A digital twin method for roadway support stress field inversion based on segmented anchor bolt stress

Abstract

A digital twin method for roadway support stress field inversion based on segmented anchor bolt stress belongs to the field of roadway surrounding rock control technology. First, the axial force at the anchor bolt end and structural installation parameters of a typical roadway cross-section are measured and collected. Then, based on the segmented mechanical characteristics of the anchor bolt, the shear stress distribution at the anchorage interface is inverted from the measured axial force. Next, the shear stress and end axial force are converted into equivalent loads, and the additional stress field of a single anchor support is reconstructed using finite element analysis. A digital twin of the stress field of a group anchor support is constructed by superimposing the stress fields. Subsequently, the spatiotemporal evolution of the stress field is dynamically updated by incorporating tunneling disturbances to identify areas of stress anomaly risk. Then, the support stress field is visualized in three dimensions and displayed interactively with AR to delineate potential instability zones in the surrounding rock. Finally, risk classification warnings and dynamic optimization of support parameters are implemented. This method inverts and reconstructs measurable point data at the anchor bolt end into a digital twin of the full-scene support stress field inside the surrounding rock, enabling quantitative evaluation, three-dimensional transparent presentation, and graded early warning of the anchor bolt group reinforcement effect.

Description

Technical Field

[0001] This invention belongs to the field of tunnel surrounding rock control technology, specifically relating to a digital twin method for tunnel support stress field inversion based on segmented anchor bolt stress. Background Technology

[0002] Rock bolt support, as a core technology for controlling surrounding rock in roadways, works by using the coordinated deformation of the rock bolts and the surrounding rock to construct a reinforced load-bearing structure within the rock mass. This inhibits further loosening and damage of the surrounding rock, effectively ensuring the stability of the roadway during excavation and use. However, the actual reinforcement effect of rock bolts on the surrounding rock—the support stress field formed by the bolts within the rock mass—remains difficult to directly observe and quantitatively assess. Engineers can only indirectly infer the support effect through force measurement data at the bolt ends or the convergence deformation on the roadway surface. They cannot accurately determine whether the compressive stress zone formed by the bolt group within the surrounding rock is continuous or whether the load-bearing arch is intact. This invisible and intangible technological status quo severely restricts the scientific and refined design of rock bolt support, making it difficult to meet the complex surrounding rock control needs of deep roadways.

[0003] Currently, domestic and international research on the effectiveness of anchor bolt support mainly focuses on two directions: First, analyzing the load transfer mechanism of a single anchor bolt through indoor pull-out tests or numerical calculations, and exploring the influence of key parameters such as anchorage length, anchor bolt diameter, and anchorage medium performance on anchorage performance; Second, based on field monitoring data, acquiring information on surrounding rock deformation and anchor bolt stress through equipment such as multi-point displacement gauges and anchor bolt force gauges, and qualitatively evaluating the working effect of the support system accordingly. Although existing research has made some progress, there are still obvious limitations. Specifically, it rarely addresses the full-field support stress distribution formed by anchor bolt groups within the surrounding rock, and lacks effective technical methods to quantitatively correlate anchor bolt stress data with the stress field inside the surrounding rock. In practical engineering applications, anchor bolt force gauges can only provide single-point stress data at the end of the anchor bolt, and cannot reflect the shear stress distribution characteristics at the interface between the anchor bolt and the surrounding rock, or the entire load transfer process; multi-point displacement gauges can only obtain displacement information at a limited depth of the surrounding rock, making it difficult to infer the true stress state inside the rock mass from the displacement data. While numerical calculation methods can simulate the stress distribution at the theoretical level, they are limited by the uncertainties of rock mechanics parameters, the limitations of rock mass constitutive models, and the complex influence of the three highs and one disturbance characteristic of deep roadways. The calculated results often deviate significantly from actual field conditions, making them unreliable as a basis for support design. This technical predicament of discrete monitoring data and distorted numerical calculations has led to a long-term reliance on qualitative judgments based on engineering experience in anchor bolt support design. This results in two extreme situations: either the design is overly conservative, leading to unnecessary waste of anchor bolt materials and construction costs; or the design has safety hazards, failing to withstand the deformation and damage of complex surrounding rock in deep areas, easily causing roadway instability, anchor bolt failure, and other safety accidents, seriously threatening mine safety production. To effectively overcome the existing technical bottlenecks and construct an effective method that can accurately reflect the internal support stress field of the surrounding rock has become a key technical problem urgently needing to be solved in the field of deep roadway anchor bolt support. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a digital twin method for roadway support stress field inversion based on segmented stress of anchor bolts. This method inverts and reconstructs measurable point data at the end of the anchor bolt into a digital twin of the full-scene support stress field inside the surrounding rock, enabling quantitative evaluation, three-dimensional transparent presentation, and graded early warning of the anchor bolt group reinforcement effect.

[0005] To achieve the above objectives, the present invention provides a digital twin method for roadway support stress field inversion based on segmented anchor bolt stress, comprising the following steps:

[0006] S1: Actual force measurement and structural parameter acquisition at the end of the anchor bolt;

[0007] A high-precision force gauge was installed at the end of the anchor bolt in a typical cross-section of the roadway to obtain the actual axial force borne by the anchor bolt at the end. Simultaneously record the anchor bolt geometry and installation parameters, and preprocess the measured data;

[0008] S2: Inversion of interface shear stress distribution based on anchor bolt segmentation characteristics;

[0009] Based on the mechanical differences between the free section and the anchored section of the anchor bolt, the measured end axial force is used... Inverting the stress distribution at the anchorage section interface The mechanical rationality of the inversion results is verified by static equilibrium.

[0010] S3: Reconstruction of the additional stress field of single anchor bolt support based on equivalent nodal force;

[0011] The interface shear stress obtained from the inversion is transformed into an equivalent nodal force acting on the borehole wall in the surrounding rock, while the end axial force is also converted. As the concentrated force of the support plate, the additional stress field generated by a single anchor rod in the surrounding rock is obtained by solving the static equilibrium equation using the finite element method. ;

[0012] S4: Superposition of group anchor effect and construction of digital twin of support stress field in all scenarios;

[0013] The additional stress fields of all individual anchor bolts are algebraically superimposed to obtain the total additional support stress field of the anchor bolt group within the monitoring section. The stress uniformity index is calculated and integrated with the roadway geometric model to form a digital twin of the support stress field at the current moment.

[0014] S5: Spatiotemporal evolution modeling and dynamic update mechanism of support stress field;

[0015] Based on the dynamic changes in the anchor bolt force during the tunneling process, the stress field digital twin is updated at a set frequency. Potential risk areas are identified by calculating the stress change rate, ensuring that the model evolves synchronously with the physical tunnel.

[0016] S6: 3D visualization and augmented reality interactive presentation;

[0017] The invisible support stress field is made transparent through multiple methods, and supports timeline playback and interactive operation;

[0018] S7: Identification of potential instability zones based on yield criterion; Defining yield proximity index using the Mohr-Coulomb yield criterion based on reconstructed support stress field and surrounding rock mechanical parameters. It identifies and highlights the yield zone, intuitively reflecting the stability state of the surrounding rock;

[0019] S8: Graded early warning and dynamic optimization of support parameters;

[0020] Based on the characteristics of the support stress field and the identification results of the yield zone, a three-level risk early warning system is established, which automatically alarms and provides targeted suggestions for optimizing support parameters. It also supports dynamic adjustment of the early warning threshold according to the importance of the roadway.

[0021] As an alternative, it also includes S9: Engineering Case Database Construction and Intelligent Recommendation System;

[0022] We collect engineering cases under different geological and support conditions, build a database, use machine learning models to intelligently recommend new roadway support parameters, and achieve model self-evolution through incremental training to assist on-site decision-making.

[0023] As a preferred option, in S1, the process of measuring the force at the anchor end and collecting structural parameters is as follows:

[0024] S11: Setting up measuring points and selecting cross sections; Select no less than 5 representative anchor bolts to set up measuring points in typical cross sections of the roadway or areas with significant geological changes, ensuring that the measuring points cover different areas of the roof and both sides, and the monitoring cycle for each cross section is no less than 7 days to cover the entire process of tunneling disturbance.

[0025] S12: Install and calibrate the force gauge; install a high-precision force gauge between the end of the anchor bolt and the support plate. Before installation, calibrate the high-precision force gauge to ensure that the measurement accuracy is not less than 1% of the full scale.

