A digital driving tunnel intelligent construction parameter real-time regulation method

Abstract

A digital-driven real-time regulation method of tunnel intelligent construction parameters, which belongs to the technical field of mine tunnel construction control, builds a tunnel construction digital twin initial model containing surrounding rock, tunneling and supporting structure, and defines adjustable construction parameters; collects and preprocesses multi-source data of tunneling, supporting equipment and surrounding rock monitoring to form stable data flow to drive virtual model evolution; uses ensemble Kalman filter algorithm for data assimilation and relies on measured data to dynamically correct model variables; uses the updated model to simulate the response of surrounding rock in advance, constructs the objective function, takes tunneling speed, anchor pre-tightening force, etc. as variables, and optimizes construction parameters with the help of genetic algorithm; converts the optimized parameters into control commands and issues them to the equipment PLC system to complete the equipment linkage regulation; each time a tunneling cycle is completed, the data collection, model updating, parameter optimization and equipment regulation process are repeated. This method can significantly improve the safety, efficiency and intelligent level of construction under complex geological conditions.

Description

Technical Field

[0001] This invention belongs to the field of mine roadway construction control technology, specifically relating to a digitally driven method for real-time control of intelligent roadway construction parameters. Background Technology

[0002] Tunnel excavation and support are core processes in underground mining engineering, and the rational selection of construction parameters directly determines the effectiveness of surrounding rock stability control and engineering excavation efficiency. Currently, tunnel construction parameters are mostly based on preliminary geological survey results and engineering design experience, with pre-determined parameters such as excavation speed, anchor bolt preload, and support lag distance. However, the geological environment of underground engineering is complex and variable, and the actual surrounding rock conditions revealed on-site often deviate significantly from design assumptions. Adverse geological bodies such as fault fracture zones and lithological abrupt change zones exhibit strong uncertainty. In traditional construction models, on-site monitoring data is mostly used for post-construction analysis and manual early warning. Adjustments to construction parameters heavily rely on the experience and judgment of technical personnel, resulting in delayed control responses and difficulty in adapting to sudden changes in geological conditions. This passive control model of designing before construction and monitoring before rectification has become a key technical bottleneck restricting the safe and efficient excavation of deep tunnels.

[0003] While some progress has been made in the research on intelligent control of tunnel construction, there are still obvious shortcomings. For example, there are few technologies that deeply integrate digital models with real-time monitoring data and coordinate the control of construction equipment. Existing technologies mostly rely on numerical simulation to analyze the stability of surrounding rock or use on-site monitoring data to carry out safety early warning. The monitoring data and numerical models are isolated from each other, and a closed-loop control mechanism with real-time interaction and dynamic updates has not been constructed.

[0004] While some studies have introduced digital twin technology into the mining field, they have only reached the level of visualization and offline simulation, lacking technical pathways for real-time monitoring, data model assimilation, and automatic optimization of construction parameters. Furthermore, existing technologies cannot achieve direct linkage between optimized parameters and construction equipment; parameter optimization results still require manual intervention for adjustment. This prevents construction parameters from dynamically and adaptively adjusting to surrounding rock conditions, easily leading to problems such as conservative parameters hindering tunneling efficiency or untimely responses to geological changes causing safety hazards. To address these technical issues, there is an urgent need for a digitally driven method for real-time control of intelligent construction parameters in roadways. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a digitally driven method for real-time control of intelligent tunnel construction parameters. This method upgrades traditional experience-based tunnel construction management to data-driven intelligent adaptive control, which can significantly improve the safety, efficiency, and intelligence level of construction under complex geological conditions.

[0006] To achieve the above objectives, the present invention provides a digitally driven method for real-time control of intelligent construction parameters in tunnels, comprising the following steps:

[0007] Step 1: Construct an initial digital twin model for tunnel construction;

[0008] Based on geological survey data and tunnel design, a three-dimensional numerical model including surrounding rock, tunneling equipment and support structure is established using FLAC3D or ABAQUS as a virtual mirror of the digital twin, and adjustable construction parameters are defined.

[0009] Step 2: Multi-source real-time data acquisition and preprocessing;

[0010] Data is collected in real time from tunneling machines, rock bolt drilling rigs, and surrounding rock monitoring sensors. Spatiotemporal synchronization, anomaly removal, and filtering smoothing are performed to form a continuous and stable data stream that drives the evolution of the digital twin.

[0011] Step 3: Dynamically update the digital twin model based on data assimilation;

[0012] Four key mechanical parameters with high uncertainty were selected as variables to be corrected. An ensemble Kalman filter algorithm was used to dynamically update the parameters using measured displacement data, so that the digital twin model could approximate the real surrounding rock condition.

[0013] Step 4: Optimization of construction parameters based on digital twin-based advanced prediction;

[0014] Based on the updated digital twin model, the surrounding rock response of the next construction cycle is simulated. A comprehensive control objective function is defined, with tunneling speed, anchor bolt preload, and support lag distance as design variables. A genetic algorithm is used to solve for the optimal combination of construction parameters, and the support effect is required to meet the minimum standard.

[0015] Step 5: Issue real-time control commands and coordinate with equipment;

[0016] The optimized construction parameters are converted into equipment control commands and sent in real time to the PLC control systems of the tunneling machine and the anchor drilling rig via industrial Ethernet, realizing the reverse control of the physical equipment by the digital twin;

[0017] Step 6: Iterate and update the digital twin;

[0018] After each tunneling cycle is completed, steps 2 to 5 are immediately repeated to achieve synchronous evolution of the digital twin model and the physical tunnel, and the construction parameters are dynamically and adaptively adjusted according to changes in the surrounding rock conditions.

[0019] As an alternative, step 7 is also included: a safety early warning and manual intervention mechanism;

[0020] When the predicted surrounding rock response exceeds the attention threshold or safety threshold, the system issues a graded warning and, if necessary, automatically suspends tunneling and prompts for manual intervention.

[0021] As a preferred option, the process of constructing the initial digital twin model of the tunnel construction in step 1 is as follows:

[0022] S11: Model Scope and Boundary Conditions; The model size is set to 5 times the width of the tunnel to ensure that boundary effects do not affect the calculation results; Boundary conditions are set as follows: the bottom is fixed with vertical displacement, the top is applied with vertical principal stress, and the sides are applied with horizontal principal stress.

[0023] S12: Constitutive model and parameter assignment of surrounding rock; The surrounding rock adopts the Mohr-Coulomb elastoplastic constitutive model, and different rock layers are assigned corresponding mechanical parameters according to the stratigraphic distribution, and an initial geostress field is set;

[0024] S13: Tunneling and Support Simulation; the tunneling equipment is simulated with a dynamic moving boundary, advancing gradually with each tunneling cycle; the anchor bolts are simulated using cable elements, and the steel arch frame is simulated using beam elements, arranged according to the design drawings; the anchor bolt element parameters are set according to the actual product specifications and calibrated through indoor pull-out tests to truly reflect the interaction characteristics between the anchor bolts and the surrounding rock;

[0025] S14: Mesh generation; densify the mesh around the roadway to ensure calculation accuracy in areas with large stress gradients in the surrounding rock, and gradually thin it outwards to improve calculation efficiency;

[0026] S15: Definition of adjustable parameters; Adjustable construction parameters are defined in the model: tunneling speed. Anchor bolt preload Support lag distance This is used for subsequent optimization and regulation.

