Collaborative management and control method and system for underground engineering geological safety risks throughout the entire life cycle

Abstract

The present invention discloses a method and system for collaboratively managing underground engineering geological safety risks throughout the entire life cycle, including: identifying the regional geological structure and geological risk markers in the target construction area and constructing a regional geological digital model; obtaining regional construction monitoring information, combining it with the regional geological digital model to perform construction risk analysis and cavity prediction, and obtaining regional construction risk analysis information; constructing a carrier risk assessment model, analyzing whether the current construction risk affects ground buildings based on the regional construction risk analysis information and regional construction monitoring information, and obtaining carrier risk assessment information; determining whether construction risk management is necessary based on the regional construction risk analysis information and carrier risk assessment information, and if so, generating a risk management plan for collaboratively managing regional construction safety risks. This method improves the timeliness, comprehensiveness, and decision-making reliability requirements of underground engineering construction management in complex geological environments.

Description

Technical Field

[0001] The present invention relates to the technical field of underground engineering risk management and control, and in particular to a method and system for collaborative management and control of underground engineering geological safety risks throughout the entire life cycle. Background Art

[0002] With the acceleration of urbanization, the scale of underground space development continues to expand, and the geological safety risks faced by projects such as tunnels and underground pipelines are becoming increasingly prominent. Traditional underground engineering risk management mainly relies on empirical judgment and static geological reports. During the construction phase, manual inspections are conducted to record the deformation of the surrounding rock, and a simplified limit equilibrium theory is used to estimate the safety of the support structure. Currently, underground engineering geological risk management faces three major technical bottlenecks: data fragmentation throughout the life cycle, delayed dynamic response, and difficulty in quantifying multi-source risk transmission. Traditional methods rely on static geological survey data to construct a two-dimensional profile model, which makes it difficult to integrate real-time monitoring information during the construction period for dynamic correction, resulting in deficiencies in construction risk identification and response. At the same time, traditional underground engineering risk management mostly focuses on geological changes, and thus only considers the control of underground projects, failing to consider disaster damage to buildings above or near the construction area.

[0003] In order to meet the requirements of timeliness, comprehensiveness and decision-making reliability for risk management throughout the entire life cycle in complex geological environments, it is urgent to build an intelligent management and control framework with multi-system collaboration. Therefore, a method and system for collaborative management of underground engineering geological safety risks throughout the entire life cycle are proposed to solve the above problems. Summary of the Invention

[0004] The present invention overcomes the shortcomings of the existing technology and provides a method and system for collaborative management of underground engineering geological safety risks throughout the entire life cycle. Its important purpose is to improve the timeliness, comprehensiveness and decision-making reliability requirements of underground engineering construction management in complex geological environments.

[0005] To achieve the above objectives, the present invention provides a first aspect of a method for collaboratively managing underground engineering geological safety risks throughout the entire life cycle, comprising:

[0006] Conduct geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model;

[0007] The construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information. Construction risk analysis and cavity prediction are carried out in combination with the regional geological digital model to obtain regional construction risk analysis information.

[0008] Constructing a load-bearing body risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining load-bearing body risk assessment information;

[0009] Based on the regional construction risk analysis information and the carrier risk assessment information, it is determined whether construction risk control is needed. If necessary, a risk control plan is generated to carry out collaborative control of regional construction safety risks.

[0010] In this solution, the geological monitoring and surface topography scanning of the target area are carried out to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model, specifically including:

[0011] Surface scanning equipment is used to scan the target area at multiple angles according to a preset scanning plan to obtain surface terrain point cloud data. An iterative closest point algorithm is introduced to align the regional point cloud scan data of the target area to a unified engineering coordinate system and eliminate interference points to construct an initial digital surface model.

[0012] Conduct geological monitoring of the target area to obtain regional geological monitoring information, and perform data preprocessing on the obtained regional geological monitoring information to eliminate abnormal data and supplement missing values;

[0013] Perform feature extraction on the pre-processed regional geological monitoring information, extract the regional geological features of the target area and convert them into a vector point set associated with geographic coordinates, and use the Kriging interpolation algorithm to perform spatial distribution calculation;

[0014] Taking the sensor coordinates as control points, the spatial autocorrelation of lithologic parameters is analyzed based on the semivariogram, and a physical property distribution matrix of each cubic meter of rock mass is established to obtain a rock mass property matrix. The rock mass property matrix is ​​then linearly weighted fused with the initial digital surface model in the depth direction to generate a three-dimensional geological grid model.

[0015] In this solution, the geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, regional geological structure identification and geological risk identification, and construction of a regional geological digital model also include:

[0016] Obtaining a three-dimensional geological grid model, inputting the three-dimensional geological grid model into a pre-trained three-dimensional convolutional neural network for structure recognition, traversing the entire three-dimensional geological grid model based on a preset sliding window through the first convolutional layer, performing feature scanning on local rock mass voxel units, and obtaining local rock mass feature vectors;

[0017] The local rock mass feature vector is input into the second layer for compression through the maximum pooling operation, and the maximum eigenvalue is taken for output. Finally, the high-dimensional feature is mapped into a lithology boundary probability map using a fully connected layer;

[0018] Obtaining the velocity difference and RQD value of each voxel unit through the lithologic boundary probability map, and identifying the adjacent voxel units as candidate points of geological discontinuity when the velocity difference or RQD value of adjacent voxel units is greater than a preset threshold;

[0019] For each candidate point on the geological discontinuity, the wave velocity and rock quality index values ​​within the preset voxel range around the point are extracted, and the parameter change rate is calculated in the three dimensions of the X, Y, and Z of the spatial rectangular coordinate system to generate a parameter gradient vector representing the direction of the geological parameter mutation;

[0020] The marching cube algorithm is introduced to connect candidate points of geological discontinuities to generate a fault surface, and a spatial triangulated grid is generated by combining the parameter gradient vector. The engineering fault distance value of each section is calculated using the spatial triangulated grid, and sections with engineering fault distance values ​​less than a preset threshold are eliminated to construct a fault grid model.