[0026] S13: Collect anchor bolt parameters; synchronously record the structural and installation parameters of each anchor bolt, including the diameter. Total length Length of free segment Anchorage section length Installation parameters include preload. Pallet dimensions;

[0027] S14: Set the sampling frequency; Set the sampling frequency according to the degree of tunneling impact: no less than once every 30 minutes during the period of severe tunneling disturbance, and reduce to 1-2 times per day during the period of stable surrounding rock.

[0028] S15: Data preprocessing; outlier removal and filtering smoothing are performed on the measured end axial force data to eliminate high-frequency noise and erroneous data.

[0029] As a preferred option, in S2, the inversion process of interface shear stress distribution based on the segmented characteristics of the anchor bolt is as follows:

[0030] S21: Segmented treatment of anchor bolts; based on the structural characteristics of the anchor bolts, the anchor bolts are divided into free sections and anchored sections for separate treatment; for the free sections, the axial force remains constant along the length of the bolt and is equal to the measured axial force at the end. For the anchorage section, the axial force is determined from the measured axial force at the end. The length gradually decreases to zero along the rod;

[0031] S22: Construct the load transfer differential equation; establish the static equilibrium differential equation for the axial force and interface shear stress in the anchorage section. Boundary conditions are and ,in, The coordinates are along the anchorage section;

[0032] S23: Select the shear stress distribution pattern; based on the ratio of anchorage length to diameter. The distribution pattern of interfacial shear stress is determined by comprehensively considering the surrounding rock conditions.

[0033] according to The surrounding rock conditions are determined to be uniform, triangular / parabolic, negative exponential, or segmented distribution patterns.

[0034] when Furthermore, when the surrounding rock is intact, a uniform distribution pattern is adopted. ;

[0035] when Or, when the surrounding rock is moderately fractured, a triangular or parabolic distribution pattern should be adopted;

[0036] when Alternatively, a negative exponential decay distribution pattern may be adopted when the surrounding rock is fractured. ;

[0037] in, The shear stress at the start of the anchorage section. The attenuation coefficient;

[0038] S24: Inversion of distributed parameters; using static equilibrium conditions The distribution parameters are obtained by inverting the selected distribution pattern.

[0039] S25: Static equilibrium check; the inverted result Substituting into the integral formula, the calculated results are verified to match the measured end axial force. If the deviation is less than 1%, adjust the distribution pattern or re-invert the parameters to ensure that the inversion results satisfy the principle of conservation of mechanics.

[0040] As a preferred option, in S3, the process of reconstructing the additional stress field of a single anchor bolt support based on equivalent nodal forces is as follows:

[0041] S31: Discretized anchorage section hole wall; The anchor bolt hole wall is discretized into several micro-segments along the length of the anchorage section, and the micro-segments are two-dimensional representations of the discrete micro-elements of the surrounding rock space on the hole wall.

[0042] S32: Calculate the equivalent nodal forces; calculate the resultant force of the shear force on each infinitesimal segment. It is converted into a concentrated nodal force, with its direction parallel to the anchor bolt axis and pointing deep into the anchorage section;

[0043] S33: Apply concentrated force to the end support plate; measure the axial force at the end. As a concentrated force, it is applied to the contact surface between the pallet and the surrounding rock, with the direction perpendicular to the roadway surface and pointing towards the interior of the surrounding rock;

[0044] S34: Establish and solve the numerical model; use the finite element method to establish the numerical model and solve the static equilibrium equations. ,in To support the additional stress tensor, The equivalent volume force vector; fixed displacement constraints or zero-stress boundaries are set at the outer boundary of the model;

[0045] S35: Mesh refinement and gradual loading; Local mesh refinement is performed around the anchor hole wall, with a minimum mesh size of no more than 0.1m, and the nodal force is applied in a step-by-step, gradual manner;

[0046] S36: Outputs the additional stress field of a single anchor bolt; the solution yields the additional stress field of a single anchor bolt. This additional stress field Using discrete micro-elements of the surrounding rock space as carriers, each discrete micro-element of the surrounding rock space stores the stress tensor information caused by the anchor.

[0047] As a preferred option, in S4, the process of superimposing the group anchor effect and constructing the digital twin of the full-scenario support stress field is as follows:

[0048] S41: Algebraic superposition of stress fields of all individual anchor bolts; The additional stress fields of individual anchor bolts calculated separately for all anchor bolts within the monitoring section are algebraically superimposed, that is, on the same set of discrete micro-elements of the surrounding rock space, the stress tensors of each anchor bolt are accumulated element by element to obtain the additional stress field of the support. ,in, This represents the total number of anchor bolts within the cross-section; after stacking, the stress in the middle area between adjacent anchor bolts is enhanced, which can form a continuous compression zone;

[0049] S42: Define the anchorage zone and calculate the average stress; define the anchorage zone as a distance from the roadway surface. The ring-shaped area ( (For the total length of the anchor bolt), calculate the average stress in this area. and stress standard deviation ;

[0050] S43: Calculate the stress variation coefficient; Calculate the stress variation coefficient And based on this, quantitatively assess the uniformity of support: when This indicates that the support stress distribution is uniform and the compressive stress zone is continuous and intact; when This indicates the presence of unevenness, requiring attention to locally weak areas; when This indicates the presence of significant stress concentration or stress gaps, resulting in uneven support effectiveness and necessitating optimization of support parameters.

[0051] S44: Integrate the roadway geometric model to form a digital twin; integrate the superimposed total support stress field with the roadway three-dimensional geometric model to form a digital twin of the support stress field at the current moment with the surrounding rock spatial discrete micro-element as the basic information carrier, and realize real-time mapping with the physical anchor bolt stress state.

[0052] As a preferred option, the spatiotemporal evolution modeling and dynamic update mechanism of the support stress field in S5 is as follows:

[0053] S51: Set the update frequency; update once every 8 hours during the tunneling impact period, and once every 24 hours during the stable period;

[0054] S52: Collect new data and recalculate; input a new batch of measured stress data at the ends of anchor bolts into the system, repeat S2 to S4, and regenerate the real-time support stress field at the current moment. ;

[0055] S53: Calculate the rate of stress change; calculate the rate of stress change at each spatial point in adjacent time steps through time series analysis. ;

[0056] S54: Mark potential risk areas; when the stress change rate of a certain area exceeds a preset threshold, the area is automatically marked as a potential risk area;

[0057] S55: Establish an automatic data interface; establish an automatic data interface between monitoring data and model calculation to realize the automatic access of new data and the real-time update of digital twin model parameters.

[0058] As a preferred option, the process of 3D visualization and augmented reality interaction in S6 is as follows:

[0059] S61: Generate a 3D stress map; Generate a 3D stress cloud map on the computer, using each discrete micro-element of the surrounding rock space as the basic rendering object, and use color mapping to represent the stress magnitude: red represents the high-pressure stress area, yellow represents the medium-stress area, and blue represents the low-stress area or tensile stress area.

[0060] S62: Draw isosurfaces and stress streamlines; draw isosurfaces to show the spatial shape of a specific stress value, and draw stress streamlines to show the direction of principal stress;

[0061] S63: Build an augmented reality positioning system; use fixed feature points in the tunnel and infrared reflective markers to perform visual positioning; equip mobile terminals with supplementary lights and infrared cameras to ensure stable recognition even in low-light environments underground.

[0062] S64: Real-time AR overlay display; The reconstructed support stress field is overlaid in real time on the real tunnel image captured by the mobile terminal camera in the form of a semi-transparent heat map, and users can switch between displaying different stress components or displaying stress profiles at different depths through gesture interaction.

[0063] S65: Supports timeline playback; supports playback of historical stress field evolution along the timeline, analyzes the full-cycle response characteristics of the support system from installation to the present, and helps to trace the root causes of surrounding rock deformation and support failure.

[0064] As a preferred option, in S7, the process for identifying potential instability zones based on the yield criterion is as follows:

[0065] S71: Define the yield proximity index; Define the yield proximity index ,in, , These are the maximum and minimum principal stresses, respectively. For cohesion, It is the internal friction angle;

[0066] S72: Determine the yield state; when When the discrete micro-element of the surrounding rock is determined to have entered the yielding state; when The time is used to determine whether the unit is in an elastic or stable state;

[0067] S73: Highlight the yield zone; In the 3D visualization model, the yield unit is highlighted as a red transparent area;

[0068] S74: Obtain mechanical parameters; cohesion and internal friction angle The values ​​are preferentially determined based on field or indoor test data; when no measured data is available, values ​​are determined based on experience with rock mass classification indicators, and are continuously corrected through data assimilation and inversion in subsequent monitoring.