[0027] As a preferred option, the multi-source real-time data acquisition and preprocessing process in step 2 is as follows:

[0028] S21: Tunneling machine parameter acquisition; Sensors are installed on the tunneling machine to collect tunneling parameters once per second, including thrust, torque, and propulsion speed;

[0029] S22: Anchor bolt drilling rig parameter acquisition; acquiring the preload applied value and drilling depth on the anchor bolt drilling rig;

[0030] S23: Surrounding rock monitoring data acquisition; deploying multi-point displacement gauges and anchor bolt force gauges in the surrounding rock of the roadway; the multi-point displacement gauges are deployed to cover the interior of the loosened zone, the middle of the anchoring zone, the boundary of the anchoring zone, and the original rock zone, with multiple measuring holes deployed at each monitoring section, and deep displacement is collected every 10 minutes; anchor bolt force gauges are installed at the ends of representative anchor bolts, evenly distributed on the roof and both sides, and the axial force of the anchor bolts is collected every 10 minutes;

[0031] S24: Data preprocessing; all collected data are spatiotemporally synchronized; outliers that are significantly outside the range are directly removed and replaced with the value of the previous valid moment; the moving average method is used for filtering and smoothing to form a continuous and stable real-time data stream.

[0032] As a preferred option, in step 3, the process of dynamically updating the digital twin model based on data assimilation is as follows:

[0033] S31: Generate the parameters to be corrected and the initial set; define the parameter vector. ,in, For elastic modulus, For cohesion, For internal friction angle, The lateral pressure coefficient; an initial set of 50 samples is generated for the parameter to be corrected. The sample value range is set based on the results of indoor tests: Take 0.5 to 1.5 times the experimental value. Take 0.3 to 2.0 times the experimental value. Take the test value ±5°. Take a value of 0.5 to 1.5;

[0034] S32: Calculate the simulated observation values; run the numerical model for each sample to obtain the simulated displacement values ​​of the monitoring point locations. Define observation vector ,in The number of monitoring points For the first Measured displacement at each monitoring point;

[0035] S33: Calculate the Kalman gain; calculate the Kalman gain matrix according to the following formula. :

[0036] ;

[0037] In the formula, This is the covariance matrix between the parameters and the simulated observations; The covariance matrix of the simulated observations; The observation error covariance matrix;

[0038] S34: Parameter update; update each parameter sample according to the following formula:

[0039] ;

[0040] In the formula, For the first The parameter values ​​of each sample before the update; For the first The simulated displacement vector corresponding to each sample; This represents the residual between the measured displacement and the simulated displacement.

[0041] S35: Iterative convergence and post-processing; iterate and update 3-5 times until the parameters converge; take the mean of the updated sample set as the true parameters of the current surrounding rock to obtain the updated digital twin model; the calculation time of each update must be completed before the next construction decision node; after the parameter update, a physical rationality test must be performed to ensure that it is within the preset range; if it exceeds the range, the boundary value is taken; the assimilation frequency is matched with the data acquisition frequency to ensure that the digital twin model always reflects the latest surrounding rock condition.

[0042] As a preferred option, the optimization process for construction parameters based on digital twin-based advance prediction in step 4 is as follows:

[0043] S41: Advanced prediction; Based on the updated digital twin model from step 3, simulate the surrounding rock response of the next tunneling cycle and predict the maximum displacement of the roof and the volume of the plastic zone;

[0044] S42: Define the integrated control objective function; define the objective function according to the following formula. :

[0045] ;

[0046] In the formula, This represents the maximum displacement of the roof predicted by the model. The allowable displacement threshold; The volume of the plastic region; This represents the total volume of the anchorage area; This is the current construction cycle time; For reference cycle time; , , Let be the weighting coefficient, satisfying + The settings are determined based on the different levels of importance that projects place on safety and efficiency.

[0047] S43: Calculate the volume of the plastic zone; calculate the volume of the plastic zone according to the following formula. :

[0048] ;

[0049] In the formula, This represents the total number of units near the anchorage zone. Let i be the volume of the i-th unit. This is the i-th indicator function;

[0050] S44: Design variables and constraints; based on tunneling speed Anchor bolt preload Support lag distance As design variables; to set engineering feasibility constraints: , , 1000; of which The safety factor for the anchor bolt preload. This refers to the maximum permissible advance speed of the tunneling machine. The minimum safe hysteresis distance; The anchor rod's yield strength; The cross-sectional area of ​​the anchor bolt; This is the safety factor for the preload force;

[0051] S45: Define the evaluation index for support effectiveness; the evaluation index for support effectiveness is defined according to the following formula:

[0052] ;

[0053] In the formula, This represents the total number of units within the anchorage zone. For the first Additional stress on the support of each unit; The vertical principal stress of the original rock;

[0054] The evaluation index corresponding to the optimized parameter combination is required. If the conditions are not met, adjust the parameters and re-optimize. The critical evaluation threshold for tunnel support stability is determined by on-site geological conditions and design specifications.

[0055] S46: Optimization Algorithm; Using a genetic algorithm to solve the objective function. The algorithm seeks the minimum optimal combination of construction parameters. During each generation of evolution, constraints are first assessed for the generated parameter combinations. The algorithm terminates when the objective function value changes less than a specified value for multiple consecutive generations or when the maximum number of iterations is reached. The final output is the optimal parameter combination. The displacement and support effects are verified through a complete numerical model forward modeling process to ensure that they meet the requirements. If the verification fails, the algorithm parameters are adjusted and re-optimized to ensure that the output results are true and reliable.

[0056] As a preferred option, in step 5, the process of issuing real-time control commands and linking them with the equipment is as follows:

[0057] S51: Command generation and issuance; The construction parameters obtained from step 4 are optimized. The commands are converted into equipment control instructions and sent in real time to the tunneling machine PLC control system and the bolt drilling rig control system via industrial Ethernet.

[0058] S52: Equipment execution; the tunneling machine automatically adjusts its advance speed according to the instruction. The anchor drilling rig automatically sets the preload to... On-site workers follow the optimized lag distance Organize support construction;

[0059] S53: Feedback and Redundancy; Real-time feedback of equipment execution status to the digital twin system; Redundant design of industrial Ethernet communication to ensure that single point of failure does not affect command issuance; The equipment control system completes parameter adjustment within the specified control cycle and feeds back the execution results; If no feedback is received from the equipment for several consecutive times, the system determines that there is a communication failure and issues an alarm, prompting on-site personnel to check the network connection and equipment status; The total delay from command issuance to equipment response is controlled within the specified time.

[0060] As a preferred embodiment, the iterative process and digital twin update in step 6 are as follows:

[0061] S61: Cyclic execution; after each tunneling cycle is completed, the latest monitoring data is immediately collected again, the digital twin model parameters are updated, the construction parameters for the next cycle are optimized, and instructions are issued;

[0062] S62: Adaptive adjustment; through repeated closed-loop iterations, an intelligent construction process control system with digital twin as its core is formed, realizing dynamic adaptive control of construction parameters.

[0063] As a preferred option, the process of the safety warning and manual intervention mechanism in step 7 is as follows:

[0064] S71: Threshold-based early warning; when the surrounding rock response predicted in step 4 exceeds the set attention threshold, the system displays an early warning on the visualization interface to prompt on-site technicians to strengthen monitoring.

[0065] S72: Safety threshold warning and intervention; when the predicted response exceeds the safety threshold, or the optimization algorithm cannot find a parameter combination that meets the constraints, the system displays a high-level warning, automatically suspends tunneling, and prompts for manual intervention.

[0066] S73: Manual decision-making; After manual intervention, on-site technicians judge based on the revealed geological conditions, and choose to continue to implement the original plan, adjust and optimize the parameter range and recalculate, or take enhanced support measures. Automatic control is restored only after safety is confirmed.

