Coal mine underground space digital twinning system based on numerical simulation

Abstract

The invention relates to the technical field of mine safety monitoring, and discloses a coal mine underground space digital twinning system based on numerical simulation, which comprises the following modules: a multi-source sensor assembly arranged in a roadway surrounding rock and used for collecting stress, displacement, anchor rod axial force and vibration data in real time; according to the coal mine underground space digital twinning system based on numerical simulation, dimension spanning from single index trend analysis to full-field stress and displacement continuous evolution process prediction is achieved, and more comprehensive stability cognition is provided. Model parameters are continuously updated through a dynamic self-calibration mechanism, a traditional method depending on fixed empirical values is replaced, and the long-term fidelity and prediction accuracy of the model under complex geological conditions are remarkably improved. The system provides a visual risk map based on simulation deduction and a multi-scene simulation function, the dependence on artificial experience is reduced, and a visual and scientific basis is provided for support scheme optimization and disaster emergency decision making.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine safety monitoring, in particular to a coal mine underground space digital twin system based on numerical simulation. BACKGROUND

[0002] Coal mine underground space stability evaluation is to evaluate the ability of underground space to resist collapse, deformation and other damage in mining activities or engineering reconstruction through geological survey, surrounding rock mechanical property test, stress state analysis and deformation monitoring, etc.

[0003] Current coal mine underground space stability evaluation mainly relies on two technical approaches. One is static numerical simulation based on previous geological data. This method cannot respond to the dynamic changes of geological and stress conditions in the mining process, resulting in large deviation between the prediction results and the actual situation. The second is to monitor through physical sensors and rely on artificial experience for safety judgment. This method can realize data visualization, but lacks the mechanism restoration ability of the continuous evolution process of the stress field and displacement field in the surrounding rock, and the prediction function is limited. The above existing technologies are difficult to realize high-precision early warning and emergency decision support. SUMMARY

[0004] To solve the problems raised in the background art, the present application provides the following technical solution: a coal mine underground space digital twin system based on numerical simulation, comprising the following modules:

[0005] A multi-source sensor assembly is arranged in the surrounding rock of the roadway for real-time acquisition of stress, displacement, anchor rod axial force and vibration data;

[0006] A data acquisition and preprocessing module is used to receive and process sensor data and output high-quality time series data sets;

[0007] A multi-source fusion algorithm module is used to reconstruct discrete sensor data into a continuous surrounding rock comprehensive state matrix;

[0008] A numerical model construction module is used to establish an initial three-dimensional surrounding rock numerical model based on geological and design parameters;

[0009] A model calibration module is used to compare measured data with simulation output and dynamically adjust model parameters through intelligent optimization algorithm;

[0010] A stability prediction module is used to predict stress and displacement evolution based on the calibrated model and generate risk probability;

[0011] A visual display module is used to display the monitoring and prediction results in three dimensions.

[0012] Preferably, the multi-source sensor assembly comprises stress sensors, displacement sensors, anchor force gauges and vibration sensors arranged in the roof, two sides and floor of the roadway.

[0013] Preferably, the data acquisition and preprocessing module adopts wavelet denoising and Kalman filtering for signal conditioning.

[0014] Preferably, the multi-source fusion algorithm module adopts a deep neural network or a Bayesian framework to realize spatio-temporal alignment and data fusion.

[0015] Preferably, the numerical model construction module adopts the finite element method or the discrete element method to establish a surrounding rock model and defines the Mohr-Coulomb constitutive relation.

[0016] Preferably, the model calibration module dynamically adjusts the elastic modulus, cohesion, internal friction angle and boundary conditions of the rock mass through a genetic algorithm or a particle swarm optimization algorithm.

[0017] Preferably, the stability prediction module integrates a reduced-order model technique to construct a lightweight surrogate model through proper orthogonal decomposition (POD) or dynamic mode decomposition (DMD), thereby improving the calculation efficiency.

[0018] Preferably, the stability prediction module supports multi-scenario deduction, including stability simulation and disaster emergency deduction under different mining and excavation speeds and support schemes.

[0019] Preferably, the surrogate model significantly improves the calculation efficiency while maintaining the prediction accuracy of key physical fields.

[0020] Preferably, the visualization display module is realized based on the Unity3D engine and supports multi-terminal access and real-time cloud map display.

[0021] Compared with the prior art, the present application provides a coal mine underground space digital twin system based on numerical simulation, which has the following beneficial effects:

[0022] 1. The coal mine underground space digital twin system based on numerical simulation realizes a dimensional leap from single index trend analysis to continuous evolution process prediction of full-field stress and displacement, providing more comprehensive stability cognition. Through a dynamic self-calibration mechanism, the model parameters are continuously updated, replacing the traditional method of relying on fixed empirical values, thereby significantly improving the long-term fidelity and prediction accuracy of the model under complex geological conditions. The system provides a visual risk map based on simulation deduction and multi-scenario simulation function, reducing the dependence on artificial experience and providing intuitive and scientific basis for support scheme optimization and disaster emergency decision-making. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 A system architecture data flow diagram is provided for the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be apparently and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those of ordinary skill in the art without creative work belong to the protection scope of the present application.