[0069] As a preferred option, the process of hierarchical early warning and dynamic optimization of support parameters in S8 is as follows:

[0070] S81: Establish a three-level risk assessment standard;

[0071] Level I risk occurs when the compressive stress zones of adjacent anchor bolts overlap to form a continuous compression zone, with a thickness of not less than 0.3 times the length of the anchor bolt, and a stress variation coefficient of... There is no yield zone or the yield zone is scattered and not continuous;

[0072] Level II risk is defined as local discontinuity in the compressive stress zone or a thickness less than 0.2 times the length of the anchor bolt. The yield zone extends locally but does not penetrate the anchoring layer;

[0073] Level III risk is characterized by significant fracture of the compressive stress zone, or The yield zone may penetrate the anchorage layer, or a significant tensile stress zone may appear.

[0074] S82: Automatic alarm trigger; For risk areas of level II and above, the system automatically triggers an alarm and displays it in a highlighted flashing display on the visual interface. The alarm information includes the risk level, location coordinates, and key judgment indicator values.

[0075] S83: Generate optimization suggestions; automatically generate support parameter optimization suggestions: install additional anchor bolts and increase the support spacing in stress gap areas; adjust anchor bolt preload; increase monitoring frequency in high-risk areas and simultaneously develop support reinforcement implementation plans; for permanent roadways, it is recommended to redesign support parameters;

[0076] S84: Dynamically adjust early warning thresholds; the graded early warning thresholds are dynamically adjusted according to the importance and service life of the roadway: for permanent roadways or important chambers, the risk assessment criteria are raised; for temporary roadways, the risk assessment criteria are relaxed.

[0077] Compared with the prior art, the present invention has the following advantages:

[0078] 1. Transparent reconstruction of full-field stress: By establishing a complete technical path of end stress measurement, interface shear stress inversion, equivalent load application, full-field stress reconstruction, and visualization early warning, the measurable point data at the end of the anchor bolt is transformed into an invisible full-field additional stress field of the surrounding rock. For the first time, the quantitative reconstruction and transparent presentation of the reinforcement effect of the anchor bolt group are realized, providing a scientific basis and practical tool for the design of anchor bolt support in deep roadways.

[0079] 2. An inversion method based on segmented mechanical characteristics: This method fully utilizes the force difference between the free section (constant axial force) and the anchored section (decreasing axial force) of the anchor bolt to establish a static equilibrium differential equation. Simultaneously, it effectively combines the measured end axial force with the anchor bolt structural parameters (diameter, anchorage length, etc.) to invert and solve the shear stress distribution at the anchored section interface. This method has clear physical meaning, the required parameters are easy to obtain, avoids complex on-site pull-out tests, and lowers the threshold for engineering applications.

[0080] 3. Static equivalent nodal force loading and high-precision numerical reconstruction: The interface shear stress obtained by inversion is discretized into equivalent nodal forces, and the static equivalence principle of the resultant force of all nodal forces being equal to the measured end axial force is satisfied; at the same time, a step-by-step gradual loading strategy is adopted to avoid abrupt changes in stress, and local mesh refinement (minimum mesh size ≤ 0.1m) is carried out around the anchor hole wall, which effectively ensures the accuracy and numerical stability of the additional stress field calculation.

[0081] 4. Quantitative superposition and uniformity evaluation index of anchor group effect: The additional stress fields of all single anchor bolts in the monitoring section are algebraically superimposed to obtain the total additional stress field of the anchor bolt group in the surrounding rock of the roadway, and the stress variation coefficient is introduced for the first time. As a quantitative evaluation indicator, it objectively reflects the continuity of the compressive stress zone and the synergistic effect of the anchor group, overcoming the subjectivity of traditional experience-based judgment.

[0082] 5. Spatiotemporal dynamic update and real-time mapping mechanism: A dynamic update mechanism based on measured data has been established (every 8 hours during the tunneling influence period, and every 24 hours during the stable period), supporting three update modes: timed, threshold-triggered, and manual-triggered. Potential risk areas are automatically marked by calculating the stress change rate of adjacent time steps, ensuring that the support stress state of the digital twin and the physical roadway evolves synchronously, truly reflecting the time-varying response characteristics of the support system.

[0083] 6. Instability zone identification and graded early warning based on yield criterion; the Mohr-Coulomb yield criterion is used to define the yield proximity index. It automatically identifies and highlights the yield zone. Combining the continuity, thickness, stress variation coefficient, and yield zone state of the compressive stress zone, a three-level risk early warning system (Level I / II / III) is established. Differentiated optimization suggestions are automatically generated for different risk levels (from daily inspections to installing additional anchor bolts, increasing preload, grouting reinforcement, and even redesigning support parameters), achieving a leap from qualitative judgment to quantitative classification and precise policy implementation.

[0084] 7. 3D visualization and augmented reality (AR) integration: Supports various visualization formats such as 3D stress cloud maps (red / yellow / blue mapping), isosurfaces, and stress streamlines. It can also be developed based on mobile AR applications to overlay the reconstructed support stress field as a semi-transparent heat map onto real-time images of the tunnel captured by a tablet computer. On-site technicians can intuitively see the continuity of the compressive stress zone within the rock mass and weak support areas, solving the engineering problem of invisible stress underground.

[0085] 8. Historical data backtracking and case database self-optimization: The system supports the inversion of support stress fields for sections that have undergone deformation or damage, summarizes failure modes, and extracts empirical values ​​for early warning thresholds to continuously optimize the early warning model. Simultaneously, an engineering case database covering geological parameters, anchor bolt stress, stress field morphology, and engineering response is established. Machine learning algorithms such as random forest / gradient boosting tree are used to intelligently recommend new roadway support parameters, and incremental training enables the model to self-evolve.

[0086] 9. Strong engineering adaptability; the required input data is only the axial force at the anchor bolt end and the anchor bolt structural parameters monitored by conventional methods, without the need for additional expensive sensors; the numerical model discretization method is clear and the computational efficiency is moderate; the output report is automatically generated, including cloud maps, yield zone distribution, risk level, and optimization suggestions, which is convenient for engineering technicians to archive and make decisions. The entire method has a clear process and is highly operable, applicable to the evaluation and dynamic design of anchor bolt support effects in underground engineering such as coal mines, metal mines, and tunnels.

[0087] This method inverts and reconstructs measurable point data at the end of the anchor bolt into a digital twin of the full-scene support stress field inside the surrounding rock, realizing quantitative evaluation, three-dimensional transparent presentation, and graded early warning of the anchor bolt group reinforcement effect. Attached Figure Description

[0088] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0089] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0090] like Figure 1 As shown, this invention provides a digital twin method for roadway support stress field inversion based on segmented anchor bolt stress, comprising the following steps:

[0091] S1: Actual force measurement and structural parameter acquisition at the end of the anchor bolt;

[0092] A high-precision force gauge was installed at the end of the anchor bolt in a typical cross-section of the roadway to obtain the actual axial force borne by the anchor bolt at the end. Simultaneously, the geometry and installation parameters of the anchor bolts are recorded, and the measured data are subjected to anomaly removal and filtering preprocessing to provide reliable input for subsequent shear stress inversion, thereby ensuring the input quality of the inversion.

[0093] As a preferred option, the process of measuring the force at the anchor bolt end and collecting structural parameters is as follows:

[0094] S11: Setting up measuring points and selecting cross sections; Select no less than 5 representative anchor bolts to set up measuring points in typical cross sections of the roadway or areas with significant geological changes (fault zones, lithological change zones, or typical representative cross sections) to ensure that the measuring points cover different areas of the roof and both sides, and the monitoring cycle for each cross section is no less than 7 days to cover the entire process of tunneling disturbance.

[0095] S12: Install and calibrate the force gauge; Install a high-precision force gauge between the end of the anchor bolt and the support plate. Before installation, the high-precision force gauge must be calibrated to ensure that the measurement accuracy is not less than 1% of the full scale. The installation should be firm and reliable to avoid loosening that could cause data drift.