[0067] This invention aims to overcome the shortcomings of existing technologies and provide a digitally driven method for real-time control of intelligent construction parameters in tunnels. By constructing a digital twin model that evolves synchronously with the physical tunnel, real-time data from tunneling equipment, support equipment, and surrounding rock monitoring are integrated. A data assimilation algorithm dynamically updates the model parameters, and based on the updated model, the surrounding rock response is predicted in advance, construction parameters are automatically optimized, and finally, the optimized commands are sent to the equipment control system for execution. This invention forms an adaptive closed-loop construction system of "monitoring-assimilation-prediction-control," significantly improving the intelligence level of tunnel construction under complex geological conditions. It effectively solves the problems of construction safety and efficiency caused by the inability of traditional methods to respond in real-time to changes in surrounding rock, achieving a leap from passive response to active control.

[0068] Compared with the prior art, the technical advantages of the present invention are as follows:

[0069] 1. Real-time synchronous evolution of the digital twin model and the physical tunnel: An initial numerical model is established based on geological survey data, and through real-time fusion of data from tunneling equipment, support equipment, and surrounding rock monitoring, the digital twin model can be continuously updated to reflect the current real state of the surrounding rock. The model is built using FLAC3D / ABAQUS, with a finer mesh around the tunnel and reasonable boundary conditions, providing a high-fidelity virtual mirror for construction parameter control.

[0070] 2. A data assimilation algorithm enables dynamic adaptive correction of model parameters. An ensemble Kalman filter algorithm is employed, driven by measured displacements from multi-point displacement gauges, to dynamically update four key mechanical parameters: elastic modulus, cohesion, internal friction angle, and lateral pressure coefficient. An initial set of 50 parameter samples is generated, and the Kalman gain matrix is ​​calculated. The residual between measured and simulated displacements is converted into parameter correction values, iterated 3-5 times until convergence. This method allows the digital twin model to gradually approximate the actual surrounding rock conditions from its initial design state, effectively overcoming the shortcomings of traditional methods where model parameters remain unchanged for a long time and are disconnected from actual measurements.

[0071] 3. Advanced prediction and multi-objective construction parameter optimization based on an updated model: Based on the updated digital twin model, the surrounding rock response (maximum roof displacement, plastic zone volume, etc.) for the next tunneling cycle (1.0 m) is predicted in advance. A comprehensive control objective function is constructed, with tunneling speed, anchor bolt preload, and support lag distance as design variables. A genetic algorithm is used to automatically solve for the optimal parameter combination, and support effect evaluation indicators are introduced as constraints. This mechanism realizes the transformation from passive response to active control, significantly improving the construction adaptability and safety under complex geological conditions.

[0072] 4. Real-time control commands are automatically issued and executed in conjunction with equipment. The optimized construction parameters are sent in real-time via industrial Ethernet to the PLC control systems of the tunneling machine and anchor drilling rig. The equipment automatically adjusts its advance speed and preload, and operators organize construction according to the optimized lag distance. Based on a communication redundancy design, the equipment execution status can be fed back in real time, and a communication fault alarm is provided, forming a real-time intelligent construction system of monitoring, assimilation, prediction, and control, reducing reliance on manual intervention.

[0073] 5. Iterative Cycles and Continuous Evolution of the Digital Twin: After each tunneling cycle is completed, the processes of data acquisition, assimilation and update, advanced prediction, parameter optimization, and command issuance are immediately repeated. Through continuous closed-loop iteration, the digital twin model evolves synchronously with the physical tunnel, and the construction parameters are dynamically and adaptively adjusted according to the changes in the exposed surrounding rock conditions, truly realizing intelligent construction process control that uses the real to drive the virtual and the virtual to control the real.

[0074] 6. Tiered safety early warning and manual intervention mechanism: When the predicted surrounding rock response exceeds the concern threshold, the system displays an early warning on the visualization interface, prompting enhanced monitoring; when the safety threshold is exceeded or the optimization algorithm cannot find a feasible solution, the system automatically suspends tunneling and issues a high-level alarm, requiring manual intervention. On-site technicians can choose to continue the original plan, adjust the optimization scope, or take enhanced support measures based on the revealed geological conditions to ensure that construction safety is always under control.

[0075] 7. An optional historical case database accelerates assimilation;

[0076] A historical construction case database can be established, recording the surrounding rock condition parameters, updated mechanical parameters, optimized construction parameters, and actual monitoring results for each cycle. When the geological conditions of a new working face are similar to those of a historical case, the system automatically calls the final parameter sample from that case as the initial population, accelerating the convergence process of the ensemble Kalman filter and improving computational efficiency. However, core updates and optimizations are still strictly driven by current real-time monitoring data to ensure relevance.

[0077] This method upgrades traditional experience-based tunnel construction management to data-driven intelligent adaptive control through a complete technology chain including digital twin modeling, data assimilation, advanced prediction, automatic optimization, closed-loop execution, iterative iteration, and hierarchical early warning. This significantly improves the safety, efficiency, and intelligence of construction under complex geological conditions. Attached Figure Description

[0078] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0079] 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.

[0080] like Figure 1 As shown, this invention provides a digitally driven method for real-time control of intelligent construction parameters in tunnels, comprising the following steps:

[0081] Step 1: Construct an initial digital twin model for tunnel construction;

[0082] Based on geological survey data and tunnel design, a three-dimensional numerical model including surrounding rock, tunneling equipment and support structure is established using FLAC3D or ABAQUS as a virtual mirror of the digital twin, and adjustable construction parameters are defined.

[0083] As a preferred option, the process of constructing the initial digital twin model for tunnel construction is as follows:

[0084] S11: Model Scope and Boundary Conditions; The model size is set to 5 times the width of the tunnel to ensure that boundary effects do not affect the calculation results; Boundary conditions are set as follows: the bottom is fixed with vertical displacement, the top is applied with vertical principal stress, and the sides are applied with horizontal principal stress.

[0085] S12: Constitutive model and parameter assignment of surrounding rock; The surrounding rock adopts the Mohr-Coulomb elastoplastic constitutive model, and different rock layers are assigned corresponding mechanical parameters (elastic modulus, Poisson's ratio, cohesion, internal friction angle, density, etc.) according to the stratigraphic distribution, and an initial geostress field is set.

[0086] S13: Tunneling and Support Simulation; the tunneling equipment is simulated with a dynamic moving boundary, advancing gradually with each tunneling cycle; the anchor bolts are simulated using cable elements, and the steel arch frame is simulated using beam elements, arranged according to the design drawings; the anchor bolt element parameters are set according to the actual product specifications and calibrated through indoor pull-out tests to truly reflect the interaction characteristics between the anchor bolts and the surrounding rock;

[0087] S14: Mesh generation; densify the mesh around the roadway to ensure calculation accuracy in areas with large stress gradients in the surrounding rock, and gradually thin it outwards to improve calculation efficiency;

[0088] S15: Definition of adjustable parameters; Adjustable construction parameters are defined in the model: tunneling speed. (Unit: m / min) Anchor bolt preload (Unit: kN) Support lag distance (Unit: m), used for subsequent optimization and control.

[0089] In this technical solution, by setting the model range to 5 times the tunnel width, fixing the vertical displacement at the bottom, and applying ground stress at the top and sides as boundary conditions, a Mohr-Coulomb elastoplastic constitutive model is adopted and different rock strata mechanical parameters are assigned. Dynamic moving boundaries are used to simulate tunneling, cable elements are used to simulate anchor bolts, and beam elements are used to simulate steel arches. The mesh is densified around the tunnel. At the same time, the tunneling speed, anchor bolt preload, and support lag distance are defined as adjustable parameters, thereby establishing an initial digital twin model that can truly reflect the interaction between the surrounding rock and the support and has optimization capabilities.