[0025] Please refer to Figure 1 The present application provides a technical solution: a coal mine underground space digital twin system based on numerical simulation, comprising the following modules:

[0026] A multi-source sensor assembly is arranged in the surrounding rock of the roadway and is used for collecting stress, displacement, anchor rod axial force and vibration data in real time.

[0027] A data acquisition and preprocessing module is used for receiving and processing sensor data and outputting high-quality time series data sets.

[0028] A multi-source fusion algorithm module is used for reconstructing discrete sensor data into a continuous surrounding rock comprehensive state matrix.

[0029] A numerical model construction module is used for establishing an initial three-dimensional surrounding rock numerical model based on geological and design parameters.

[0030] A model calibration module is used for comparing measured data and simulation output and dynamically adjusting model parameters through an intelligent optimization algorithm.

[0031] A stability prediction module is used for predicting stress and displacement evolution based on the calibrated model and generating a risk probability.

[0032] A visual display module is used for three-dimensionally displaying monitoring and prediction results.

[0033] The multi-source sensor assembly comprises stress sensors, displacement sensors, anchor rod dynamometers and vibration sensors arranged in the roof, two sides and floor of the roadway.

[0034] The data acquisition and preprocessing module uses wavelet denoising and Kalman filtering for signal conditioning.

[0035] The multi-source fusion algorithm module uses a deep neural network or a Bayesian framework to realize spatio-temporal alignment and data fusion.

[0036] The numerical model construction module uses a finite element method or a discrete element method to establish a surrounding rock model and defines a Mohr-Coulomb constitutive relationship.

[0037] The model calibration module dynamically adjusts the elastic modulus of the rock mass, cohesion, internal friction angle and boundary conditions through genetic algorithm or particle swarm optimization algorithm.

[0038] The stability prediction module integrates reduced-order model technology to construct a lightweight proxy model through proper orthogonal decomposition (POD) or dynamic mode decomposition (DMD), improving computational efficiency.

[0039] The stability prediction module supports multi-scenario deduction, including stability simulation under different mining and excavation speeds, support schemes and disaster emergency deduction.

[0040] The proxy model significantly improves computational efficiency while maintaining key physical field prediction accuracy.

[0041] The visualization module is implemented based on the Unity3D engine, supporting multi-terminal access and real-time cloud map display

[0042] I. System preparation and initial modeling phase

[0043] Geomechanical parameter acquisition and uncertainty quantification:

[0044] Collect geological exploration reports, drilling data and geological radar scan results in the target roadway area to obtain macro geological information such as rock layer distribution, inclination, thickness, etc.

[0045] Key steps: Quantify the uncertainty of rock mass mechanical parameters (such as elastic modulus E, cohesion c, internal friction angle φ, Poisson's ratio μ). Use triangular distribution or normal distribution function to describe the value range (for example, the initial value of elastic modulus E is set to [5, 15] GPa), rather than a single fixed value, to provide optimization space for subsequent parameter inversion. This overcomes the defect of absolute value of traditional model parameter.

[0046] High-fidelity initial numerical model construction:

[0047] Based on the real three-dimensional design drawings of the roadway, unstructured tetrahedral mesh is used for discretization, with a grid element number of not less than 120,000, ensuring local grid refinement in stress concentration areas such as roadway corners and support structure contact surfaces.

[0048] Use the modeling method of coupling elastic-plastic constitutive model and damage mechanics. Specifically, the rock mass constitutive relationship uses the Mohr-Coulomb criterion considering plastic hardening, and embeds a damage factor D (value range 0-1) related to plastic strain to simulate the progressive failure process of rock mass under mining stress. The initial stress field is applied by superimposing the self-weight stress field and the tectonic stress field.

[0049] II. Real-time monitoring and dynamic data fusion phase

[0050] Cooperative deployment and synchronous acquisition of heterogeneous sensor network:

[0051] A multi-source heterogeneous sensor network is constructed in the roadway. The specific deployment scheme is as follows: a monitoring section is arranged every 20 meters along the strike of the roadway; a microseismic sensor and two fiber Bragg grating displacement sensors are installed at the center of the roof of each section; one three-dimensional stress sensor is installed at the waist line of each side; and two anchor stress gauges are installed on the key anchor rods.

[0052] Innovative details: All sensors use a unified hardware clock synchronization protocol (such as PTP1588 protocol) to ensure the spatiotemporal synchronization accuracy of different physical quantity acquisition better than 1ms, laying a foundation for subsequent multi-source data spatiotemporal fusion.

[0053] Data reconstruction based on physical mechanism and AI fusion:

[0054] Data preprocessing: For the original signal, the improved wavelet packet transform is used to filter out high-frequency noise such as tunneling machine vibration; the Kalman filter algorithm based on temperature compensation is used to eliminate thermal drift for stress data.

[0055] Core innovative steps: The discrete point data after preprocessing is input into a physics-informed neural network (PINN) for physical field reconstruction. The loss function of the network not only includes the data fitting term Loss_data, but also adds the control equation term Loss_PDE (such as stress balance equation), i.e. Loss_total = Loss_data + λ·Loss_PDE. Through training the network, the final output is the continuous full-field stress tensor and displacement vector distribution covering the entire roadway, rather than just the discrete values at the measurement points.