[0096] S13: Collect anchor bolt parameters; synchronously record the structural and installation parameters of each anchor bolt, including the diameter. Total length Length of free segment Anchorage section length Installation parameters include preload. The dimensions of the tray are used for subsequent segmented modeling and parameter inversion;

[0097] S14: Set the sampling frequency; Set the sampling frequency according to the degree of tunneling impact: no less than once every 30 minutes during the period of severe tunneling disturbance, and reduce to 1-2 times per day during the period of stable surrounding rock, so as to ensure that dynamic changes can be captured while reducing data redundancy;

[0098] S15: Data preprocessing; outlier removal (e.g., using the 3σ criterion) and filtering and smoothing (e.g., low-pass filtering) are performed on the measured end axial force data to eliminate high-frequency noise and erroneous data, ensuring that the quality of the input inversion data meets the accuracy requirements.

[0099] In this technical solution, by selecting representative cross-sections and setting up measuring points on both sides of the roof, calibrating the force gauge with high precision, comprehensively collecting anchor bolt structure and installation parameters, dynamically adjusting the sampling frequency according to the degree of tunneling influence, and using statistical and filtering methods for data preprocessing, the representativeness, accuracy, completeness, and reliability of the measured end axial force data are ensured, providing a high-quality input foundation for subsequent interface shear stress inversion and stress field reconstruction.

[0100] S2: Inversion of interface shear stress distribution based on anchor bolt segmentation characteristics;

[0101] Based on the mechanical difference between the free section (constant axial force) and the anchored section (decreasing axial force) of the anchor bolt, the measured end axial force is used... Inverting the stress distribution at the anchorage section interface The mechanical rationality of the inversion results is verified by static equilibrium.

[0102] As a preferred option, the inversion process of interface shear stress distribution based on the segmented characteristics of the anchor bolt is as follows:

[0103] S21: Segmented treatment of anchor bolts; based on the structural characteristics of the anchor bolts, the anchor bolts are divided into free sections and anchored sections for separate treatment; for the free sections ( The anchor bolt is not bonded to the surrounding rock, and the axial force remains constant along the length of the bolt and is equal to the measured axial force at the end. ;

[0104] For the anchorage section ( (The origin of the coordinate system is taken as the boundary between the free section and the anchored section). The anchor rod is bonded to the surrounding rock through anchoring agent, and the axial force is measured from the end. The shear stress gradually decreases to zero along the length of the rod, and the rate of decrease depends on the distribution of the interfacial shear stress.

[0105] S22: Construct the differential equation for load transfer; based on the load transfer mechanism of anchor bolts, establish the static equilibrium differential equation for axial force and interface shear stress in the anchorage section. Boundary conditions are and (The axial force at the end of the anchorage section drops to zero), where, The coordinates are along the anchorage section (the origin is taken as the boundary point between the free section and the anchorage section).

[0106] S23: Select the shear stress distribution pattern; based on the ratio of anchorage length to diameter. The distribution pattern of interfacial shear stress is determined by comprehensively considering the surrounding rock conditions.

[0107] according to The surrounding rock conditions are determined to be uniform, triangular / parabolic, negative exponential, or segmented distribution patterns.

[0108] when Furthermore, when the surrounding rock is intact, a uniform distribution pattern is adopted. ;

[0109] when Or, when the surrounding rock is moderately fractured, a triangular or parabolic distribution pattern should be adopted;

[0110] when Alternatively, a negative exponential decay distribution pattern may be adopted when the surrounding rock is fractured. ;

[0111] in, This refers to the shear stress at the start of the anchorage section. The attenuation coefficient;

[0112] For deep high ground stress conditions, if there are multiple neutral points along the rod length, a piecewise inversion method is adopted;

[0113] S24: Inversion of distributed parameters; using static equilibrium conditions The distribution parameters (such as...) are obtained by inversion using the selected distribution model. , Taking the negative exponential distribution as an example, The specific values ​​can be determined in advance by experience or by on-site pull-out tests;

[0114] S25: Static equilibrium check; the inverted result Substituting into the integral formula, the calculated results are verified to match the measured end axial force. If the deviation is less than 1%, adjust the distribution pattern or re-invert the parameters to ensure that the inversion results satisfy the principle of conservation of mechanics.

[0115] In this technical solution, by dividing the anchor bolt into a free section and an anchored section and establishing axial force transmission models for each, constructing differential control equations based on static equilibrium, adaptively selecting various shear stress distribution modes such as uniform, triangular / parabolic, negative exponential, or segmented distributions according to the length-to-diameter ratio of the anchored section and the surrounding rock conditions, utilizing the measured axial force at the end to invert the distribution parameters and providing an explicit calculation formula for the negative exponential distribution, and employing integral verification to ensure that the deviation between the inversion results and the measured axial force is less than 1%, a high-precision inversion from easily measurable anchor bolt end point data to the interface shear stress distribution, which is difficult to measure directly, is achieved, ensuring the mechanical rationality and engineering applicability of the inversion process.

[0116] S3: Reconstruction of the additional stress field of single anchor bolt support based on equivalent nodal force;

[0117] The inverted interfacial shear stress is transformed into an equivalent nodal force acting on the borehole wall in the surrounding rock, while the end axial force is... As the concentrated force of the support plate, the additional stress field generated by a single anchor rod in the surrounding rock is obtained by solving the static equilibrium equation using the finite element method. ;

[0118] As a preferred option, the process of reconstructing the additional stress field of single anchor bolt support based on equivalent nodal forces is as follows:

[0119] S31: Discretization of the anchorage section wall; The anchorage section wall is discretized into several micro-segments along the length of the anchorage section. The micro-segments are two-dimensional representations of the discrete micro-elements of the surrounding rock space on the hole wall, and are the basis for subsequent three-dimensional micro-element division. The length of the micro-segments is 0.1 to 0.2 m to ensure the continuous distribution of shear stress and that the numerical calculation accuracy meets the requirements.

[0120] S32: Calculate the equivalent nodal forces; calculate the resultant force of the shear force on each infinitesimal segment. It is converted into a concentrated nodal force, with its direction parallel to the anchor bolt axis and pointing deep into the anchorage section;

[0121] Within the free section, there is no shear stress and no nodal forces are applied.

[0122] S33: Apply concentrated force to the end support plate; measure the axial force at the end. As a concentrated force, it is applied to the contact surface between the support plate and the surrounding rock, with the direction perpendicular to the roadway surface and pointing into the surrounding rock. This force, together with the nodal forces distributed in the anchoring section, constitutes a static equilibrium system.

[0123] S34: Establish and solve the numerical model; use the finite element method to establish the numerical model and solve the static equilibrium equations. ,in To support the additional stress tensor, The equivalent volume force vector (represented as nodal forces in the discrete model) is used; fixed displacement constraints or zero-stress boundaries are set at the outer boundary of the model (3 to 5 times the width from the center of the roadway);

[0124] S35: Mesh refinement and gradual loading; Local mesh refinement is performed around the anchor hole wall, with a minimum mesh size of no more than 0.1m, and the nodal forces are applied in a step-by-step gradual manner (e.g., in 3 to 5 steps) to avoid stress abrupt changes that could lead to non-convergence of the calculation;

[0125] S36: Outputs the additional stress field of a single anchor bolt; the solution yields the additional stress field of a single anchor bolt. This additional stress field Using discrete micro-elements of the surrounding rock space as a carrier, each discrete micro-element stores the stress tensor information caused by the anchor. This stress field intuitively reflects the support influence range of a single anchor: compressive stress bubbles or stress cones are formed around the anchored section, and their shape and strength depend on the anchor diameter, anchorage length, and the interface shear stress distribution characteristics obtained by inversion; the stress response is weaker at the corresponding position of the free section due to the absence of shear stress.

[0126] In this technical solution, by discretizing the hole wall of the anchoring section into micro-segments and calculating the equivalent nodal force, without applying nodal force to the free section, and applying concentrated force to the end plate, combined with a finite element numerical model and reasonable external boundary constraints, and by adopting a local mesh refinement of the hole wall (minimum mesh size ≤ 0.1m) and a step-by-step gradual loading strategy, the additional stress field generated by a single anchor rod in the surrounding rock can be reconstructed with high precision and stability. The compressive stress bubble / stress cone morphology around the anchoring section and the weak stress response of the free section are clearly presented, realizing the quantitative and visual characterization of the actual reinforcement effect of the anchor rod on the surrounding rock.

