighted method is used to transform it into a single objective, namely minimizing... ,in and The weights are set by the decision-maker. When solving the model, if both the constraints and the objective are linear, the simplex method is used. However, in most cases, the predicted function... For nonlinear problems, nonlinear programming algorithms such as interior point method, sequential quadratic programming, or genetic algorithm are required.

[0076] The solver receives the latest rock displacement parameters. The system monitors and inverts the rock strata state, then substitutes the values ​​into the mathematical expressions of constraints and objectives for iterative calculations. Several different weight combinations can be preset (e.g., emphasizing resource recovery or production stability) to solve for multiple Pareto optimal solutions at once, allowing decision-makers to choose.

[0077] Its working principle is to apply multi-objective decision-making theory, incorporating multiple factors such as economy and production into a unified optimization framework, and using algorithms to efficiently explore trade-off solutions in a complex solution space. The beneficial effect is that the optimization result is no longer just a technically "optimal" solution, but a "satisfactory solution" that integrates multiple factors such as technology, economy, and production, thereby improving the scientific and comprehensive nature of decision-making and making it more in line with the actual situation of mine production.

[0078] As a preferred technical solution, the data collected by the dynamic sensing network in step S2 includes the three-dimensional coordinates and energy of rock mass fracture events obtained by the microseismic monitoring system, the surrounding rock separation and deformation rate monitored by the roadway roof delamination instrument and multi-point displacement gauge, the surface movement time series data obtained by the high-precision level and GNSS monitoring station, and the hydrological dynamic data obtained by the underground borehole seepage monitoring equipment. After preprocessing and spatiotemporal registration, these data are not only used to update the state snapshot, but also used for inversion analysis of the stress distribution evolution in front of the working face and inside the coal pillar, the fracture state of the key overburden layer, and the development height of water-conducting fractures, thereby providing a more refined and forward-looking correction basis for the right-hand side of the constraint conditions in step S3.

[0079] Spatial clustering and energy release rates of microseismic events can be used to invert stress concentration zones ahead of the working face and damage zones within the coal pillar. The stress concentration factor K can be estimated using rock mechanics principles. This K value can be used to dynamically correct calculations regarding coal pillar strength in stability constraints. Abrupt changes in roadway deformation rates may indicate the failure of key strata, affecting the dynamics of surface subsidence and potentially requiring temporary adjustments to the predicted subsidence factor q. The height of water-conducting fracture development... The parameters in the predicted formula can be calibrated online based on the real-time fracture development height observed during downhole drilling. All these inversion analyses essentially transform the monitored "phenomenon" data into quantitative information used to correct the "parameters" or "right-hand terms" of the constraint inequalities in S3, through theoretical models such as rock mechanics and mining subsidence physics.

[0080] Its working principle is to construct a quantitative mapping relationship between "monitoring data - rock mass mechanical state - design model parameters" to achieve data-driven model self-adaptation. The beneficial effect is that it greatly improves the "forward-looking" capability and refinement of the optimization model, enabling the adjustment of coal pillar width to be based on a deep understanding of the internal state of the rock strata, rather than just a response to surface displacement, thereby making more forward-looking optimization decisions.

[0081] As a preferred technical solution, the warning threshold in step S5 is set in stages according to the safety margin of the constraint. The system sets a green safety zone, a yellow warning zone, and a red alarm zone. When the monitoring data shows that the safety margin of a certain constraint drops to the yellow warning zone, the system automatically marks the risk and prompts the engineering technicians to pay attention. It can also trigger a low-priority optimization recalculation to assess potential adjustment directions. When the safety margin enters the red alarm zone, the system immediately triggers a high-priority optimization recalculation and automatically generates emergency plan suggestions such as temporary reinforcement of coal pillars, control of mining speed, or local increase of coal pillars. At the same time, the alarm and plan are pushed to relevant decision-makers for emergency approval and execution.

