Tunnel excavation face back break real-time efficient identification method based on fusion of point cloud data and handheld scanner

Abstract

The invention discloses a real-time efficient identification method for back break of a tunnel excavation face based on fusion of point cloud data and a handheld scanner, and relates to the technical field of tunnel construction quality safety monitoring, and the method comprises the following steps: coupling data collection, coupling modeling, coupling determination, and real-time mutual feedback closed loop. According to the method, energy field parameter quantification water seepage noise, four-dimensional dynamic tensor model construction, back break-water seepage-stress coupling judgment and a real-time mutual feedback closed loop mechanism are utilized, so that the technical problems that point cloud details are lost, multiple tasks are isolated and uncorrelated, and a recognition result is lack of risk guidance due to the water seepage noise in water-rich stratum tunnel back break recognition are solved; the method achieves the precise calculation of the back break amount, the dynamic distinguishing of the risk priorities, and the integration of real-time recognition, risk tracing and precise disposal, and remarkably improves the construction safety coefficient and disposal efficiency.

Description

Technical Field

[0001] This invention relates to the field of tunnel construction quality and safety monitoring technology, specifically a real-time and efficient method for identifying over-excavation and under-excavation at the tunnel excavation face based on the fusion of point cloud data and a handheld scanner. Background Technology

[0002] Identifying over-excavation and under-excavation at the tunnel excavation face is a crucial step in ensuring construction quality and safety, directly impacting tunnel structural stability, construction costs, and schedule. With the development of 3D scanning technology, the identification method that integrates point cloud data with handheld scanners has gradually become mainstream. Point cloud data provides global 3D information about the excavation face, while handheld scanners offer the advantage of flexible and convenient supplementation of local details. The fusion of these two methods can balance comprehensiveness and accuracy in identification, meeting the demands of real-time and efficient construction.

[0003] However, existing over- and under-excavation identification technologies based on the fusion of point cloud data and handheld scanners still have significant limitations. First, water seepage is particularly prevalent in water-rich strata at the tunnel face. The water mist generated by seepage introduces regional noise into the fused data. While denoising algorithms remove this noise, they neglect the underlying physical environment information. Furthermore, the denoising process can easily lead to the loss of point cloud details in the seepage area, affecting the accuracy of over- and under-excavation calculations. Second, existing technologies treat over- and under-excavation identification, seepage treatment, and surrounding rock stability assessment as independent tasks. They obtain over- and under-excavation amounts solely through geometric comparison between the design contour and measured data, without considering the impact of surrounding rock saturation and softening caused by seepage on the risk level of over- and under-excavation, and thus cannot distinguish the safety priorities of over- and under-excavation in different areas. Third, the fusion of handheld scanners and point cloud data focuses only on data complementarity to improve geometric accuracy, without exploring the environmental and structural correlation features hidden in the fused data. This results in identification results that only reflect geometric deviations and cannot provide accurate risk-oriented references for construction decisions.

[0004] The aforementioned deficiencies make it difficult for existing fusion identification methods to meet the core requirements of real-time identification, risk prediction, and precise handling of tunnels in water-rich strata. There is an urgent need for a new identification technology that can fully utilize data features and correlate physical environment with risk level. In view of this, a real-time and efficient identification method for over-excavation and under-excavation of tunnel excavation face based on the fusion of point cloud data and handheld scanner is provided to overcome the above problems. Summary of the Invention

[0005] The purpose of this invention is to provide a real-time and efficient method for identifying over-excavation and under-excavation of tunnel excavation faces based on the fusion of point cloud data and handheld scanners, so as to solve the problems mentioned in the background art.

[0006] To address the aforementioned technical problems, this invention provides a real-time and efficient method for identifying over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner, comprising the following steps:

[0007] Coupled data acquisition: By embedding a photon energy monitoring module into a 3D laser point cloud acquisition instrument and a handheld scanner, a bidirectional mechanism for energy sensing and acquisition parameter self-adaptation is constructed to convert water seepage noise into quantifiable energy field parameters, complete the synchronous binding of geometric and energy data, perform energy compensation only for invalid noise, and retain all valid water seepage noise signals.