[0127] S4: Superposition of group anchor effect and construction of digital twin of support stress field in all scenarios;

[0128] The additional stress fields of all individual anchor bolts are algebraically superimposed to obtain the total additional stress field of the anchor bolt group within the monitoring section, and the stress uniformity index (coefficient of variation) is calculated. It is then integrated with the roadway geometry model to form a digital twin of the support stress field at the current moment.

[0129] As a preferred option, the process of superimposing the group anchor effect and constructing a digital twin of the support stress field in the entire scenario is as follows:

[0130] S41: Algebraic superposition of stress fields of all individual anchor bolts; The additional stress fields of individual anchor bolts calculated separately for all anchor bolts within the monitoring section are algebraically superimposed, that is, on the same set of discrete micro-elements of the surrounding rock space, the stress tensors of each anchor bolt are accumulated element by element to obtain the additional stress field of the support. ,in, This represents the total number of anchor bolts within the cross-section; after stacking, the stress in the middle area between adjacent anchor bolts is enhanced, which can form a continuous compression zone (bearing arch).

[0131] S42: Define the anchorage zone and calculate the average stress; define the anchorage zone as a distance from the roadway surface. The ring-shaped area ( (For the total length of the anchor bolt), calculate the average stress in this area. and stress standard deviation This is used for subsequent uniformity evaluation;

[0132] S43: Calculate the stress variation coefficient; Calculate the stress variation coefficient And based on this, quantitatively assess the uniformity of support: when This indicates that the support stress distribution is uniform, the group anchoring effect is fully utilized, and the compressive stress zone is continuous and intact; when This indicates a certain degree of unevenness, requiring attention to locally weak areas; when This indicates the presence of significant stress concentration or stress gaps, resulting in uneven support effectiveness and necessitating optimization of support parameters.

[0133] S44: A digital twin is formed by integrating the tunnel geometry model; the superimposed total support stress field is integrated with the tunnel's three-dimensional geometry model to form a digital twin of the current support stress field, using discrete micro-elements of the surrounding rock space as the basic information carrier, achieving real-time mapping with the physical anchor stress state. This integration of the total support stress field with the tunnel geometry model constitutes the current support stress field digital twin, which is mapped in real-time to the anchor stress state in the physical world.

[0134] In this technical solution, the quantitative characterization of the synergistic effect of the group anchors and the explicit reconstruction of the compressive stress zone (bearing arch) are achieved by algebraically superimposing the additional stress fields of all individual anchors within the monitoring section; furthermore, the mean stress and standard deviation of the anchorage zone are defined, and the stress variation coefficient is introduced. As a quantitative evaluation indicator, a uniform ( ), there is unevenness ( Significantly uneven () The three-level hierarchical evaluation standard provides an objective and quantifiable basis for judging the support effect. Finally, the superimposed total support stress field is integrated with the three-dimensional geometric model of the roadway to form a digital twin that is mapped in real time with the stress state of the physical anchor bolts, realizing a leap from the local effect of a single anchor bolt to the transparent presentation of the stress field of the whole scene group anchor support.

[0135] S5: Spatiotemporal evolution modeling and dynamic update mechanism of support stress field;

[0136] Based on the dynamic changes in the anchor bolt force during the tunneling process, the stress field digital twin is updated at a set frequency. Potential risk areas are identified by calculating the stress change rate, ensuring that the model evolves synchronously with the physical tunnel.

[0137] During tunnel excavation, as the excavation face advances and the surrounding rock deforms, the stress at the ends of the anchor bolts changes dynamically. As a preferred method, the spatiotemporal evolution modeling and dynamic updating mechanism of the support stress field is as follows:

[0138] S51: Set the update frequency; update once every 8 hours during the tunneling influence period and once every 24 hours during the stable period. The specific frequency can be adjusted appropriately according to the actual deformation rate on site and the monitoring system capability.

[0139] S52: Collect new data and recalculate; input a new batch of measured stress data at the ends of anchor bolts into the system, repeat S2 to S4, and regenerate the real-time support stress field at the current moment. ;

[0140] S53: Calculate the rate of stress change; calculate the rate of stress change at each spatial point in adjacent time steps through time series analysis. To reflect the time-varying nature of the support stress field;

[0141] S54: Mark potential risk areas; when the stress change rate of a certain area exceeds a preset threshold (such as continuous increase or abnormal decay, the example threshold is 0.1MPa / h), the area is automatically marked as a potential risk area; the threshold can be dynamically adjusted according to the importance of the roadway and the allowable deformation.

[0142] S55: Establish an automatic data interface; establish an automatic data interface between monitoring data and model calculation to realize the automatic access of new data and the real-time update of digital twin model parameters; the system supports three modes: timed update (at a set time interval), threshold-triggered update (automatically triggered when the change in anchor force exceeds the set value), and manual-triggered update (manually triggered by on-site personnel as needed).

[0143] This dynamic update mechanism ensures that the digital twin evolves synchronously with the physical tunnel, truly reflecting the time-varying response characteristics of the support system.

[0144] In this technical solution, by setting differentiated update frequencies (every 8 hours during the tunneling influence period and every 24 hours during the stable period) and supporting on-demand adjustments, combined with the automatic access of new data and the repeated calculations in steps 2-4, the synchronous evolution of the digital twin of the support stress field and the physical roadway is achieved; the stress change rate is calculated through time series analysis. It also sets preset thresholds (such as 0.1 MPa / h) to automatically mark potential risk areas, providing a quantitative basis for dynamic early warning; at the same time, it establishes an automatic interface between monitoring data and model calculation, supporting three update modes: timed, threshold-triggered, and manual-triggered, which significantly improves the timeliness, automation, and engineering adaptability of the model, and truly reflects the time-varying response characteristics of the support system.

[0145] S6: 3D visualization and augmented reality interactive presentation;

[0146] The invisible support stress field is made transparent through various methods such as 3D cloud maps, isosurfaces, streamlines and augmented reality (AR) overlay, and supports timeline playback and interactive operation, which makes it easier for technicians to intuitively understand the stress distribution inside the surrounding rock.

[0147] As a preferred approach, the process of interactive presentation of 3D visualization and augmented reality is as follows:

[0148] S61: Generate 3D stress map; Generate 3D stress cloud map on the computer, using discrete micro-elements of each surrounding rock space as the basic rendering object, and use color mapping to represent the stress magnitude: red represents high-pressure stress zone (effective support zone), yellow represents medium-stress zone, and blue represents low-stress zone or tensile stress zone; Supports rotation, scaling, and sectioning operations, and can display the stress distribution details of any cross section.

[0149] S62: Draw isosurfaces and stress streamlines; draw isosurfaces to show the spatial shape of a specific stress value (such as 0.5 times the average stress), and draw stress streamlines to show the direction of the principal stress, thereby intuitively reflecting the continuity of the compressive stress zone and the stress transmission path;

[0150] S63: Build an augmented reality positioning system; In order to solve the engineering problems of insufficient light and scarce feature points in the mine, use fixed feature points such as anchor bolt brackets, cable hooks, and signs in the roadway, as well as infrared reflective markers (spaced no more than 5m apart) for visual positioning; The mobile terminal is equipped with a supplementary light and an infrared camera to ensure stable recognition in the low-light environment in the mine.

[0151] S64: Real-time AR overlay display; Develop a mobile AR application to overlay the reconstructed support stress field in a semi-transparent heatmap format (transparency set to 40%–60% to ensure no obscuring of the actual structure) onto real tunnel footage captured by a mobile terminal camera in real time. On-site technicians can scan the tunnel using a tablet to intuitively see the stress distribution within the rock mass, identify whether the compressive stress zone is continuous, and whether there are weak support areas. Simultaneously, users can use gestures to switch between displaying different stress components (such as maximum principal stress, minimum principal stress, and shear stress) or displaying stress profiles at different depths.

[0152] S65: Supports timeline playback; supports playback of historical stress field evolution along the timeline, analyzes the full-cycle response characteristics of the support system from installation to the present, and helps to trace the root causes of surrounding rock deformation and support failure.