[0082] During system initialization, a safety margin is set for each safety constraint. Set three thresholds. For example, define... , ,as well as The system's background service continuously monitors the real-time margins calculated in step S4.

[0083] When a margin falls into the yellow zone, the system highlights the risk on the human-machine interface and can schedule a "simulation optimization" during system idle time to calculate the possible outcome of the next formal optimization if the current trend continues, but no instructions are output. When the margin falls into the red zone, indicating an imminent risk, the system interrupts the current non-urgent calculation, immediately collects the latest data, and calls a fast optimization algorithm (e.g., setting stricter iteration termination conditions) to provide a new coal pillar boundary scheme within minutes. Simultaneously, the knowledge base matches preset emergency text templates based on the risk type (deformation, water inrush, instability) and severity, generating suggestions such as "It is recommended to reinforce the coal pillar by constructing grouting anchor bolts at coordinates (x, y)." Alerts and solutions are pushed through multiple channels, including SMS and in-app messages.

[0084] Its working principle is to program and automate the "dual control" mechanism in safety management, establishing a real-time risk perception-assessment-response cycle. The beneficial effect is a significant improvement in the mine's ability to cope with dynamic risks, transforming post-event remediation into pre-event early warning and rapid in-event intervention, greatly enhancing the initiative and timeliness of production safety management.

[0085] As a preferred technical solution, this method is suitable for optimizing the protection coal pillars of roadways affected by the lateral support pressure of adjacent goaf areas. In this scenario, the engineering stability constraints in step S3 need to focus on the long-term strength stability and creep characteristics of the coal pillar. By combining the roadway deformation rate and coal pillar stress data monitored in step S2, the damage and plastic zone range of the coal pillar are dynamically evaluated, and the calculation model of the effective bearing width of the coal pillar is corrected accordingly. The optimization goal is to minimize the width of the section coal pillar as much as possible while ensuring that the cross-section meets the requirements during the roadway's use. The optimization results guide the dynamic adjustment of the working face stop line and roadway support scheme.

[0086] The core constraint of the stability of a coal pillar section is to ensure the existence and stability of its effective bearing core area.

[0087] The constraints can be modeled as follows: .in It is the width of the coal pillar. It is the calculated minimum stable width, which is the coal pillar strength. Stress ,time and the creep rate obtained by monitoring The function. This may be based on the limit equilibrium theory or the plastic zone theory. The deformation rate of the coal pillar sidewall is monitored through step S2. By combining borehole stress gauge data, the current plastic zone width and effective core zone stress of the coal pillar can be determined. Substituting this dynamic information into a function... Real-time updates The calculation.

[0088] The optimization goal is The decision variable is the width of the coal pillar. Constraints are (Dynamic) and tunnel deformation The optimal width is obtained by solving the problem. Then, the location of the stop line at the working face can be determined. If Smaller than the original design means that more coal pillar resources can be recovered.

[0089] Its working principle combines the time-varying theory of coal pillar stability in rock mechanics with online monitoring data to achieve real-time diagnosis of the "health status" of the coal pillar and precise calculation of the "dosage". The beneficial effect is that it can scientifically and dynamically determine the minimum safe size of coal pillars in a section, effectively improving the recovery rate of the mining area while ensuring the long-term availability of the roadway, directly generating huge economic benefits.

[0090] As a preferred technical solution, the executable design results output in step S5 are issued in the form of a digital mining engineering plan, a three-dimensional geological model update layer, and a construction parameter form. The system automatically compares the coal pillar boundaries of the old and new versions and calculates the increase or decrease of coal volume in the changed area and evaluates the safety and economic benefits brought about by the adjustment. All optimization decisions, triggering conditions, input data, output results, and approval processes are completely recorded in the blockchain or a database with anti-tampering function, forming a traceable and auditable complete decision chain.