[0090] Step 2: Multi-source real-time data acquisition and preprocessing;

[0091] Data is collected in real time from tunneling machines, rock bolt drilling rigs, and surrounding rock monitoring sensors. Spatiotemporal synchronization, anomaly removal, and filtering smoothing are performed to form a continuous and stable data stream that drives the evolution of the digital twin.

[0092] As a preferred option, the multi-source real-time data acquisition and preprocessing process is as follows:

[0093] S21: Tunneling machine parameter acquisition; Sensors are installed on the tunneling machine to collect tunneling parameters once per second. The tunneling parameters include thrust (unit: kN), torque (unit: kN·m), and propulsion speed (unit: m / min).

[0094] S22: Anchor bolt drilling rig parameter acquisition; collect the preload applied value (in kN) and drilling depth (in m) on the anchor bolt drilling rig.

[0095] S23: Surrounding rock monitoring data acquisition; deploying multi-point displacement gauges and anchor bolt force gauges in the surrounding rock of the roadway; the multi-point displacement gauges are deployed to cover the interior of the loosened zone, the middle of the anchoring zone, the boundary of the anchoring zone, and the original rock zone, with multiple measuring holes (roof and sidewalls) deployed at each monitoring section, and collecting deep displacement (in mm) every 10 minutes; anchor bolt force gauges are installed at the ends of representative anchor bolts, evenly distributed in the roof and sidewalls, and collecting anchor bolt axial force (in kN) every 10 minutes.

[0096] S24: Data preprocessing; Spatiotemporal synchronization of all collected data (time synchronization error does not exceed the specified value, time is synchronized via GPS or industrial network); for outliers that are significantly outside the range, they are directly removed and replaced with the value of the previous valid moment; filtering and smoothing are performed using the moving average method to form a continuous and stable real-time data stream, which serves as the input to drive the evolution of the digital twin;

[0097] In this technical solution, the tunneling machine's sensors collect thrust, torque, and propulsion speed every second, while the anchor drilling rig collects preload and drilling depth. Multiple displacement gauges (covering the loosened zone to the original rock area, collecting data every 10 minutes) and anchor bolt force gauges (collecting axial force every 10 minutes) are deployed in the surrounding rock of the tunnel. All data are then subjected to spatiotemporal synchronization, outlier removal (replacing with adjacent valid values), and moving average filtering to form a continuous and stable real-time data stream, providing highly reliable driving input for the digital twin model.

[0098] Step 3: Dynamically update the digital twin model based on data assimilation;

[0099] Four key mechanical parameters with relatively large uncertainties (elastic modulus, cohesion, internal friction angle, and lateral pressure coefficient) were selected as variables to be corrected. An ensemble Kalman filter algorithm was used to dynamically update the parameters using measured displacement data, so that the digital twin model could approximate the real surrounding rock condition.

[0100] As a preferred approach, the dynamic update process of a digital twin model based on data assimilation is as follows:

[0101] S31: Generate the parameters to be corrected and the initial set; define the parameter vector. ,in, Elastic modulus (unit: GPa) Cohesion (unit: MPa) The internal friction angle (in degrees). The lateral pressure coefficient (dimensionless); an initial set of 50 samples is generated for the parameter to be corrected. The sample value range is set based on the results of indoor tests: Take 0.5 to 1.5 times the experimental value. Take 0.3 to 2.0 times the experimental value. Take the test value ±5°. The value should be between 0.5 and 1.5; the sample set size should be an optimal value that balances computational accuracy and speed.

[0102] S32: Calculate the simulated observation values; run the numerical model for each sample to obtain the simulated displacement values ​​of the monitoring point locations. (Unit: mm); Define observation vector ,in The number of monitoring points For the first Measured displacement at each monitoring point;

[0103] S33: Calculate the Kalman gain; calculate the Kalman gain matrix according to the following formula. :

[0104] ;

[0105] In the formula, The covariance matrix of the parameters and simulated observations is used to quantitatively characterize the degree of influence of parameter changes on the displacement of monitoring points. The covariance matrix of the simulated observations reflects the correlation between the displacements of different monitoring points; The observation error covariance matrix is ​​determined based on the sensor accuracy.

[0106] S34: Parameter update; update each parameter sample according to the following formula:

[0107] ;

[0108] In the formula, For the first The parameter values ​​of each sample before the update; For the first The simulated displacement vector corresponding to each sample; The residual between the measured displacement and the simulated displacement is the core basis for driving the update of digital twin parameters;

[0109] S35: Iterative convergence and post-processing; iterate and update 3-5 times until the parameters converge (the change in parameters after two consecutive updates is less than the specified value, indicating that the model has stabilized); take the mean of the updated sample set as the true parameters of the current surrounding rock to obtain the updated digital twin model; the calculation time for each update must be completed before the next construction decision node; after the parameter update, a physical rationality test must be performed to ensure that it is within the preset range; if it exceeds the range, the boundary value is taken to prevent parameter divergence; the assimilation frequency is matched with the data acquisition frequency to ensure that the digital twin model always reflects the latest surrounding rock condition;

[0110] S36: Acceleration Strategy; To achieve fast computation, a surrogate model (such as a neural network or Gaussian process) can be used to replace the complete numerical model for a large number of forward calculations. Cross-validation is used to ensure that the prediction accuracy of the surrogate model meets the requirements and that the surrogate model is accurate enough to replace the original numerical model, thereby accelerating the update speed of the digital twin.

[0111] S37: Assimilation method; a sequential assimilation method is adopted, and each update only uses the data collected in the latest time period, discarding historical outdated data to avoid model response lag;

[0112] S38: Accelerated Assimilation Using Historical Case Database (Preferred Implementation); To accelerate the convergence speed of ensemble Kalman filtering, a historical construction case database can be established; this database records the surrounding rock condition parameters (such as rock strength, integrity coefficient, etc.) and updated mechanical parameters for each tunneling cycle. ), and optimized construction parameters ( , , The system monitors the geological conditions encountered at the new working face and the key indicators of a historical case in the database. When the deviation between these geological conditions and the database's key indicators is small (e.g., the relative deviation of each indicator is less than 10%), the system automatically calls upon the parameter samples updated from that historical case as the initial population for the current loop's Kalman filter, thereby reducing the parameter search range and accelerating the assimilation and convergence process. It is important to emphasize that historical cases are only used to provide reasonable initial guesses; model updates and parameter optimization are still strictly driven by real-time monitoring data from the current working face, ensuring that the optimization results of each loop are targeted and adaptable, truly reflecting the essential characteristics of a digital twin: using the real to drive the virtual and using the virtual to control the real.

[0113] The core idea of ​​this algorithm is to use the statistical relationship of multiple parameter samples to transform the difference between measured data and model simulation values ​​into parameter correction amounts, thereby enabling the digital twin model to gradually approach the real surrounding rock state from the initial design state.

[0114] In this technical solution, by defining elastic modulus, cohesion, internal friction angle, and lateral pressure coefficient as parameters to be corrected, an initial set of 50 samples is generated, and a numerical model is run to obtain the simulated displacement of each monitoring point. Based on the residual between the measured displacement and the simulated displacement, the Kalman gain matrix is ​​calculated, and the parameter samples are iteratively updated 3 to 5 times until convergence. The mean of the samples is taken as the current true parameters of the surrounding rock, and a digital twin model that evolves synchronously with the physical tunnel is obtained. At the same time, a surrogate model is used to accelerate the calculation, and sequential assimilation discards outdated data. It is also possible to use a historical construction case database to provide a similar initial population to accelerate assimilation convergence. However, the core update is always driven by the current real-time monitoring data to ensure that the model truly reflects the state of the surrounding rock.