[0008] Coupled modeling: Construct a four-dimensional dynamic tensor model that includes a dynamic coupling sub-module of energy field, pore water pressure and stress field. Establish a three-dimensional calibration tensor database of energy-environment-mechanics through cross-scene calibration experiments to complete the dynamic mapping of energy field parameters to physical environment and stress field.

[0009] Coupling determination: Based on the four-dimensional dynamic tensor model, the composite evaluation unit is divided according to the two dimensions of seepage intensity and stress level. The regional over-excavation and under-excavation volume and the dynamic bearing capacity of the surrounding rock are calculated. The risk priority is determined by the risk correlation factor and the risk root cause label is marked.

[0010] Real-time feedback closed loop: Set dual anomaly trigger conditions to trigger handheld scanner fixed-point enhanced scanning and update tensor model, fit trend curve based on model data to predict risk changes, output accurate construction suggestions and optimize data acquisition resource allocation.

[0011] Furthermore, the photon energy monitoring module achieves the following functions: real-time capture of energy attenuation coefficient, energy distribution entropy, and energy time attenuation rate; the self-adaptive rules for acquisition parameters are: point cloud acquisition frequency of 10Hz and handheld scanning resolution of 0.5mm in weak seepage scenarios, point cloud acquisition frequency of 15Hz and handheld scanning resolution of 0.3mm in medium seepage scenarios, and point cloud acquisition frequency of 20Hz and handheld scanning resolution of 0.2mm in strong seepage scenarios, with dynamic focusing mode activated; geometric-energy data are bound through a photon propagation path tracing algorithm with a synchronization error ≤5ms, and energy compensation is only performed for invalid noise with an energy distribution entropy ≥1.2.

[0012] Furthermore, the cross-scenario calibration experiment selected five types of surrounding rocks: sandstone, shale, granite, limestone, and mudstone. For each type, three sets of 50cm×50cm×50cm standard samples were prepared to simulate a combination of three seepage intensities, three types of surrounding rock saturation, and three types of pore water pressure. After stabilizing for 30 minutes, the energy field parameters, measured values ​​of pore water pressure, and measured values ​​of stress were recorded simultaneously. The database clearly defines the correspondence between seepage intensity levels and energy field parameters. A surrounding rock saturation calculation model was constructed based on the energy time decay rate, and a pore water pressure coupling model was constructed based on the energy decay coefficient and the energy time decay rate.

[0013] Furthermore, the dynamic mapping of the stress field includes: via the formula:

[0014] σ total =γ×h,

[0015] The total natural stress is calculated using the formula: γ is the unit weight of the surrounding rock, and h is the tunnel depth.

[0016] σ eff =σ total -P×k σ ;

[0017] Calculate the effective stress, where P is the pore water pressure and k is the pore water pressure. σ The stress correction factor is used to generate the global stress field distribution using an energy-weighted kriging interpolation algorithm, with a stress value resolution ≤ 0.01 MPa.

[0018] Furthermore, the geometric correction of the four-dimensional dynamic tensor model is achieved through the formula:

[0019] Δd = Δd1 + Δd2,

[0020] The original three-dimensional coordinates are corrected based on the total geometric deviation. The seepage intensity, surrounding rock saturation, pore water pressure, and stress value are used as attribute dimensions and dynamically bound to the corrected coordinates and energy field parameters through a data linked list structure.

[0021] Furthermore, the composite evaluation unit is divided into 9 categories, and the stress level classification standard is high stress zone ≥0.5MPa, medium stress zone 0.2-0.5MPa, and low stress zone <0.2MPa. The over-excavation and under-excavation amounts are calculated using the energy-stress dual-weighted volume algorithm combined with the three-dimensional mesh integration method, with a calculation error ≤1.5%. Over-excavation is taken as a positive value, and under-excavation is taken as a negative value.