[0091] Specific technical details involve system output and auditability. The system has a built-in or integrated geological survey GIS system and 3D modeling software. After the optimization results are generated, the API is automatically called to modify the coal pillar boundary lines in the mining engineering drawing in an incremental update manner, and the coal and rock strata properties of the corresponding area are updated in the 3D geological model. At the same time, a structured report is generated, explaining the reasons for the adjustment, the area of ​​the coal pillar before and after, the amount of coal released, the expected recoverable value, and changes in safety margin.

[0092] Every operation, including every optimization triggered manually, on a timed basis, or by an alert, has its inputs (monitoring data snapshot ID, constraint parameter values), optimization configuration, solution log, output results, solution issuer, and issue time written as an immutable record to a log database based on hash chaining or similar technology. Each record has a unique hash value, and any tampering will cause hash verification to fail.

[0093] Its working principle is to use digital tools to ensure the accurate transmission of design changes and to use the evidence-keeping characteristics of blockchain to ensure the authenticity and transparency of the decision-making process. The beneficial effects include the precise and seamless implementation of optimization results into production practice, and the construction of a complete trust and accountability system, ensuring that every adjustment to the protective coal pillar is "evidence-based, reasonable, and traceable," meeting the requirements of high-level enterprise management and external supervision.

[0094] As a preferred technical solution, the method is implemented through an integrated intelligent optimization system platform. This intelligent optimization system platform includes geological data management, real-time monitoring access, model calculation engine, optimization solver, visualization display and workflow driving module. The intelligent optimization system platform provides application programming interfaces to enable data integration and command interaction with the mine's geological spatial management information system, safety monitoring and control system and production execution system, so as to realize automated and intelligent closed-loop management from data perception to design optimization to production command issuance.

[0095] The system architecture is described in detail. The intelligent optimization system platform is a modular software system. The geological data management module is responsible for managing the basic data of S1. The real-time monitoring access module extracts data from various monitoring systems (S2) through protocols such as OPCUA and MQTT. The model calculation engine encapsulates the code of various prediction models and inversion algorithms in S3. The optimization solver module integrates open-source or commercial solver libraries such as IPOPT and GEKKO to perform calculations in S4. The visualization module provides a graphical interface to display monitoring curves, optimization results, safety margin dashboards, etc. The workflow-driven module connects the above modules according to preset rules (timed, event-triggered) and manages the early warning and approval processes in S5. The platform interfaces with external systems through standard APIs, such as reading coal seam floor contour lines from the geological survey system and issuing new mining boundary coordinates to the production execution system.

[0096] Its working principle employs a loosely coupled, modular design, decomposing complex business logic into various professional modules, scheduling them through a workflow engine, and collaborating with other information systems in the mine through an enterprise service bus. The beneficial effect is that it provides a fully functional, scalable, and easily integrated engineering software platform with existing mine information systems, transforming the innovative concepts of this approach into intelligent tools that are operational and usable in daily mine operations, serving as an important vehicle for technology implementation.

[0097] This application also provides an intelligent dynamic retention optimization system for mine protective coal pillars, used to implement the above-mentioned intelligent dynamic retention optimization method for mine protective coal pillars. The system includes:

[0098] The Geological and Design Fundamentals Database module is used to store and manage coal seam geological information, rock movement parameters, protected object information, and initial coal pillar design schemes.

[0099] The dynamic sensing and data fusion module is used to access and process real-time and historical data from various sensors such as surface subsidence monitoring, underground surrounding rock deformation monitoring, microseismic monitoring, and hydrological monitoring, in order to generate a spatiotemporally unified state snapshot.

[0100] The safety constraint modeling engine is used to formalize deformation constraints, watertightness constraints, and stability constraints into parameterized, machine-readable inequality models based on the type of protected object and safety standards. It can also receive state snapshot data to dynamically correct the model parameters.