[0115] Step 4: Optimization of construction parameters based on digital twin-based advanced prediction;

[0116] Based on the updated digital twin model, the surrounding rock response of the next construction cycle (1.0m tunneling distance) is simulated. A comprehensive control objective function is defined, with tunneling speed, anchor bolt preload and support lag distance as design variables. The optimal combination of construction parameters is solved by a genetic algorithm, and the support effect is required to meet the minimum standard.

[0117] As a preferred option, the optimization process for construction parameters based on digital twin-based advanced prediction is as follows:

[0118] S41: Advanced prediction; Based on the updated digital twin model in step 3, simulate the surrounding rock response of the next tunneling cycle (one tunneling distance, taken as 1.0 m), and predict indicators such as the maximum displacement of the roof and the volume of the plastic zone.

[0119] S42: Define the integrated control objective function; define the objective function according to the following formula. :

[0120] ;

[0121] In the formula, The maximum displacement of the roof predicted by the model (in mm) directly reflects the stability of the surrounding rock. The allowable displacement threshold (in mm) is determined based on engineering specifications or experience from similar projects. The volume of the plastic zone (in m³) reflects the extent of surrounding rock failure. This refers to the total volume of the anchorage zone (in m³), ​​which is the volume of the effective control area of ​​the anchor bolt. This represents the current construction cycle time (in minutes), reflecting construction efficiency. The reference cycle time (in minutes) serves as a benchmark for efficiency evaluation. , , Let be the weighting coefficient, satisfying + The settings are determined based on the different levels of importance that projects place on safety and efficiency.

[0122] objective function Used to quantitatively evaluate the merits of a construction plan, the smaller the value, the better the plan balances safety and efficiency;

[0123] S43: Calculate the volume of the plastic zone; calculate the volume of the plastic zone according to the following formula. :

[0124] ;

[0125] In the formula, This represents the total number of units near the anchorage zone. Let i be the volume of the i-th unit. This is the i-th indicator function, which takes the value 1 when the element is in the yield state, and 0 otherwise;

[0126] S44: Design variables and constraints; based on tunneling speed (Unit: m / min) Anchor bolt preload (Unit: kN) Support lag distance (Unit: m) as design variables; set engineering feasibility constraints: , , 1000; of which The safety factor for the anchor bolt preload is less than 1, ensuring that the preload does not exceed the yield bearing capacity as required by the standard. This refers to the maximum permissible advance speed of the tunneling machine. To minimize the safety lag distance and prevent excessive open roof area from causing roof collapse risk; The anchor bolt's yield strength (in MPa). The cross-sectional area of ​​the anchor bolt (in mm) 2 ); To ensure the anchor bolts do not break due to the preload safety factor;

[0127] S45: Define the evaluation indicators for support effectiveness; evaluation indicators for support effectiveness This value is used to quantify the quality of the compressive stress zone formed by the anchor bolt group in the surrounding rock. The larger the value, the more effective the anchor bolts are in reinforcing the surrounding rock. The average additional support stress can be defined first. Then the evaluation index of support effect Specifically, the evaluation index for support effectiveness can be defined according to the following formula:

[0128] ;

[0129] In the formula, The total number of units within the anchorage zone is obtained through grid statistics from the digital twin model; For the first The additional stress of the support unit (in MPa) is calculated by subtracting the stress field without anchor bolts from the stress field after anchor bolts are applied. This difference directly reflects the reinforcement contribution of the anchor bolts to the surrounding rock. The original rock vertical principal stress (in MPa) is used as the normalization benchmark.

[0130] The evaluation index corresponding to the optimized parameter combination is required. This means that the additional stress on the support needs to reach a certain level to ensure the effectiveness of the support; if this is not met, the parameters need to be adjusted and re-optimized. The critical evaluation threshold for tunnel support stability is determined by on-site geological conditions and design specifications.

[0131] Support effect evaluation indicators The calculation requires extracting the additional support stress of all units within the anchorage zone during the post-processing of the digital twin model. The anchorage zone is defined as a ring-shaped area extending from the roadway surface at a certain multiple of the anchor bolt length (e.g., 0.5 to 1.5 times the anchor bolt length). The area and volume of this zone are calculated by integrating the dimensions according to the roadway cross-section. Total number of elements. This is obtained by statistically analyzing all grid cells within the region, ensuring the completeness of the statistics.

[0132] S46: Optimization Algorithm; Genetic Algorithm is a global optimization method that simulates the biological evolution process. It searches for the optimal solution in the parameter space step-by-step through selection, crossover, and mutation operations. Using a genetic algorithm to solve the objective function... The algorithm seeks the minimum optimal combination of construction parameters. During each generation of evolution, the generated parameter combinations are first evaluated for constraints to ensure the optimization results are practically feasible in engineering. The algorithm terminates when the objective function value changes less than a specified value for multiple consecutive generations or when the maximum number of iterations (50 generations) is reached. The final output is the optimal parameter combination. The displacement and support effects are verified through a complete numerical model forward modeling process to ensure that they meet the requirements. If the verification fails, the algorithm parameters are adjusted (such as increasing the population size or the number of iterations) and then re-optimized to ensure that the output results are true and reliable.

[0133] This technical solution utilizes an updated digital twin model to predict the surrounding rock response in advance, constructing a multi-objective function that integrates roof displacement, plastic zone volume, and construction cycle time. With tunneling speed, prestressing force, and lag distance as design variables, a genetic algorithm automatically solves for the optimal parameter combination. Furthermore, an evaluation index for support effectiveness is introduced to ensure the reinforcement effect meets standards. This method transforms traditional experience-based decision-making into data-driven multi-objective optimization, proactively adjusting construction efficiency while ensuring surrounding rock safety, significantly improving the synergy and intelligence of tunneling and support under complex geological conditions.

[0134] Step 5: Issue real-time control commands and coordinate with equipment;

[0135] The optimized construction parameters are converted into equipment control commands and sent in real time to the PLC control systems of the tunneling machine and the anchor drilling rig via industrial Ethernet, realizing the reverse control of the physical equipment by the digital twin;

[0136] As a preferred option, the process of issuing real-time control commands and linking them with equipment is as follows:

[0137] S51: Command generation and issuance; The construction parameters obtained from step 4 are optimized. The commands are converted into equipment control instructions and sent in real time to the tunneling machine PLC control system and the bolt drilling rig control system via industrial Ethernet.

[0138] S52: Equipment execution; the tunneling machine automatically adjusts its advance speed according to the instruction. The anchor drilling rig automatically sets the preload to... On-site workers follow the optimized lag distance Organize support construction;

[0139] S53: Feedback and Redundancy; Real-time feedback of equipment execution status to the digital twin system; Redundant design of industrial Ethernet communication to ensure that single-point failures do not affect command issuance; The equipment control system completes parameter adjustment within the specified control cycle and feeds back the execution results; If no feedback is received from the equipment for several consecutive times, the system determines a communication failure and issues an alarm, prompting on-site personnel to check the network connection and equipment status; The total delay from command issuance to equipment response is controlled within the specified time to ensure that control commands take effect in a timely manner, realizing reverse control of physical equipment by the digital twin;

[0140] In this technical solution, the optimized construction parameters are converted into equipment control commands and sent in real time to the PLC control systems of the tunneling machine and the anchor drilling rig via industrial Ethernet. This enables automatic adjustment of the advance speed and pre-tensioning force, as well as standardized execution of the support lag distance. At the same time, the use of communication redundancy design, real-time feedback of execution status, and fault alarm mechanism ensures the reliability of command transmission and the timeliness of response. This forms a closed loop of reverse control of the physical equipment by the digital twin, which significantly improves the automation level of construction control and the reliability of command execution.