[0022] Furthermore, the formula for the risk association factor is:

[0023]

[0024] ΔV represents the over- or under-excavation amount. For safety margin, K is the stress concentration factor.

[0025] Furthermore, the dual anomaly triggering conditions are an energy distribution entropy change rate ≥30% or a stress concentration coefficient change rate ≥25%, and the fixed-point enhancement scanning parameters are a scanning frequency of 30Hz, a resolution of 0.1mm, and a scanning range extended outward by 50cm; the trend curve is based on the tensor model data of the last 5 times and uses cubic polynomial fitting to predict the risk changes in the next 3 minutes; the acquisition resource optimization strategy is a point cloud acquisition frequency of 5Hz for low-risk and low-stress areas, a handheld scanning resolution of 0.5mm, and a full-process recognition cycle of ≤2min.

[0026] Compared with the prior art, the beneficial effects of the present invention are:

[0027] Completely solves the defects in noise processing: by quantifying seepage noise through energy field parameters, it not only fully preserves the details of the point cloud, but also penetrates and obtains dual information of environmental and mechanical parameters, significantly improving the accuracy of over-excavation and under-excavation calculation;

[0028] Completely solve the problem of isolated tasks: Over-excavation and under-excavation identification, seepage treatment and surrounding rock stability assessment are linked in real time through dynamic tensor models. Risk judgment takes into account geometric deviation, seepage impact and stress concentration at the same time, and accurately distinguishes the safety priority of different areas.

[0029] Completely solves the lack of risk orientation: The identification results not only include geometric deviation data, but also provide stress field distribution, risk root cause labels and combined construction suggestions, which fully meet the core needs of real-time identification, risk tracing and precise treatment of tunnels in water-rich strata.

[0030] Stress field penetration correlation effect: It realizes cross-physical domain mapping between photon energy signals and rock mass stress field, without the need to deploy additional stress sensors. The global stress distribution can be obtained at low cost by simply upgrading the algorithm of existing scanning equipment.

[0031] Risk Root Cause Tracing Effect: Through over- and under-excavation-seepage-stress coupling analysis, the dominant risk factors are accurately identified, avoiding blind handling during construction and reducing ineffective construction costs;

[0032] Dual geometric correction effect: Simultaneously eliminates water mist interference and measurement deviations caused by rock mass micro-deformation, breaking through the limitation of existing technologies that can only correct environmental interference, and significantly improving geometric measurement accuracy compared to existing technologies;

[0033] Stress concentration early warning effect: Predict areas of rapid stress concentration in advance, avoid the risk of surrounding rock collapse caused by over- or under-excavation due to stress superposition, and significantly improve the construction safety factor. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the principle of the real-time and efficient identification method for over-excavation and under-excavation of tunnel excavation face based on the fusion of point cloud data and handheld scanner of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figure 1 The present invention provides a technical solution:

[0037] See Figure 1As shown, an embodiment of a real-time and efficient method for identifying over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner is presented:

[0038] This invention transforms seepage noise into an energy field signal through noise energization, coupled modeling of energy and stress fields, coupled structural response, and real-time feedback optimization. This signal further penetrates to the rock mechanics level, simultaneously achieving geometric correction, environmental perception, stress assessment, and risk prediction, as detailed below:

[0039] I. Coupled Data Acquisition: Bidirectional Feedback Acquisition of Noise-Energy-Geometry:

[0040] A self-adaptive two-way mechanism for energy sensing and parameter acquisition is constructed to convert seepage noise into quantifiable energy field parameters, thereby addressing the root causes of noise rejection issues such as loss of detail and neglect of environmental information.