[0101] The intelligent optimization solution module incorporates various mathematical programming and intelligent optimization algorithms to solve for the optimal coal pillar boundary scheme and calculate the real-time safety margin of each constraint within the feasible region defined by the safety constraint model, with the goal of minimizing the coal pillar size or maximizing resource recovery.

[0102] The design deliverables generation and publishing module is used to automatically convert optimization results into updated engineering drawings, control coordinates, and construction guidance documents, and to drive their review and publication through a workflow.

[0103] The risk warning and closed-loop control module continuously monitors safety margins and automatically triggers process re-optimization and manages emergency adjustment plans when warning thresholds are reached. All modules are connected and collaborate via a system bus and data services.

[0104] In the specific system architecture, this is a software system based on microservices or a layered architecture. The geological and design foundation database module can be a relational database. The dynamic perception and data fusion module includes a streaming data processing engine. The safety constraint modeling engine and intelligent optimization solution module are core computing services that can be deployed on high-performance computing nodes. The design result generation and publishing module includes graphics processing services and message queues. The risk warning and closed-loop control module is a rules engine and a process engine. All modules communicate through RESTful APIs or message middleware, and a data bus ensures that information flows in an orderly and efficient manner between modules. The entire system is deployed in the mine's private cloud or data center and securely interconnected with the production control system through a firewall.

[0105] Its working principle is to instantiate each functional step in the methodology as an independent, reusable software service, and to automate the entire business process through service orchestration. The beneficial effect is that it provides a standardized, productized solution, enabling the intelligent dynamic retention optimization of mine protective coal pillars to be deployed and applied in various mines as a software system, reducing implementation complexity.

[0106] It should be noted that the terms "first," "second," and similar terms used in this application specification and claims do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, "an" or "a" and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. "A plurality" or "several" indicates at least two. Unless otherwise stated, terms such as "front," "back," "left," "right," "lower," and / or "upper" are for illustrative purposes only and are not limited to a location or spatial orientation. Terms such as "comprising" or "including" indicate that the elements or objects preceding "comprising" encompass the elements or objects listed following "comprising" or "including" and their equivalents, and do not exclude other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0107] The singular forms “a,” “the,” and “the” used in this application specification and appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0108] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for intelligent dynamic retention and optimization of mine protective coal pillars, characterized in that, Includes the following steps: S1. Obtain and establish a basic design dataset, which includes basic geological and mining data of the target area, permissible critical indicators of the protected objects, and a formalized expression of the initial design boundary of the protective coal pillar and the corresponding safety constraints determined according to current specifications and measured rock movement parameters of the mining area. Among them, the basic geological and mining data includes at least the coal seam occurrence, overlying lithological stratification and key layer location, hydrological and aquitard structure, and geostress background. The measured rock movement parameters include the subsidence coefficient, the tangent of the main influence angle, the inflection point offset distance, and the influence propagation angle. S2. During the mining project, a dynamic sensing network is deployed and operated to continuously collect dynamic variable data related to the impact of mining, and the dynamic variable data is fused according to a unified spatiotemporal benchmark to generate a quantifiable snapshot reflecting the current state of the impact of mining. The dynamic variable data includes at least the real-time advance and exposed area of ​​the working face, the convergence of the surrounding rock in adjacent roadways, deformation monitoring data of key measuring points on the surface or slope, and hydrological dynamic data under the influence of mining as needed. S3. Based on the formalized expression of the safety constraints in step S1, construct a machine-readable, parameterized safety constraint inequality model, where the decision variable is the boundary position or width of the protective coal pillar to be optimized; the safety constraint inequality model integrates at least the deformation constraints derived from the allowable deformation index of the protected object, the water-proof constraints derived from the overburden failure height and the waterproof safe coal-rock pillar conditions, and the engineering stability constraints derived from the self-stabilizing conditions of the surrounding rock; the parameters of the safety constraint inequality model can be dynamically corrected based on the latest state snapshot generated in step S2; S4. Using the safety constraint inequality model constructed in step S3 and dynamically corrected as hard constraint conditions, with the optimization objective of minimizing the width of the protective coal pillar or maximizing the amount of recoverable coal, solve for the optimal solution that satisfies all constraints, output the optimal width of the protective coal pillar or the boundary adjustment amount under the current mining stage, and simultaneously calculate and output the current safety margin of each constraint as a visual risk indicator. S5. The optimization results output in step S4 are converted into executable engineering design instructions, including updated coal pillar boundary coordinates, allowable mining operations limits, and key areas for enhanced monitoring. When the monitoring data collected in step S2 indicates that the safety margin has reached the preset warning threshold, the recalculation and optimization process in steps S3 to S4 is automatically triggered to generate a new version of the retention plan, thereby realizing dynamic, closed-loop, and auditable optimization adjustment of the coal pillar boundary, and retaining complete plan versions and decision records throughout the process.

2. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The measured rock movement parameters mentioned in step S1 are a set of parameters obtained through long-term monitoring and data analysis by setting up surface and underground rock strata movement observation stations in the target mining area. This set of parameters includes at least the subsidence coefficient, the tangent of the main influence angle, the horizontal movement coefficient, the inflection point offset distance, the boundary angle, and the movement angle. When making the initial design of the protective coal pillar, this set of measured parameters is used to replace the empirical parameter values ​​in the specifications. In the continuous mining process in step S2, new rock movement and deformation data are collected through the monitoring network, and the set of measured rock movement parameters is periodically corrected and updated. The updated parameters are then fed back into the safety constraint inequality model in step S3 to achieve dynamic calibration of the key calculation coefficients of the model.

3. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The deformation constraints derived from the allowable deformation index of the protected object in step S3 are constructed separately for different types of protected objects such as buildings, railways and water bodies. For the protective coal pillars of buildings and railways, the prediction model based on the probability integral method is used to calculate the maximum deformation value of the protected object affected by mining and ensure that it is less than the critical allowable value specified in the standard, thereby forming a deformation constraint inequality. For the protective coal pillars under water bodies, the maximum height of the fracture zone is calculated according to the empirical formula for the development height of the water-conducting fracture zone in the fully mechanized mining of thick coal seams, and a rock pillar with a thickness greater than that of the waterproof safety coal-rock pillar is retained between it and the bottom of the protected water body to form a water-proof constraint. All constraints are expressed as calculable mathematical inequalities with the width of the coal pillar or the boundary coordinates as variables.

4. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1 or 3, characterized in that, The safety constraint inequality model described in step S3 is a multi-objective, multi-constraint optimization mathematical model. Its decision variables are the retention width of each side of the coal pillar or the coordinates of the key control points of the coal pillar boundary. In addition to minimizing the coal pillar width, the objective function can also be set as maximizing the value of coal resource extraction over the entire life cycle or minimizing the impact on mining progress under the premise of satisfying safety constraints. The model is solved using linear programming, nonlinear programming, or intelligent optimization algorithms. Each time optimization is triggered, the algorithm searches for the optimal solution in the feasible region based on the latest dynamic variable data and the corrected constraint parameters, and provides a comparison of schemes under different optimization objective weights.

5. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The data collected by the dynamic sensing network in step S2 includes the three-dimensional coordinates and energy of rock mass fracture events obtained by the microseismic monitoring system, the surrounding rock separation and deformation rate monitored by the roadway roof delamination instrument and multi-point displacement gauge, the surface movement time series data obtained by the high-precision level and GNSS monitoring station, and the hydrological dynamic data obtained by the underground borehole seepage monitoring equipment. After preprocessing and spatiotemporal registration, these data are not only used to update the state snapshot, but also used for inversion analysis of the stress distribution evolution in front of the working face and inside the coal pillar, the fracture state of the key overburden layer, and the development height of water-conducting fractures, thereby providing a more refined and forward-looking basis for the correction of the right-hand side of the constraint conditions in step S3.

6. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The warning threshold mentioned in step S5 is set in stages according to the safety margin of the constraints. The system sets green safety zone, yellow warning zone and red alarm zone. When the monitoring data shows that the safety margin of a certain constraint drops to the yellow warning zone, the system automatically marks the risk and prompts the engineering technicians to pay attention. It can also trigger low-priority optimization recalculation to evaluate potential adjustment directions. When the safety margin enters the red alarm zone, the system immediately triggers high-priority optimization recalculation and automatically generates emergency plan suggestions for temporary reinforcement of coal pillars, control of mining speed or local increase of coal pillars. At the same time, the alarm and plan are pushed to relevant decision-makers for emergency approval and execution.

7. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, This method is applicable to the optimization of protective coal pillars in roadways affected by lateral support pressure from adjacent goaf areas. In this scenario, the engineering stability constraints in step S3 need to focus on the long-term strength stability and creep characteristics of the coal pillar. By combining the roadway deformation rate and coal pillar stress data monitored in step S2, the damage and plastic zone range of the coal pillar are dynamically evaluated, and the calculation model of the effective bearing width of the coal pillar is corrected accordingly. The optimization goal is to minimize the width of the section coal pillar as much as possible while ensuring that the cross-section meets the requirements during the roadway's use. The optimization results guide the dynamic adjustment of the working face stop line and roadway support scheme.

8. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The executable design results output in step S5 are distributed in the form of digital mining engineering plan, updated 3D geological model layers and construction parameter forms. The system automatically compares the coal pillar boundaries of the old and new versions and calculates the increase or decrease of coal volume in the changed area and evaluates the safety and economic benefits brought about by the adjustment. All optimization decisions, triggering conditions, input data, output results and approval processes are completely recorded in the blockchain or a database with anti-tampering function, forming a complete decision chain that is traceable and auditable.

9. The intelligent dynamic retention and optimization method for mine protective coal pillars according to claim 1, characterized in that, The method is implemented through an integrated intelligent optimization system platform, which includes geological data management, real-time monitoring access, model calculation engine, optimization solver, visualization and workflow driving modules. The intelligent optimization system platform provides application programming interfaces to enable data integration and command interaction with the mine's geological spatial management information system, safety monitoring system and production execution system, so as to realize automated and intelligent closed-loop management from data perception to design optimization to production command issuance.

10. A smart dynamic retention and optimization system for mine protective coal pillars, used to implement the smart dynamic retention and optimization method for mine protective coal pillars as described in any one of claims 1-9, characterized in that, The system includes: The geological and design foundation database module is used to store and manage coal seam geological information, rock movement parameters, protected object information, and initial coal pillar design schemes. The dynamic sensing and data fusion module is used to access and process real-time and historical data from various sensors such as surface subsidence monitoring, underground surrounding rock deformation monitoring, microseismic monitoring, and hydrological monitoring, in order to generate a spatiotemporally unified state snapshot; The safety constraint modeling engine is used to formalize deformation constraints, watertightness constraints, and stability constraints into parameterized machine-readable inequality models based on the type of protected object and safety standards, and can receive state snapshot data to dynamically correct the model parameters. The intelligent optimization solution module incorporates various mathematical programming and intelligent optimization algorithms to solve for the optimal coal pillar boundary scheme and calculate the real-time safety margin of each constraint within the feasible region defined by the safety constraint model, with the goal of minimizing the coal pillar size or maximizing resource recovery. The design results generation and publishing module is used to automatically convert optimization results into updated engineering drawings, control coordinates and construction guidance documents, and drive them for review and publication through a workflow. The risk warning and closed-loop control module is used to continuously monitor the safety margin and automatically trigger process re-optimization and manage emergency adjustment plans when the warning threshold is reached. Each module is connected and collaborates with the data service via a system bus.

Images