[0141] Step 6: Iterate and update the digital twin;

[0142] After each tunneling cycle is completed, steps 2 to 5 are immediately repeated to achieve synchronous evolution of the digital twin model and the physical tunnel, and the construction parameters are dynamically and adaptively adjusted according to changes in the surrounding rock conditions.

[0143] As a preferred approach, the iterative process and digital twin update process are as follows:

[0144] S61: Execute in cycles; after each tunneling cycle (one tunneling distance) is completed, immediately collect the latest monitoring data again (step 2), update the digital twin model parameters (step 3), optimize the construction parameters for the next cycle (step 4), and issue instructions (step 5).

[0145] S62: Adaptive adjustment; Through repeated closed-loop iterations, the digital twin model and the physical tunnel are synchronously evolved, and the construction parameters are dynamically and adaptively adjusted according to the changes in the exposed surrounding rock conditions, forming an intelligent construction process control system with digital twin as the core, and realizing dynamic adaptive regulation of construction parameters.

[0146] In this technical solution, by repeatedly collecting data, updating the model, optimizing parameters, and issuing instructions after each tunneling cycle, a closed-loop iterative mechanism is formed, which enables the digital twin model to evolve synchronously with the physical tunnel. The construction parameters can be dynamically and adaptively adjusted according to changes in the surrounding rock conditions, effectively ensuring the continuous optimization capability and long-term adaptability of the construction process.

[0147] Step 7: Safety early warning and manual intervention mechanism;

[0148] When the predicted surrounding rock response exceeds the attention threshold or safety threshold, the system issues a graded warning and, if necessary, automatically suspends tunneling and prompts for manual intervention.

[0149] As a preferred approach, the process of a safety early warning and manual intervention mechanism is as follows:

[0150] S71: Threshold-based early warning; when the surrounding rock response predicted in step 4 (e.g.) or When the threshold is exceeded, the system will display a warning on the visual interface, prompting on-site technicians to strengthen monitoring.

[0151] S72: Safety threshold warning and intervention; when the predicted response exceeds the safety threshold, or the optimization algorithm cannot find a parameter combination that meets the constraints, the system displays a high-level warning, automatically suspends tunneling, and prompts for manual intervention.

[0152] S73: Manual decision-making; After manual intervention, on-site technicians can judge based on the revealed geological conditions and choose to continue the original plan, adjust and optimize the parameter range (such as relaxing constraints), recalculate, or take enhanced support measures (such as densifying anchor bolts, grouting, etc.). Automatic control will be restored after safety is confirmed.

[0153] In this technical solution, a two-level early warning mechanism with a focus threshold and a safety threshold is set. When the predicted response exceeds the standard, a graded alarm is triggered and a prompt to strengthen monitoring is given. When the response exceeds the safety threshold or there is no solution for optimization, the tunneling is automatically suspended and manual intervention is requested. On-site technicians can flexibly choose to continue execution, adjust the optimization range, or take reinforcement measures according to the revealed geological conditions, so as to ensure that construction safety is under control under high-risk conditions, and to balance the efficiency of automation with the reliability of human decision-making.

[0154] Implementation and verification:

[0155] The application scenario is a tunneling project in a mine's main transport roadway. The roadway is designed with a straight-walled, semi-circular arch cross-section, with a net width of 5.0m and a net height of 4.0m. It employs a combined anchor-mesh-cable support system, with anchor bolts measuring Φ22×2400mm and spaced 800×800mm apart. The tunneling equipment is an EBZ-200 cantilever roadheader. The roadway traverses both sandstone and mudstone strata, resulting in significant variations in surrounding rock conditions.

[0156] To verify the effectiveness of the method of this invention, a 40-day continuous tracking test was conducted during the tunnel excavation process, covering two complete construction sections. The four schemes in Table 1 were also deployed for comparison.

[0157] Table 1: Comparison of Schemes

[0158]

[0159] I. Sandstone Section Construction Stage (Days 1-20, Excavation Mileage 0-60m);

[0160] The surrounding rock in this section is moderately weathered sandstone. The elastic modulus was measured in laboratory tests. =18.5GPa, cohesion =8.2MPa, internal friction angle =42°.

[0161] This invention establishes an initial digital twin model according to step 1. The model size is five times the width of the tunnel. The surrounding rock adopts a Mohr-Coulomb elastoplastic constitutive model, and values ​​are assigned according to sandstone mechanical parameters. The tunneling equipment is simulated using a dynamically moving boundary, the anchor bolts are simulated using cable elements, and the steel arch frame is simulated using beam elements. The mesh is refined around the tunnel perimeter and gradually thinned outwards. The tunneling speed is defined. Anchor bolt preload Support lag distance It is an adjustable parameter.

[0162] Following step 2, a multi-source real-time data acquisition system was deployed: sensors were installed on the tunneling machine to collect thrust, torque, and propulsion speed per second; preload and borehole depth were collected on the anchor drilling rig; multiple displacement gauges were deployed in the surrounding rock, covering four depths: inside the loosened zone, in the middle of the anchoring zone, at the boundary of the anchoring zone, and in the original rock zone; one measuring hole was deployed on each of the roof and two sidewalls, collecting deep displacement data every 10 minutes; anchor bolt force gauges were evenly installed on the roof and two sidewalls, collecting anchor bolt axial force data every 10 minutes. All data were processed through spatiotemporal synchronization, outlier removal, and moving average filtering to form a continuous and stable data stream.

[0163] Following step 3, the digital twin model is dynamically updated based on data assimilation. The elastic modulus E, cohesion c, internal friction angle φ, and lateral pressure coefficient λ are selected as parameters to be corrected, generating an initial set of 50 samples. The ensemble Kalman filter algorithm is used to dynamically update the model parameters using measured displacement data. After three iterations, the parameters converge, and the updated parameters are: =17.8GPa =7.6MPa =41°、 =0.82.

[0164] The performance of the updated digital twin model in displacement prediction accuracy is shown in Table 2.

[0165] Table 2: Comparison of displacement prediction errors before and after digital twin model update (sandstone section)

[0166]

[0167] As shown in Table 2, after data assimilation and updating, the prediction accuracy of the digital twin model improved by more than 60%, verifying the effectiveness of the Kalman filter parameter update method in step 3.

[0168] Following step 4, construction parameters are optimized based on the updated digital twin model. For the predicted surrounding rock response in the next tunneling cycle, a genetic algorithm is used to solve for the optimal combination of construction parameters. The weight coefficients of the objective function are set according to the mine's emphasis on safety and efficiency. =0.5、 =0.3、 =0.2. The allowable displacement threshold U_lim is set to 50mm according to engineering specifications. The genetic algorithm population size is set to 40, and the maximum number of iterations is 50 generations.

[0169] After 20 iterations and convergence, the optimal combination of construction parameters was obtained: tunneling speed. =2.4m / h, anchor bolt preload =85kN, support lag distance =1.2m. Evaluation index of support effect under this parameter combination. =0.31, which is greater than the minimum allowable value. =0.25.

[0170] Table 3 shows a comparison of the construction effects of various schemes in the sandstone section from day 1 to day 20.

[0171] Table 3 Comparison of construction effects of various schemes in the sandstone section (days 1-20)

[0172]

[0173] As shown in Table 3, all three schemes can meet the basic safety requirements in the sandstone section. However, the method of the present invention has shown advantages in terms of surrounding rock control and construction efficiency. The roof displacement is reduced by 25.5% compared with Comparative Example 1, and the cycle time is shortened by 15.6%.