[0041] Data Acquisition System Upgrade: A photon energy monitoring module was implanted into the 3D laser point cloud acquisition instrument and handheld scanner, which are responsible for global data acquisition and local detail acquisition, respectively. This module is implemented through software algorithm upgrades of existing equipment without the need for additional hardware, and specifically achieves two main functions:

[0042] Energy field parameter acquisition: Real-time capture of the energy attenuation coefficient, energy distribution entropy, and energy time decay rate of photons during propagation due to water seepage and mist. The acquisition frequency remains consistent with the original acquisition frequency of the equipment. Parameter self-adaptation: Dynamically adjust the equipment's acquisition parameters based on the energy attenuation coefficient μ. The specific parameter adjustment rules are as follows:

[0043] When μ∈[0.1-0.3]: In a weak seepage scenario, the point cloud acquisition frequency is maintained at 10Hz, and the handheld scanning resolution is set to 0.5mm;

[0044] When μ∈[0.3-0.6]: In the water seepage scenario, the point cloud acquisition frequency is increased to 15Hz, and the handheld scanning resolution is increased to 0.3mm;

[0045] When μ∈[0.6-1.0]: In strong water seepage scenarios, the point cloud acquisition frequency is increased to 20Hz, the handheld scanning resolution is increased to 0.2mm, and the dynamic focusing mode of the handheld scanner is activated to focus on the energy attenuation abnormal area, that is, the area where μ exceeds the mean of the adjacent area by ±30%.

[0046] Geometric-energy data synchronization binding: Data binding is achieved through a photon propagation path tracing algorithm. This algorithm calculates the propagation distance based on the difference between photon emission time and reception time. Combined with a three-dimensional coordinate system, each three-dimensional coordinate point is uniquely bound to the corresponding energy field parameters and noise grayscale value. The synchronization error is controlled within ≤5ms. After binding, only invalid noise with energy distribution entropy ≥1.2 is processed for energy compensation, and all valid energy attenuation signals >1.2 are retained. For water seepage noise, no rejection operation is performed.

[0047] It should be added that: during the acquisition process, all signals related to seepage are fully preserved, their physical characteristics are quantified through energy parameters, and the acquisition parameters are dynamically adapted to changes in the energy field to avoid the loss of geometric details from the source.

[0048] II. Energy-Geometry-Environment-Stress Coupling Modeling: Construction of a Four-Dimensional Dynamic Tensor Model

[0049] Based on the energy-geometry-environment coupling, a dynamic coupling submodule of energy field-pore water pressure-stress field is pre-defined to construct a four-dimensional dynamic tensor model. By penetrating to the rock mechanics level through energy field parameters, it fundamentally solves the problems of missing environmental-structural mechanics correlation and single fusion dimension.

[0050] Energy field parameter-physical environment mapping mechanism:

[0051] Cross-scenario calibration experiment:

[0052] Experimental preparation: Five typical water-rich strata surrounding rocks, namely sandstone, shale, granite, limestone and mudstone, were selected. Three sets of standard samples with a size of 50cm×50cm×50cm were made for each type of surrounding rock. A simulated tunnel excavation face test bench was set up in the laboratory, surrounding rock samples were laid out, and a controllable seepage device, pore water pressure sensor, stress sensor and energy field parameter acquisition device were configured.

[0053] Experimental procedure:

[0054] Three combined scenarios were simulated: seepage intensities of 0.2-1 L / min, 1-3 L / min, and 3-6 L / min; three surrounding rock saturations of 30%, 60%, and 90%; and three pore water pressures of 0.1 MPa, 0.5 MPa, and 1.0 MPa. After each scenario stabilized for 30 minutes, the energy field parameters were recorded simultaneously. Measured values ​​of pore water pressure and stress;

[0055] Database establishment: A three-dimensional calibration tensor database of energy-environment-mechanics was established based on experimental data, and the parameter mapping relationship was clarified.