[0153] This technical solution generates a 3D stress cloud map (red / yellow / blue mapping stress magnitude) and supports rotation, scaling, and sectioning. It then draws isosurfaces and stress streamlines to display the spatial morphology of compressive stress zones and the principal stress transmission path, achieving a high-dimensional and intuitive expression of the support stress field. Simultaneously, addressing the engineering challenges of insufficient underground lighting and scarce feature points, an augmented reality positioning system is built using fixed feature points and infrared reflective markers (spacing ≤5m) in conjunction with supplementary lighting and infrared cameras. A mobile AR application is developed to overlay a semi-transparent heat map (40%–60% transparency) onto the actual tunnel image in real time. This allows on-site technicians to directly visualize the stress distribution within the rock mass, identify the continuity of compressive stress zones and weak support areas via tablet computers, and supports gesture interaction to switch stress components and stress profiles. Furthermore, the timeline playback function traces the full-cycle response characteristics of the support system, assisting in the analysis of deformation and failure root causes. This step transforms abstract stress data into an interactive, transparent, and traceable immersive visualization experience, significantly improving on-site personnel's perception of the surrounding rock support status and decision-making efficiency.

[0154] S7: Identification of potential instability zones based on yield criterion; Defining yield proximity index using the Mohr-Coulomb yield criterion based on reconstructed support stress field and surrounding rock mechanical parameters. It identifies and highlights the yield zone, intuitively reflecting the stability state of the surrounding rock;

[0155] As a preferred approach, the process for identifying potential instability zones based on the yield criterion is as follows:

[0156] S71: Define the yield proximity index; Define the yield proximity index ,in, , These represent the maximum and minimum principal stresses (positive tension, negative compressive stress), respectively. For cohesion, It is the internal friction angle;

[0157] S72: Determine the yield state; when The criterion determines that the discrete micro-element of the surrounding rock has entered the yielding state; this criterion comprehensively considers the combined effect of normal stress and shear stress.

[0158] S73: Highlight the yield zone; In the 3D visualization model, the yield unit is highlighted as a red transparent area. The range and shape of the yield zone can intuitively reflect the location and scale of possible failure of the surrounding rock.

[0159] S74: Obtain mechanical parameters; cohesion and internal friction angle The values ​​are preferentially determined based on field or indoor test data; when no measured data is available, values ​​are determined based on rock mass classification indicators (such as the geological strength index GSI) and continuously corrected through data assimilation and inversion in subsequent monitoring.

[0160] In this technical solution, a yield proximity index based on the Mohr-Coulomb criterion is defined. (Taking into account the maximum and minimum principal stresses, cohesion, and internal friction angle), a system was established. The quantitative criteria for determining yield enable the identification of the mechanical driving force of potential instability zones in the surrounding rock; furthermore, the yielding element is highlighted as a red transparent area in the 3D model, making the location and scale of failure immediately apparent; simultaneously, mechanical parameters... , This method prioritizes the use of measured data. When measured data is unavailable, empirical values ​​can be derived from rock mass classification indices such as GSI and corrected through data assimilation and inversion, balancing accuracy and engineering feasibility. This approach elevates traditional empirical assessment of surrounding rock stability to a quantitative and visual evaluation based on yield criteria, providing a reliable mechanical basis for accurately identifying weak support zones.

[0161] S8: Graded early warning and dynamic optimization of support parameters;

[0162] Based on the characteristics of the support stress field and the identification results of the yield zone, a three-level risk early warning system is established, which automatically alarms and provides targeted suggestions for optimizing support parameters. It also supports dynamic adjustment of the early warning threshold according to the importance of the roadway.

[0163] As a preferred option, the process of hierarchical early warning and dynamic optimization of support parameters is as follows:

[0164] S81: Establish a three-level risk assessment standard;

[0165] Level I risk (low risk) occurs when the compressive stress zones of adjacent anchor bolts overlap to form a continuous compression zone, the thickness of which is not less than 0.3 times the length of the anchor bolt, and the stress variation coefficient is [not specified]. There is no yield zone or the yield zone is scattered and not continuous;

[0166] Level II risk (medium risk) is characterized by local discontinuities in the compressive stress zone (stress gaps between adjacent anchor bolts) or a thickness less than 0.2 times the length of the anchor bolt. The yield zone extends locally but does not penetrate the anchoring layer;

[0167] Level III risk (high risk) is characterized by obvious fracture of the compressive stress zone (the width of the blank area exceeds 30% of the anchor bolt spacing), or The yield zone may penetrate the anchorage layer, or a significant tensile stress zone may appear.

[0168] S82: Automatic alarm trigger; For risk areas of level II and above, the system automatically triggers an alarm and displays it in a highlighted flashing display on the visual interface. The alarm information includes the risk level, location coordinates, and key judgment indicator values.

[0169] S83: Generate optimization suggestions; automatically generate support parameter optimization suggestions: install additional anchor bolts and increase support spacing in stress-deficient areas; adjust anchor bolt preload (recommend preload increment based on current stress level); increase monitoring frequency in high-risk areas and simultaneously develop support reinforcement implementation plans; if necessary, such as after two consecutive iterations of calculation using the digital twin model and the risk level has not decreased, directly adopt reinforcement measures such as grouting; for permanent roadways, it is recommended to redesign support parameters;

[0170] Specifically, when the risk level is Level I (low risk), the following recommendations are given: maintain the existing support scheme and conduct routine inspections at the usual frequency; stress field data can be selectively recorded for subsequent data analysis, without the need for active reinforcement measures.

[0171] When the risk level is Level II (medium risk), the following recommendations are provided: ① Install additional anchor bolts in stress gaps or discontinuous locations in compressive stress zones, and increase the spacing between supports (it is recommended to reduce the spacing to 70%–80% of the original spacing); ② Appropriately increase the preload of adjacent anchor bolts (the recommended preload increment is based on the current stress level, generally 20%–30% of the original preload); ③ Increase the monitoring frequency of this risk area and its affected area (from the usual once every 24 hours to once every 8 hours), and closely track the stress evolution trend; ④ If necessary, grout the locally fractured areas to improve the overall integrity of the surrounding rock.

[0172] When the risk level is Level III (high risk), the following recommendations are provided: ① Immediately stop tunneling or mining operations and reinforce the high-risk area with temporary support; ② In sections where the compressive stress zone is clearly fractured or the yield zone is penetrated, re-install high-strength long anchor bolts (or anchor cables) and increase the spacing to 50%–60% of the original spacing; ③ Significantly increase the preload (recommended to increase to 1.5–2.0 times the original preload) and use yield support to accommodate large deformations; ④ Reinforce the high-risk area with full-section grouting or chemical grouting to form an integral load-bearing structure; ⑤ For permanent roadways (service life > 10 years) or important chambers, the support parameters must be redesigned, the surrounding rock stability reassessed, and a special design plan submitted; ⑥ List the risk section as a key monitoring target, increase the monitoring frequency to once every 2 hours, and install real-time displacement and stress sensing early warning devices if necessary.

[0173] S84: Dynamically adjust early warning thresholds; tiered early warning thresholds are dynamically adjusted based on the importance and service life of the roadway: for permanent roadways (service life > 10 years) or important chambers, the risk assessment criteria are appropriately tightened (e.g., This is considered medium risk, and the required thickness of the compressive stress zone is increased to 0.4 times the anchor bolt length; for temporary roadways (service life < 2 years), the risk assessment criteria can be appropriately relaxed to reduce support costs (e.g., (This is considered high-risk).

[0174] In this technical solution, the continuity, thickness, and stress variation coefficient of the compressive stress zone are comprehensively considered. Based on the yield zone status, a three-tiered risk quantification standard (Levels I / II / III) was established, automatically triggering alarms and highlighting risk areas. Precise optimization suggestions were provided for different risk levels: low risk maintained the status quo; medium risk implemented measures such as additional anchor bolts, increased preload, intensified monitoring, and localized grouting; high risk required immediate work stoppage, temporary reinforcement, significantly increased preload, or even full-section grouting and support redesign. Simultaneously, the system supports dynamic adjustment of early warning thresholds based on the roadway's service life (tightening for permanent roadways and widening for temporary roadways). This method upgrades traditional empirical support evaluation into a quantitative, graded, automatic alarm, tiered policy implementation, and threshold-adaptive control system, significantly improving the scientific rigor of support effectiveness assessment and the targeted nature of on-site treatment.

[0175] S9: Engineering Case Database Construction and Intelligent Recommendation System;

[0176] We collect engineering cases under different geological and support conditions, build a database, use machine learning models to intelligently recommend new roadway support parameters, and achieve model self-evolution through incremental training to assist on-site decision-making.