[0174] II. Construction stage of mudstone section (days 21-40, tunneling distance 60-120m);

[0175] From day 21 onwards, the tunnel entered the mudstone section. The mudstone had an elastic modulus of only 3.2 GPa, a cohesion of 1.5 MPa, and an internal friction angle of 28°, indicating a significant deterioration in lithology. Comparative Example 1 was constructed according to the original design parameters, while in Comparative Example 2, the technicians, based on experience, reduced the tunneling speed to 1.5 m / h and increased the preload to 100 kN.

[0176] The digital twin system of this invention automatically triggers an ensemble Kalman filter update after collecting initial monitoring data of the mudstone section on day 21. The updated surrounding rock parameters are as follows: =2.9GPa =1.3MPa =26° =1.05. The results of the parameter update accuracy verification are shown in Table 4.

[0177] Table 4. Parameter Update Accuracy of Digital Twin Model for Mudstone Section

[0178]

[0179] As shown in Table 4, the relative error between the updated surrounding rock parameters and the indoor test values ​​is within 15%, and the model can effectively approximate the real surrounding rock condition.

[0180] Based on the updated model, advanced predictions were made. The results showed that if the original parameters were maintained during construction, the roof displacement in the next cycle would reach 63.5 mm, exceeding the allowable threshold of 50 mm, and the volume of the plastic zone would increase to 8.75 m³. The genetic algorithm was then re-optimized to obtain a new parameter combination: tunneling speed. =1.2m / h, anchor bolt preload =120kN, support lag distance =0.8m. Evaluation indicators for support effectiveness. =0.28, which meets the requirements.

[0181] Table 5 shows a comparison of the construction effects of different schemes in the mudstone section from day 21 to day 40.

[0182] Table 5 Comparison of construction effects of different schemes in mudstone section (days 21-40)

[0183]

[0184] Table 5 shows that the differences between the various schemes widened significantly after entering the mudstone section. In Comparative Example 1, due to parameter solidification, the roof displacement reached 72.3 mm, severely exceeding the limit, with 8 instances of displacement exceeding the limit and 6 instances of anchor bolt axial force exceeding the limit. Although Comparative Example 2 underwent manual adjustment, the roof displacement still reached 54.6 mm, exceeding the 50 mm threshold, due to limitations in experience-based judgment, and the cycle time was extended to 52 minutes. The method of this invention resulted in a roof displacement of only 42.1 mm, a reduction of 41.8% compared to Comparative Example 1, a 63.4% reduction in the plastic zone volume, and a shortened surrounding rock deformation convergence time to 8 days. No exceedances occurred throughout the entire construction process.

[0185] III. Verification of security early warning function;

[0186] To further verify the safety warning and manual intervention mechanism in step 7 of the present invention, a special verification of the warning function was carried out on the 35th to 40th day when traversing the fault fracture zone. The results are shown in Table 6.

[0187] Table 6. Verification Results of Security Early Warning Function

[0188]

[0189] As shown in Table 6, the system triggered a Level II warning on day 36, issuing a roof displacement warning four cycles in advance; and a volume ratio warning for the plastic zone was triggered on day 37. On day 38, due to the extremely fractured surrounding rock in the fault fracture zone, the optimization algorithm could not find a parameter combination that satisfied the constraints, and the system automatically triggered a Level III warning and suspended tunneling. After on-site inspection, the technical personnel decided to take measures such as increasing the density of anchor bolts and grouting reinforcement. After reinforcement, the system was re-optimized and automatic control was restored, successfully passing through the fault section.

[0190] IV. Comprehensive Comparative Analysis;

[0191] The overall results of the 40-day continuous tracking experiment are shown in Table 7.

[0192] Table 7 Comparison of the overall effects of each scheme (40-day trial period)

[0193]

[0194] V. Conclusion;

[0195] The results of a 40-day continuous tracking test show that the method of the present invention has the following significant advantages:

[0196] 1. Parameter update accuracy: The relative error between the updated surrounding rock parameters and the indoor test values ​​is less than 15%, and the displacement prediction accuracy of the digital twin model is improved by more than 60% compared with that before the update.

[0197] 2. Surrounding rock control effect: When geological conditions deteriorate significantly, the roof displacement is reduced by 41.8% compared with the fixed parameter method, the volume of the plastic zone is reduced by 63.4%, and the displacement exceedance rate is 0.

[0198] 3. Improved construction efficiency: The average cycle time is shortened by 4.4% compared to the fixed parameter method and by 8.5% compared to the experience-based control method, while saving 12.8% of support materials.

[0199] 4. Safety early warning capability: On average, early warnings are issued 3.5 cycles in advance, providing sufficient time for on-site handling; when optimization fails, tunneling is automatically suspended and manual intervention is requested to ensure that high-risk working conditions are under safe control.

[0200] 5. Dynamic adaptability: The parameter adjustment response time is less than 10 minutes, which is far better than the approximately 4 hours of manual experience-based control, realizing real-time adaptive adjustment of construction parameters to changes in surrounding rock conditions.

[0201] 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 digitally driven method for real-time control of intelligent construction parameters in roadways, characterized in that, Includes the following steps: Step 1: Construct an initial digital twin model for tunnel construction; establish a three-dimensional numerical model including surrounding rock, tunneling equipment, and support structure as a virtual mirror of the digital twin, and define adjustable construction parameters; Step 2: Multi-source real-time data acquisition and preprocessing; real-time data acquisition from tunneling machines, anchor drilling rigs and surrounding rock monitoring sensors, preprocessed to form a continuous and stable data stream to drive the evolution of the digital twin; Step 3: Dynamic update of the digital twin model based on data assimilation; Select the variables to be corrected, use the ensemble Kalman filter algorithm, and dynamically update the parameters using measured displacement data to make the digital twin model approximate the real surrounding rock condition; Step 4: Optimization of construction parameters based on digital twin advance prediction; Based on the updated digital twin model, simulate the surrounding rock response of the next construction cycle, define the comprehensive control objective function, and use tunneling speed, anchor bolt preload, and support lag distance as design variables, and use a genetic algorithm to solve for the optimal combination of construction parameters; Step 5: Real-time control command issuance and equipment linkage; convert the optimized construction parameters into equipment control commands and send them to the PLC control systems of the tunneling machine and anchor drilling rig in real time to realize the reverse control of the physical equipment by the digital twin; Step 6: Iterative Cycle and Digital Twin Update; After each tunneling cycle is completed, steps 2 to 5 are immediately repeated to achieve synchronous evolution of the digital twin model and the physical tunnel, and the construction parameters are dynamically and adaptively adjusted according to changes in the surrounding rock conditions.

2. The method for real-time control of intelligent tunnel construction parameters driven by digital technology according to claim 1, characterized in that, It also includes step 7: a safety early warning and manual intervention mechanism; When the predicted surrounding rock response exceeds the attention threshold or safety threshold, the system issues a graded warning and, if necessary, automatically suspends tunneling and prompts for manual intervention.

3. The method for real-time control of intelligent tunnel construction parameters driven by digital technology according to claim 1, characterized in that, In step 1, the process of constructing the initial digital twin model of the tunnel construction is as follows: S11: The model size range is 5 times the width of the tunnel to ensure that the boundary effect does not affect the calculation results; Boundary condition settings: the bottom is fixed with vertical displacement, the top is applied with vertical principal stress, and the side is applied with horizontal principal stress; S12: The surrounding rock adopts the Mohr-Coulomb elastoplastic constitutive model, and different rock layers are assigned corresponding mechanical parameters according to the stratigraphic distribution, and an initial geostress field is set. S13: The tunneling equipment is simulated using dynamic moving boundaries, advancing gradually with each tunneling cycle; the anchor bolts are simulated using cable elements, and the steel arch frame is simulated using beam elements, arranged according to the design drawings. The anchor bolt unit parameters are set according to the actual product specifications and calibrated through indoor pull-out tests to truly reflect the interaction characteristics between the anchor bolt and the surrounding rock. S14: Densify the grid around the tunnel to ensure the calculation accuracy in areas with large stress gradients in the surrounding rock, and gradually thin it outwards to improve calculation efficiency. S15: The model defines adjustable construction parameters: tunneling speed. Anchor bolt preload Support lag distance This is used for subsequent optimization and regulation.