[0056] Permeability strength grade Q:

[0057] Q1 (0.2-1 L / min) corresponds to μ∈[0.1-0.3] and H∈[0.5-0.8];

[0058] Q2 (1-3 L / min) corresponds to μ∈[0.3-0.6] and H∈[0.8-1.0];

[0059] Q3 (3-6 L / min) corresponds to μ∈[0.6-1.0] and H∈[1.0-1.2];

[0060] Surrounding rock saturation S: based on energy time decay rate Construct a computational model, The larger the value, the more photon energy the surrounding rock absorbs, and the higher the saturation S; where S is the current saturation of the surrounding rock, and S0 is the natural saturation of the surrounding rock.

[0061] Pore ​​water pressure P: Constructing a coupled model of energy parameters and pore water pressure:

[0062]

[0063] In the formula: μ is the energy decay coefficient, k p Energy attenuation - water pressure coefficient k is the energy decay rate over time. e The energy time decay factor is the water pressure coefficient.

[0064] Real-time dynamic mapping: This involves mapping the collected energy field parameters μ, H, ... Input a preset three-dimensional calibration tensor database of energy, environment, and mechanics, and perform inverse operations on the multidimensional parameters through tensor decomposition algorithm to inversely deduce the seepage intensity Q, surrounding rock saturation S, and pore water pressure P corresponding to each three-dimensional coordinate point, with the mapping error controlled within ≤3%.

[0065] Energy field-stress field dynamic coupling submodule:

[0066] Stress field inference mechanism: Based on the effective stress principle (Terzaghi principle), combined with the pre-set natural total stress σ of the surrounding rock type. total Back-calculation of the effective stress σ at each coordinate point eff Natural total stress σ total Calculated based on tunnel depth:

[0067] σ total =γ×h,

[0068] γ is the unit weight of the surrounding rock, and h is the tunnel depth; the effective stress calculation formula is:

[0069] σ eff =σ total -P×k σ ;

[0070] In the formula: P is the pore water pressure, k σ Stress correction factors (determined by lithology: granite 0.9, sandstone 0.85, shale 0.75, limestone 0.8, mudstone 0.7);

[0071] Stress field distribution generation: The global stress field distribution σ of the excavation face is generated using an energy-weighted kriging interpolation algorithm. stress The algorithm uses the energy attenuation coefficient μ as the weight, and the larger the μ, the higher the interpolation weight of the region, to ensure the accuracy of stress calculation in the seepage area; the stress value resolution is set to ≤0.01MPa, and the stress field boundary is adjusted synchronously with the boundary of the seepage intensity Q level region.

[0072] Generation of four-dimensional dynamic tensor models: A geometric correction algorithm driven by energy and stress is used to construct four-dimensional dynamic tensor models. The specific implementation method is as follows:

[0073] Geometric Correction: Incorporating Stress Influence: The geometric deviation Δd1 caused by water mist is calculated by using the energy attenuation coefficient μ, and simultaneously by using the stress value σ stress The geometric deviation Δd2 caused by the micro-deformation of the rock mass is calculated, and the total geometric deviation is:

[0074] Δd = Δd1 + Δd2,

[0075] The original 3D coordinates X, Y, Z are corrected based on the total geometric deviation to obtain the corrected coordinates X+Δd. x ,Y+Δd y ,Z+Δd z This completely eliminates the geometric measurement deviation caused by the combined effects of water mist interference and micro-deformation of the rock mass; where Δd1 is determined by μ and the photon propagation distance, and Δd2 is determined by σ stress Determined by the stress-deformation coefficient;

[0076] Feature binding: The seepage intensity Q, surrounding rock saturation S, pore water pressure P, and stress value are used as attribute dimensions of the tensor and dynamically bound to the corrected geometric coordinates and energy field parameters. The binding is implemented using a data linked list structure.