[0177] As a preferred approach, the process of building an engineering case database and developing an intelligent recommendation system is as follows:

[0178] S91: Establish an engineering case database; establish an engineering case database covering different geological conditions, different burial depths, and different support parameters. Each case records: original geological parameters (rock uniaxial compressive strength, RQD index, rock mass integrity coefficient, groundwater conditions), tunnel geometric parameters (cross-sectional shape, size, burial depth), measured data sequence of anchor bolt stress, inverted interface shear stress distribution pattern, reconstructed support stress field and compressive stress zone morphology parameters, risk classification results, and actual engineering response (stability / deformation / failure).

[0179] S92: Training the machine learning model; Training the machine learning prediction model based on historical data (using random forest or gradient boosting tree algorithm), with input features including: uniaxial compressive strength of rock, RQD index, burial depth, empirical range of lateral pressure coefficient, and tunnel width-to-height ratio; outputs include: recommended anchor spacing, length, preload, expected thickness and continuity evaluation of the compressive stress zone, and prediction confidence interval;

[0180] S93: Ensure sufficient training sample size; initial training requires at least 50 high-quality historical data from engineering cases. If the sample size is insufficient, synthetic data or transfer learning should be used to supplement it.

[0181] S94: Achieve model self-evolution; perform incremental training regularly using newly accumulated engineering cases (e.g., once per quarter) to achieve model self-evolution and continuous improvement in prediction accuracy;

[0182] S95: Assisting on-site decision-making; when the model outputs recommended parameters, it also includes reference information on similar engineering cases, allowing on-site engineers to make comprehensive decisions based on their own experience, avoiding complete reliance on the black-box model.

[0183] This technical solution establishes a multi-dimensional engineering case database encompassing geological parameters, tunnel geometry, anchor bolt stress, shear stress distribution, stress field morphology, and engineering response. A machine learning prediction model is trained based on random forest or gradient boosting tree algorithms. Inputting rock strength, RQD, burial depth, lateral pressure coefficient, and tunnel width-to-height ratio, the model outputs recommended anchor bolt spacing, length, preload, expected compressive stress zone thickness, continuity evaluation, and confidence intervals. This represents a leap from "experience-based analogy" to "data-driven intelligent recommendation." The solution also specifies that initial training requires at least 50 high-quality samples, supplemented by synthetic data or transfer learning. Regular incremental training enables model self-evolution and continuous accuracy improvement. Similar engineering cases are included when outputting recommended parameters for comprehensive decision-making by on-site engineers. This method significantly reduces reliance on human experience, improves the scientific rigor and adaptability of support parameter design under different geological conditions, and demonstrates good generalization ability and practical engineering value.

[0184] System integration and functional expansion: In actual engineering deployments, this method can be integrated into existing mine safety monitoring and control systems (such as roadway surrounding rock dynamic monitoring platforms, mine pressure monitoring systems, etc.) as a software module, operating as a dedicated functional module for roadway support effectiveness evaluation. This module has the following extended functions:

[0185] Historical data retrospective analysis: It supports retrieving anchor bolt end force data for any historical period and reconstructing the digital twin of the support stress field in reverse along the time axis for that period, thereby performing post-event inversion analysis on sections that have undergone excessive deformation or support failure.

[0186] Failure Mode Summary and Threshold Self-Optimization: Based on the inversion results of multiple historical failure cases, typical failure modes of support stress fields under different geological conditions (such as local fracture of compressive stress zone, penetration of anchorage layer in yield zone, etc.) are summarized, and statistically significant early warning threshold empirical values ​​are extracted from them to dynamically correct the graded early warning criteria in step 8, so as to realize the continuous self-optimization of the early warning model.

[0187] Automatic Engineering Report Generation: After completing the stress field reconstruction and risk assessment at the current moment, the system can automatically output a standardized engineering report. The report includes: a 3D cloud map of the additional stress field of the support (with color coding and cross-sectional views), a distribution map of the yield zone of the surrounding rock (marking the volume percentage of yield units), a detailed table of risk level determination for each risk section / area, and a list of support parameter optimization suggestions generated based on the rule base. The report can be exported as a PDF or image for engineering technicians to archive, review, or use as a basis for support quality acceptance.

[0188] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A digital twin method for roadway support stress field inversion based on segmented anchor bolt stress, characterized in that, Includes the following steps: S1: Actual measurement of anchor bolt end force and acquisition of structural parameters; High-precision force gauges are installed at the anchor bolt ends in typical roadway cross-sections to obtain the actual axial force borne by the anchor bolt at the ends. ; S2: Inversion of interface shear stress distribution based on anchor bolt segment characteristics; based on the mechanical differences between the free section and the anchored section of the anchor bolt, using the measured end axial force... Inverting the stress distribution at the anchorage section interface ; S3: Reconstruction of the additional stress field of single anchor bolt support based on equivalent nodal forces; transforming interface shear stress into equivalent nodal forces acting on the surrounding rock borehole wall, while simultaneously converting end axial forces. As the concentrated force of the support plate, the additional stress field of a single anchor rod is obtained by solving the static equilibrium equation using the finite element method. ; S4: Superposition of anchor group effect and construction of digital twin of support stress field in the whole scene; algebraically superimpose all additional stress fields to obtain the total additional stress field of anchor group in the monitoring section, and form a digital twin of support stress field at the current moment. S5: Spatiotemporal evolution modeling and dynamic updating mechanism of support stress field; based on the dynamic changes of anchor bolt force during tunneling, update the digital twin of stress field and identify potential risk areas by calculating stress change rate; S6: 3D visualization and augmented reality interactive presentation; making the invisible support stress field transparent; S7: Identification of potential instability zones based on yield criterion; Defining yield proximity index using the Mohr-Coulomb yield criterion based on reconstructed support stress field and surrounding rock mechanical parameters. Identify and highlight the yield zone; S8: Graded early warning and dynamic optimization of support parameters; Based on the characteristics of the support stress field and the identification results of the yield zone, a three-level risk early warning system is established, which automatically alarms and provides targeted suggestions for optimizing support parameters.

2. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, It also includes S9: Engineering Case Database Construction and Intelligent Recommendation System; We collect engineering cases under different geological and support conditions, build a database, use machine learning models to intelligently recommend support parameters for new tunnels, and achieve model self-evolution through incremental training to assist on-site decision-making.

3. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S1, the process of measuring the force at the anchor end and collecting structural parameters is as follows: S11: Setting up measuring points and selecting cross sections; Select no less than 5 representative anchor bolts to set up measuring points in typical cross sections of the roadway or areas with significant geological changes, ensuring that the measuring points cover different areas of the roof and both sides, and the monitoring cycle for each cross section is no less than 7 days to cover the entire process of tunneling disturbance. S12: Install and calibrate the force gauge; install a high-precision force gauge between the end of the anchor bolt and the support plate. Before installation, calibrate the high-precision force gauge to ensure that the measurement accuracy is not less than 1% of the full scale. S13: Collect anchor bolt parameters; synchronously record the structural and installation parameters of each anchor bolt, including the diameter. Total length Length of free segment Anchorage section length Installation parameters include preload. Pallet dimensions; S14: Set the sampling frequency; Set the sampling frequency according to the degree of tunneling impact: no less than once every 30 minutes during the period of severe tunneling disturbance, and reduce to 1-2 times per day during the period of stable surrounding rock. S15: Data preprocessing; outlier removal and filtering smoothing are performed on the measured end axial force data to eliminate high-frequency noise and erroneous data.

4. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S2, the inversion process of interface shear stress distribution based on the segmented characteristics of the anchor bolt is as follows: S21: Segmented treatment of anchor bolts; based on the structural characteristics of the anchor bolts, the anchor bolts are divided into free sections and anchored sections for separate treatment; for the free sections, the axial force remains constant along the length of the bolt and is equal to the measured axial force at the end. For the anchorage section, the axial force is determined from the measured axial force at the end. The length gradually decreases to zero along the rod; S22: Construct the load transfer differential equation; establish the static equilibrium differential equation for the axial force and interface shear stress in the anchorage section. Boundary conditions are and ,in, The coordinates are along the anchorage section; S23: Select the shear stress distribution pattern; based on the ratio of anchorage length to diameter. The distribution pattern of interfacial shear stress is determined by comprehensively considering the surrounding rock conditions. according to The surrounding rock conditions are determined to be uniform, triangular / parabolic, negative exponential, or segmented distribution patterns. when Furthermore, when the surrounding rock is intact, a uniform distribution pattern is adopted. ; when Or, when the surrounding rock is moderately fractured, a triangular or parabolic distribution pattern should be adopted; when Alternatively, a negative exponential decay distribution pattern may be adopted when the surrounding rock is fractured. ; in, This refers to the shear stress at the start of the anchorage section. The attenuation coefficient; S24: Inversion of distributed parameters; using static equilibrium conditions The distribution parameters are obtained by inverting the selected distribution pattern. S25: Static equilibrium check; the inverted result Substituting into the integral formula, the calculated results are verified to match the measured end axial force. If the deviation is less than 1%, adjust the distribution pattern or re-invert the parameters to ensure that the inversion results satisfy the principle of conservation of mechanics.

5. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S3, the process of reconstructing the additional stress field of a single anchor bolt support based on equivalent nodal forces is as follows: S31: Discretized anchorage section hole wall; The anchor bolt hole wall is discretized into several micro-segments along the length of the anchorage section, and the micro-segments are two-dimensional representations of the discrete micro-elements of the surrounding rock space on the hole wall. S32: Calculate the equivalent nodal forces; Calculate the resultant force of the shear force on each infinitesimal segment. It is converted into a concentrated nodal force, with its direction parallel to the anchor bolt axis and pointing deep into the anchorage section; S33: Apply concentrated force to the end support plate; measure the axial force at the end. As a concentrated force, it is applied to the contact surface between the pallet and the surrounding rock, with the direction perpendicular to the roadway surface and pointing towards the interior of the surrounding rock; S34: Establish and solve the numerical model; use the finite element method to establish the numerical model and solve the static equilibrium equations. ,in To support the additional stress tensor, The equivalent volume force vector; fixed displacement constraints or zero-stress boundaries are set at the outer boundary of the model; S35: Mesh refinement and gradual loading; Local mesh refinement is performed around the anchor hole wall, with a minimum mesh size of no more than 0.1m, and the nodal force is applied in a step-by-step, gradual manner; S36: Outputs the additional stress field of a single anchor bolt; the solution yields the additional stress field of a single anchor bolt. This additional stress field Using discrete micro-elements of the surrounding rock space as carriers, each discrete micro-element of the surrounding rock space stores the stress tensor information caused by the anchor.

6. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S4, the process of superimposing the group anchor effect and constructing a digital twin of the full-scenario support stress field is as follows: S41: Algebraic superposition of stress fields of all individual anchor bolts; The additional stress fields of individual anchor bolts calculated separately for all anchor bolts within the monitoring section are algebraically superimposed, that is, on the same set of discrete micro-elements of the surrounding rock space, the stress tensors of each anchor bolt are accumulated element by element to obtain the additional stress field of the support. ,in, This represents the total number of anchor bolts within the cross-section; after stacking, the stress in the middle area between adjacent anchor bolts is enhanced, which can form a continuous compression zone; S42: Define the anchorage zone and calculate the average stress; define the anchorage zone as a distance from the roadway surface. Calculate the average stress within the annular region. and stress standard deviation ; S43: Calculate the stress variation coefficient; Calculate the stress variation coefficient And based on this, quantitatively assess the uniformity of support: when This indicates that the support stress distribution is uniform and the compressive stress zone is continuous and intact; when This indicates the presence of unevenness, requiring attention to locally weak areas; when This indicates the presence of significant stress concentration or stress gaps, resulting in uneven support effectiveness and necessitating optimization of support parameters. S44: Integrate the roadway geometric model to form a digital twin; integrate the superimposed total support stress field with the roadway three-dimensional geometric model to form a digital twin of the support stress field at the current moment with the surrounding rock spatial discrete micro-element as the basic information carrier, and realize real-time mapping with the physical anchor bolt stress state.

7. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S5, the spatiotemporal evolution modeling and dynamic update mechanism of the support stress field is as follows: S51: Set the update frequency; update once every 8 hours during the tunneling impact period, and once every 24 hours during the stable period; S52: Collect new data and recalculate; input a new batch of measured stress data at the ends of anchor bolts into the system, repeat S2 to S4, and regenerate the real-time support stress field at the current moment. ; S53: Calculate the rate of stress change; calculate the rate of stress change at each spatial point in adjacent time steps through time series analysis. ; S54: Mark potential risk areas; when the stress change rate of a certain area exceeds a preset threshold, the area is automatically marked as a potential risk area; S55: Establish an automatic data interface; establish an automatic data interface between monitoring data and model calculation to realize the automatic access of new data and the real-time update of digital twin model parameters.

8. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S6, the process of 3D visualization and augmented reality interaction is as follows: S61: Generate a 3D stress map; Generate a 3D stress cloud map on the computer, using each discrete micro-element of the surrounding rock space as the basic rendering object, and use color mapping to represent the stress magnitude: red represents the high-pressure stress area, yellow represents the medium-stress area, and blue represents the low-stress area or tensile stress area. S62: Draw isosurfaces and stress streamlines; draw isosurfaces to show the spatial shape of a specific stress value, and draw stress streamlines to show the direction of principal stress; S63: Build an augmented reality positioning system; use fixed feature points in the tunnel and infrared reflective markers to perform visual positioning; equip mobile terminals with supplementary lights and infrared cameras to ensure stable recognition even in low-light environments underground. S64: Real-time AR overlay display; The reconstructed support stress field is overlaid in real time on the real tunnel image captured by the mobile terminal camera in the form of a semi-transparent heat map, and users can switch between displaying different stress components or displaying stress profiles at different depths through gesture interaction. S65: Supports timeline playback; supports playback of historical stress field evolution along the timeline, analyzes the full-cycle response characteristics of the support system from installation to the present, and helps to trace the root causes of surrounding rock deformation and support failure.

9. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S7, the process for identifying potential instability zones based on the yield criterion is as follows: S71: Define the yield proximity index; Define the yield proximity index ,in, , These are the maximum and minimum principal stresses, respectively. For cohesion, It is the internal friction angle; S72: Determine the yield state; when When the discrete micro-element of the surrounding rock is determined to have entered the yielding state; when The time is used to determine whether the unit is in an elastic or stable state; S73: Highlight the yield zone; In the 3D visualization model, the yield unit is highlighted as a red transparent area; S74: Obtain mechanical parameters; cohesion and internal friction angle The values ​​are preferentially determined based on field or indoor test data; when no measured data is available, values ​​are determined based on experience with rock mass classification indicators, and are continuously corrected through data assimilation and inversion in subsequent monitoring.

10. The method for digital twinning of roadway support stress field by segmented force inversion of anchor bolts according to claim 1, characterized in that, In S8, the process of graded early warning and dynamic optimization of support parameters is as follows: S81: Establish a three-level risk assessment standard; Level I risk occurs when the compressive stress zones of adjacent anchor bolts overlap to form a continuous compression zone, with a thickness of not less than 0.3 times the length of the anchor bolt, and a stress variation coefficient of... There is no yield zone or the yield zone is scattered and not continuous; Level II risk is defined as local discontinuity in the compressive stress zone or a thickness less than 0.2 times the length of the anchor bolt. The yield zone extends locally but does not penetrate the anchoring layer; Level III risk is characterized by significant fracture of the compressive stress zone, or The yield zone may penetrate the anchorage layer, or a significant tensile stress zone may appear. S82: Automatic alarm trigger; For risk areas of level II and above, the system automatically triggers an alarm and displays it in a highlighted flashing display on the visual interface. The alarm information includes the risk level, location coordinates, and key judgment indicator values. S83: Generate optimization suggestions; Automatically generate support parameter optimization suggestions: Install additional anchor bolts and increase the support spacing in stress-deficient areas; Adjust anchor bolt preload; Increase the frequency of monitoring in high-risk areas and simultaneously develop a support reinforcement implementation plan; for permanent roadways, it is recommended to redesign the support parameters. S84: Dynamically adjust the early warning threshold; The risk assessment thresholds are dynamically adjusted based on the importance and service life of the roadway: for permanent roadways or important chambers, the risk assessment criteria are raised; for temporary roadways, the risk assessment criteria are relaxed.

Images