4. The method for real-time control of intelligent tunnel construction parameters driven by digital technology according to claim 1, characterized in that, In step 2, the multi-source real-time data acquisition and preprocessing process is as follows: S21: Sensors are installed on the tunneling machine to collect tunneling parameters once per second, including thrust, torque, and propulsion speed; S22: Collect the preload applied value and drilling depth on the anchor drilling rig; S23: Install multi-point displacement gauges and anchor bolt force gauges in the surrounding rock of the roadway; the multi-point displacement gauges are installed to cover the interior of the loosened zone, the middle of the anchoring zone, the boundary of the anchoring zone, and the original rock zone. Multiple measuring holes are installed in each monitoring section, and deep displacement is collected every 10 minutes; the anchor bolt force gauges are installed at the ends of representative anchor bolts and are evenly distributed on the roof and both sides, and the axial force of the anchor bolts is collected every 10 minutes. S24: All collected data are synchronized in time and space. For outliers that are significantly outside the range, they are directly removed and replaced with the value of the previous valid moment. The moving average method is used for filtering and smoothing to form a continuous and stable real-time data stream.

5. The method for real-time control of intelligent construction parameters for roadways driven by digital technology according to claim 1, characterized in that, In step 3, the process of dynamically updating the digital twin model based on data assimilation is as follows: S31: Define the parameter vector ,in, For elastic modulus, For cohesion, For internal friction angle, The lateral pressure coefficient; an initial set of 50 samples is generated for the parameter to be corrected. The sample value range is set based on the results of indoor tests: Take 0.5 to 1.5 times the experimental value. Take 0.3 to 2.0 times the experimental value. Take the test value ±5°. Take a value of 0.5 to 1.5; S32: Run the numerical model for each sample to obtain the simulated displacement value of the monitoring point location. Define observation vector ,in The number of monitoring points For the first Measured displacement at each monitoring point; S33: Calculate the Kalman gain matrix according to the following formula. : ； In the formula, This is the covariance matrix between the parameters and the simulated observations; The covariance matrix of the simulated observations; The observation error covariance matrix; S34: Update each parameter sample according to the following formula: ； In the formula, For the first The parameter values ​​of each sample before the update; For the first The simulated displacement vector corresponding to each sample; This represents the residual between the measured displacement and the simulated displacement. S35: Iterate and update 3 to 5 times until the parameters converge; take the mean of the updated sample set as the true parameters of the current surrounding rock to obtain the updated digital twin model; the calculation time of each update must be completed before the next construction decision node; after the parameter update, a physical rationality test must be performed to ensure that it is within the preset range; if it exceeds the range, the boundary value is taken; the assimilation frequency is matched with the data acquisition frequency to ensure that the digital twin model always reflects the latest surrounding rock condition.

6. The method for real-time control of intelligent tunnel construction parameters driven by digital technology according to claim 1, characterized in that, In step 4, the optimization process of construction parameters based on digital twin-based advanced prediction is as follows: S41: Based on the updated digital twin model from step 3, simulate the surrounding rock response of the next tunneling cycle and predict the maximum displacement of the roof and the volume of the plastic zone; S42: Define the objective function according to the following formula. : ； In the formula, This represents the maximum displacement of the roof predicted by the model. The allowable displacement threshold; The volume of the plastic region; This represents the total volume of the anchorage area; This is the current construction cycle time; For reference cycle time; , , Let be the weighting coefficient, satisfying + The settings are determined based on the different levels of importance that projects place on safety and efficiency. S43: Calculate the volume of the plastic zone according to the following formula. : ； In the formula, This represents the total number of units near the anchorage zone. Let i be the volume of the i-th unit. This is the i-th indicator function; S44: at the tunneling speed Anchor bolt preload Support lag distance As design variables; to set engineering feasibility constraints: , , 1000; of which The safety factor for the anchor bolt preload. This refers to the maximum permissible advance speed of the tunneling machine. The minimum safe hysteresis distance; The anchor rod's yield strength; The cross-sectional area of ​​the anchor bolt; This is the safety factor for the preload force; S45: The support effectiveness evaluation index is defined according to the following formula: ； In the formula, This represents the total number of units within the anchorage zone. For the first Additional stress on the support of each unit; The vertical principal stress of the original rock; The evaluation index corresponding to the optimized parameter combination is required. If the conditions are not met, adjust the parameters and re-optimize. The critical evaluation threshold for roadway support stability; S46: Use a genetic algorithm to solve the objective function. The minimum optimal combination of construction parameters; in each generation of evolution, the generated parameter combination is first judged for constraints; the algorithm terminates when the change of the objective function value is less than a specified value for multiple consecutive generations or when the maximum number of iterations is reached; For the optimal combination of parameters in the final output The displacement and support effects are verified through a complete numerical model forward modeling process to ensure that they meet the requirements. If the verification fails, the algorithm parameters are adjusted and re-optimized to ensure that the output results are true and reliable.

7. The method for real-time control of intelligent construction parameters for roadways driven by digital technology according to claim 1, characterized in that, In step 5, the process of issuing real-time control commands and coordinating with the equipment is as follows: S51: Optimize the construction parameters obtained in step 4. The commands are converted into equipment control instructions and sent in real time to the tunneling machine PLC control system and the bolt drilling rig control system via industrial Ethernet. S52: The tunneling machine automatically adjusts its propulsion speed according to instructions. The anchor drilling rig automatically sets the preload to... On-site workers follow the optimized lag distance Organize support construction; S53: The equipment execution status is fed back to the digital twin system in real time; the industrial Ethernet communication adopts a redundant design to ensure that a single point of failure does not affect the issuance of commands; the equipment control system completes parameter adjustment within the specified control cycle and feeds back the execution results; If no feedback is received from the device multiple times in a row, the system determines that there is a communication failure and issues an alarm, prompting on-site personnel to check the network connection and device status; the total delay from the issuance of the command to the device response is controlled within the specified time.

8. The method for real-time control of intelligent construction parameters for roadways driven by digital technology according to claim 1, characterized in that, In step 6, the iterative process and digital twin update are as follows: S61: After each tunneling cycle is completed, immediately collect the latest monitoring data, update the digital twin model parameters, optimize the construction parameters for the next cycle, and issue instructions; S62: Through closed-loop iteration, an intelligent construction process control system with digital twin as its core is formed, realizing dynamic adaptive control of construction parameters.

9. The method for real-time control of intelligent construction parameters for roadways driven by digital technology according to claim 2, characterized in that, In step 7, the process of the safety early warning and manual intervention mechanism is as follows: S71: When the surrounding rock response predicted in step 4 exceeds the set attention threshold, the system displays an early warning on the visualization interface, prompting on-site technicians to strengthen monitoring and attention. S72: When the predicted response exceeds the safety threshold, or the optimization algorithm cannot find a parameter combination that satisfies the constraints, the system displays a high-level warning, automatically suspends tunneling, and prompts for manual intervention. S73: After manual intervention, on-site technicians, based on the revealed geological conditions, decide whether to continue with the original plan, adjust and optimize the parameter range for recalculation, and take enhanced support measures. Automatic control will be restored only after safety is confirmed.

Images