[0077] III. Coupling Determination of Over-excavation / Under-excavation, Structural Response, and Risk Priority:

[0078] Dynamic calculation of core parameters:

[0079] Regional over- or under-excavation amount ΔV:

[0080] Assessment area delineation: Based on a four-dimensional dynamic tensor model, the assessment area is delineated according to two dimensions: seepage intensity and stress level. The stress level classification criteria are: high stress zone ≥ 0.5 MPa, medium stress zone 0.3 MPa ≤ σ. stress 0.5MPa, low stress zone σstress 0.3MPa, forming 9 types of composite evaluation units such as Q1-low stress, Q1-medium stress, and Q1-high stress, with the unit boundaries adjusted in real time according to water seepage diffusion and stress redistribution;

[0081] Over- and under-excavation calculation: The energy-stress dual-weighted volumetric algorithm is used for calculation, with energy attenuation coefficient μ and stress value σ. stress The weights are dynamically allocated based on the importance of the area to ensure the highest calculation accuracy for areas with superimposed seepage and high stress. The volume calculation adopts the three-dimensional grid integration method, which compares the corrected three-dimensional coordinates with the tunnel design outline, and the ΔV calculation error is controlled within ≤1.5%. ΔV is taken as a positive value when over-excavation occurs and as a negative value when under-excavation occurs.

[0082] Dynamic bearing capacity of surrounding rock σ d Introducing the saturation-water pressure-stress coupling coefficient, β, which is determined by the surrounding rock saturation S, pore water pressure P, and stress value σ. stress With natural total stress σ total Together, the larger β is, the greater σ is. d The lower; σ d σ0 represents the dynamic bearing capacity of the surrounding rock, which is updated in real time with the tensor model, while σ0 represents the natural bearing capacity of the surrounding rock.

[0083] Structural response threshold σ th Based on the dynamic setting of tunnel design support strength, surrounding rock dynamic bearing capacity and stress concentration coefficient, the more severe the stress concentration, the more stringent the structural response threshold.

[0084] Risk priority is dynamically determined:

[0085] Risk-related factor R f Defined as over- or under-excavation amount ΔV, safety margin The coupling result with the stress concentration factor K, the three factors synergistically amplify the risk weight of high-risk areas, as shown in the formula:

[0086]

[0087] Priority division:

[0088] High-risk priority: satisfying R f ≥σ th ×0.8 and K≥1.2 correspond to a scenario with large over- or under-excavation deviation and severe stress concentration, requiring immediate work stoppage and handling; simultaneously label the risk root cause, with the judgment logic as follows: when Q≥Q2 and P≥0.5MPa, label it as seepage-dominant, and when K≥1.2, label it as superposition-dominant;

[0089] Medium risk priority: satisfying σ th ×0.4≤R f <σ th×0.8 corresponds to scenarios with small over- or under-excavation deviations and no stress concentration, and should be handled according to standard procedures;

[0090] Output results: Synchronously output corrected over- and under-excavation amounts, seepage intensity, surrounding rock saturation, pore water pressure, stress field distribution, dynamic bearing capacity, structural response threshold, risk priority, and risk root cause labels to form a dynamic risk report. The report includes a composite area boundary label map, risk development trend curve, and stress concentration hotspot map. All output data are in CSV format, and graphic files are in PNG format.

[0091] IV. Real-time Feedback Closed Loop: Energy Field - Stress Field - Acquisition - Judgment - Optimization:

[0092] By incorporating stress field anomalies into the triggering conditions, a real-time feedback closed loop is constructed, integrating energy field and stress field dual monitoring, self-adaptive acquisition parameters, optimized judgment results, and precise output of construction suggestions. This achieves integrated identification, prediction, source tracing, and handling.

[0093] Closed-loop operation mechanism:

[0094] Dual anomaly trigger: Set an anomaly trigger condition. When a certain area meets any of the following conditions, the handheld scanner's fixed-point intensified scanning mode will be immediately triggered: ① Energy distribution entropy change rate ② Stress concentration factor variation rate The fixed-point enhancement scanning parameters were set as follows: scanning frequency 30Hz, scanning resolution 0.1mm, and scanning range extending 50cm outward from the abnormal area; after the scan was completed, all mechanical and environmental parameters in the tensor model were updated synchronously.