[0056] III. Model closed-loop calibration and advanced prediction phase

[0057] Triggered dynamic calibration mechanism and multi-parameter intelligent inversion:

[0058] The system calculates the error between the monitoring data and the simulation results in real time. A double-threshold triggering mechanism is set: when the average relative error (MRE) of the displacement field is greater than 5% or the absolute error of the stress field in the key area is greater than 0.8MPa, the calibration process is automatically started.

[0059] An improved multi-objective particle swarm optimization (MOPSO) algorithm is used for parameter inversion. The target parameters for inversion include not only the traditional E, c, φ, μ, but also the evolution parameters of the damage factor D mentioned above. The optimization objective function is to minimize the error between the monitoring value and the simulation value, while maximizing the consistency between different sensor data. This process is automatically executed every 6 hours or when triggered, realizing the "dynamic growth" of the model.

[0060] Ultra-real-time multi-scenario simulation based on reduced order model (ROM):

[0061] Surrogate model construction: Collect calibrated full-order model stress-displacement field snapshot data under 200 different mining and excavation progress, support working conditions. Use proper orthogonal decomposition (POD) and radial basis function (RBF) interpolation to construct reduced order model. Specifically: use POD to extract dominant modes (first 150 modes, energy ratio > 99%), then use RBF to establish a nonlinear mapping relationship from input parameters (such as mining and excavation distance, support resistance) to POD modal coefficients.

[0062] Ultra-real-time simulation application:

[0063] Production plan optimization: Use the surrogate model to simulate the stability of surrounding rock under different mining and excavation speeds for the next 3 days within 30 seconds, quickly evaluate the effect of different support schemes (such as the influence of adjusting anchor rod spacing from 0.8m to 1.0m on roof subsidence), and output the optimal scheme.

[0064] Disaster emergency simulation: If a sudden increase in microseismic energy or a sudden drop in anchor rod stress is detected, the system will immediately trigger a water inrush or roof fall disaster simulation scenario. The surrogate model simulates the disaster evolution process (such as the expansion of the caving range and the stress change in the escape passage) within seconds, and automatically generates the optimal disaster avoidance route and emergency material allocation scheme based on the simulation results. The results are pushed to the underground terminal and the ground command center in real time through the visualization module.

[0065] Four, visualization and human-computer interaction stage

[0066] Augmented reality (AR) risk visualization:

[0067] Build a three-dimensional visualization interface through WebGL engine or Unity3D. Innovatively introduce AR technology, support underground personnel to scan the roadway through smart safety helmet or mobile terminal camera, and display stress contour, displacement contour and virtual dangerous area warning box generated by digital twin system on the real scene, greatly improve the intuitiveness of risk cognition.

[0068] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A digital twin system for underground coal mine space based on numerical simulation, characterized in that, Includes the following modules: Multi-source sensor components are deployed in the surrounding rock of the tunnel to collect stress, displacement, anchor bolt axial force and vibration data in real time; The data acquisition and preprocessing module is used to receive and process sensor data and output high-quality time-series datasets. The multi-source fusion algorithm module is used to reconstruct discrete sensor data into a continuous comprehensive state matrix of the surrounding rock. The numerical model building module establishes an initial three-dimensional surrounding rock numerical model based on geological and design parameters. The model calibration module is used to compare measured data with simulation output and dynamically adjust model parameters through intelligent optimization algorithms; The stability prediction module predicts stress and displacement evolution based on the calibrated model and generates risk probabilities. The visualization module is used to display monitoring and prediction results in three dimensions.

2. The digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The multi-source sensor assembly includes stress sensors, displacement sensors, anchor bolt force gauges, and vibration sensors arranged in the roof, sides, and floor of the tunnel.

3. The digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The data acquisition and preprocessing module uses wavelet denoising and Kalman filtering for signal conditioning.

4. The digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The multi-source fusion algorithm module uses deep neural networks or Bayesian frameworks to achieve spatiotemporal alignment and data fusion.

5. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The numerical model construction module uses the finite element method or discrete element method to establish the surrounding rock model and defines the Mohr-Coulomb constitutive relation.

6. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The model calibration module dynamically adjusts the rock mass's elastic modulus, cohesion, internal friction angle, and boundary conditions using a genetic algorithm or particle swarm optimization algorithm.

7. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The stability prediction module integrates order reduction model technology, constructs a lightweight surrogate model through intrinsic orthogonal decomposition (POD) or dynamic mode decomposition (DMD), and improves computational efficiency.

8. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The stability prediction module supports multi-scenario simulations, including stability simulations and disaster emergency simulations under different mining speeds and support schemes.

9. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The surrogate model significantly improves computational efficiency while maintaining the accuracy of predictions for key physics fields.

10. A digital twin system for underground coal mine space based on numerical simulation according to claim 1, characterized in that, The visualization module is implemented based on the Unity3D engine and supports multi-terminal access and real-time cloud map display.

Images