[0095] Risk prediction and precise construction recommendations: Based on the data from the last 5 tensor models at 0.5-second intervals, a cubic polynomial was used to fit the three-dimensional trend curve of over- and under-excavation volume-stress-risk. The risk changes within the next 3 minutes were predicted by extrapolating the curve. The prediction results correspond to two types of construction recommendations: ① If the predicted R... f ≥σ th ×0.8 and K≥1.2, the recommended combination of output shutdown + local grouting reinforcement (for water seepage) + stress relief holes (for stress concentration) is specified, with the grouting pressure set at 1.5-2.0MPa, the stress relief hole diameter set at 50mm, and the hole depth set at 1.5m; ② If the predicted R f The risk level has been raised to medium, and recommendations for increased monitoring and support density have been issued. The monitoring frequency has been increased to once every 30 seconds, and the support density has been increased by 20% based on the original design.

[0096] Resource optimization: The acquisition frequency is automatically reduced in low-risk and low-stress areas, set to 5Hz for point cloud and 0.5mm for handheld scanning resolution; the enhanced acquisition mode is maintained in high-risk and high-stress areas; the entire recognition cycle is controlled within ≤2min, improving computing power utilization.

[0097] Summarize:

[0098] Completely solves the defects in noise processing: by quantifying seepage noise through energy field parameters, it not only fully preserves the details of the point cloud, but also penetrates and obtains dual information of environmental and mechanical parameters, significantly improving the accuracy of over-excavation and under-excavation calculation;

[0099] Completely solve the problem of isolated tasks: Over-excavation and under-excavation identification, seepage treatment and surrounding rock stability assessment are linked in real time through dynamic tensor models. Risk judgment takes into account geometric deviation, seepage impact and stress concentration at the same time, and accurately distinguishes the safety priority of different areas.

[0100] Completely solves the lack of risk orientation: The identification results not only include geometric deviation data, but also provide stress field distribution, risk root cause labels and combined construction suggestions, which fully meet the core needs of real-time identification, risk tracing and precise treatment of tunnels in water-rich strata.

[0101] Stress field penetration correlation effect: It realizes cross-physical domain mapping between photon energy signals and rock mass stress field, without the need to deploy additional stress sensors. The global stress distribution can be obtained at low cost by simply upgrading the algorithm of existing scanning equipment.

[0102] Risk Root Cause Tracing Effect: Through over- and under-excavation-seepage-stress coupling analysis, the dominant risk factors are accurately identified, avoiding blind handling during construction and reducing ineffective construction costs;

[0103] Dual geometric correction effect: Simultaneously eliminates water mist interference and measurement deviations caused by rock mass micro-deformation, breaking through the limitation of existing technologies that can only correct environmental interference, and significantly improving geometric measurement accuracy compared to existing technologies;

[0104] Stress concentration early warning effect: Predict areas of rapid stress concentration in advance, avoid the risk of surrounding rock collapse caused by over- or under-excavation due to stress superposition, and significantly improve the construction safety factor.

Claims

1. A real-time and efficient method for identifying over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner, characterized in that... Includes the following steps: Coupled data acquisition: By embedding a photon energy monitoring module into a 3D laser point cloud acquisition instrument and a handheld scanner, a bidirectional mechanism for energy sensing and acquisition parameter self-adaptation is constructed to convert water seepage noise into quantifiable energy field parameters, complete the synchronous binding of geometric and energy data, perform energy compensation only for invalid noise, and retain all valid water seepage noise signals. Coupled modeling: Construct a four-dimensional dynamic tensor model that includes a dynamic coupling sub-module of energy field, pore water pressure and stress field. Establish a three-dimensional calibration tensor database of energy-environment-mechanics through cross-scene calibration experiments to complete the dynamic mapping of energy field parameters to physical environment and stress field. Coupling determination: Based on the four-dimensional dynamic tensor model, the composite evaluation unit is divided according to the two dimensions of seepage intensity and stress level. The regional over-excavation and under-excavation volume and the dynamic bearing capacity of the surrounding rock are calculated. The risk priority is determined by the risk correlation factor and the risk root cause label is marked. Real-time feedback closed loop: Set dual anomaly trigger conditions to trigger handheld scanner fixed-point enhanced scanning and update tensor model, fit trend curve based on model data to predict risk changes, output accurate construction suggestions and optimize data acquisition resource allocation.

2. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: The photon energy monitoring module performs the following functions: real-time capture of energy decay coefficient, energy distribution entropy, and energy time decay rate; the self-adaptive rules for acquisition parameters are: point cloud acquisition frequency of 10Hz and handheld scanning resolution of 0.5mm in weak seepage scenarios, point cloud acquisition frequency of 15Hz and handheld scanning resolution of 0.3mm in medium seepage scenarios, and point cloud acquisition frequency of 20Hz and handheld scanning resolution of 0.2mm in strong seepage scenarios, with dynamic focusing mode activated; geometric-energy data are bound through a photon propagation path tracing algorithm with a synchronization error ≤5ms, and energy compensation is only performed for invalid noise with energy distribution entropy ≥1.

2.

3. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: Cross-scenario calibration experiments selected five types of surrounding rocks: sandstone, shale, granite, limestone, and mudstone. Three sets of 50cm×50cm×50cm standard samples were made for each type. The experiments simulated a combination of three permeability intensities, three types of surrounding rock saturation, and three types of pore water pressure. After stabilization for 30 minutes, the energy field parameters, measured values ​​of pore water pressure, and measured values ​​of stress were recorded simultaneously. The database clearly defines the correspondence between permeability intensity levels and energy field parameters. A surrounding rock saturation calculation model was constructed based on the energy time decay rate, and a pore water pressure coupling model was constructed based on the energy decay coefficient and the energy time decay rate.

4. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: Dynamic mapping of stress field includes: via the formula: s total =γ×h, The total natural stress is calculated using the formula: γ is the unit weight of the surrounding rock, and h is the tunnel depth. s eff =s total -P×k σ ; Calculate the effective stress, where P is the pore water pressure and k is the pore water pressure. σ The stress correction factor is used to generate the global stress field distribution using an energy-weighted kriging interpolation algorithm, with a stress value resolution ≤ 0.01 MPa.

5. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: The geometric correction of the four-dimensional dynamic tensor model is achieved through the formula: Δd = Δd1 + Δd2, The original three-dimensional coordinates are corrected based on the total geometric deviation. The seepage intensity, surrounding rock saturation, pore water pressure, and stress value are used as attribute dimensions and dynamically bound to the corrected coordinates and energy field parameters through a data linked list structure.

6. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: The composite evaluation unit is divided into 9 categories. The stress level classification standard is high stress zone ≥0.5MPa, medium stress zone 0.2-0.5MPa, and low stress zone <0.2MPa. The over-excavation and under-excavation volume are calculated by energy-stress dual-weighted volume algorithm combined with three-dimensional mesh integration method. The calculation error is ≤1.5%. Over-excavation is taken as positive value and under-excavation is taken as negative value.

7. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: The formula for risk association factors is: ΔV represents the over- or under-excavation amount. For safety margin, K is the stress concentration factor.

8. The method for real-time and efficient identification of over-excavation and under-excavation at tunnel excavation faces based on the fusion of point cloud data and a handheld scanner as described in claim 1, characterized in that: The dual anomaly triggering conditions are an energy distribution entropy change rate ≥ 30% or a stress concentration coefficient change rate ≥ 25%. The fixed-point enhancement scanning parameters are a scanning frequency of 30Hz, a resolution of 0.1mm, and a scanning range extended outward by 50cm. The trend curve is based on the tensor model data of the last 5 times and uses cubic polynomial fitting to predict the risk changes in the next 3 minutes. The acquisition resource optimization strategy is to acquire point cloud data of low-risk and low-stress areas at a frequency of 5Hz and a handheld scanning resolution of 0.5mm, with a full-process recognition cycle of ≤ 2 minutes.

Images