[0021] Based on the fault grid model, taking the endpoints where the fault lines intersect as seed source points, a three-dimensional spatial extension calculation is performed along the dominant direction of the rock mass joint surface, and a risk score for each voxel unit is generated by traversing each voxel and extracting risk assessment parameters for weighted calculation, wherein the risk assessment parameters include wave velocity anomaly, RQD attenuation rate, crack density, and seepage convergence intensity;

[0022] The geological risk level is identified based on the calculated risk score, and the final regional geological digital model is output after the geological risk identification is completed.

[0023] In this solution, the construction area is monitored in real time by a set heterogeneous sensor network to obtain regional construction monitoring information. Construction risk analysis and cavity prediction are performed in combination with the regional geological digital model to obtain regional construction risk analysis information, specifically including:

[0024] Based on a preset construction plan, the target area is constructed and the construction data of the construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information, and data preprocessing is performed on the regional construction monitoring information;

[0025] Obtaining a regional geological digital model, inputting preprocessed regional construction monitoring information into the regional geological digital model, and using a spatial hashing algorithm to match the sensor's three-dimensional coordinates with each risk unit in the regional geological model to generate a sequence of construction disturbance vectors with geological attribute labels;

[0026] Performing coupled simulation analysis of geological responses based on the construction disturbance vector sequence, constructing an energy transfer model based on a discrete element-finite element hybrid calculation framework and inputting the construction disturbance vector sequence, wherein the energy transfer model includes a discrete element module and a finite element module;

[0027] In the discrete element module, the rock mass risk unit is used as a basic particle group to simulate the propagation and attenuation process of drilling impact according to the Hertz contact theory, where the energy attenuation rate is dynamically controlled by the rock mass damping coefficient.

[0028] In the finite element model, the interface between the support structure and the surrounding rock is discretized into shell-body coupling elements. The contact stress redistribution under the action of the perturbation load is dynamically solved based on the Lagrangian algorithm. When the peak shear stress of the coupling interface exceeds the rock mass shear strength stored in the geological model, the corresponding interface is calibrated as an abnormal interface.

[0029] Obtaining geological response coupling simulation results, combining them with regional construction monitoring information to input a gradient boosting decision tree model for cavity prediction, analyzing the probability of cavities occurring in the current construction area, and outputting regional construction cavity prediction information;

[0030] Based on the results of geological response coupling simulation, a stress redistribution cloud map is generated, and the area where the disturbance stress increase exceeds the preset value is marked as a risk area. Based on the construction void prediction information of the said area, the risk level is calibrated according to the void probability to obtain regional construction risk analysis information.

[0031] In this solution, the carrier risk assessment model is constructed to analyze whether the current construction risk affects the ground buildings through the regional construction risk analysis information and regional construction monitoring information, and obtain the carrier risk assessment information, specifically including:

[0032] Introducing a data network, obtaining historical construction risk instances through the big data network, extracting features from the historical construction risk instances, classifying the extracted historical construction risk instance features according to construction risk categories, generating several clusters, and constructing a historical construction risk feature set;

[0033] A load-bearing risk assessment model is built based on a Bayesian inference network. During the model training process, a node conditional probability table is introduced to optimize the load-bearing risk assessment model. A training dataset is constructed using a historical construction risk feature set, and a variational Bayesian inference algorithm is used to train the node conditional probability table.

[0034] Establishing a validation data set based on the historical construction risk feature set to validate the load-bearing body risk assessment model; when the model output results meet expectations, retaining the model parameters and outputting the trained load-bearing body risk assessment model;

[0035] Obtain regional construction risk analysis information and regional construction monitoring information, and input them into the trained load-bearing body risk assessment model to analyze the risk probability of regional ground load-bearing bodies when construction risks occur;

[0036] Using the regional construction risk analysis information and regional construction monitoring information input into the model, a current construction risk characteristic map is generated, and the spatial relationship between construction risk and ground bearing bodies is calculated, and topological association is performed with the current construction risk characteristic map;

[0037] The risk nodes are initialized through the current construction risk characteristic graph after topological association, and the node transfer probability distribution is obtained from the node conditional probability table. The node transfer probability distribution is used to infer risk conduction and obtain the carrier risk assessment information.

[0038] In this solution, the need for construction risk control is determined based on regional construction risk analysis information and carrier risk assessment information. If necessary, a risk control plan is generated to conduct collaborative control of regional construction safety risks, specifically including:

[0039] Through historical data retrieval, several historical construction risk control instances are obtained, and features are extracted for each historical construction risk control instance to obtain the historical regional construction risk type, historical regional geological risk characteristics, historical construction risk control scheme characteristics, and historical risk control timeliness characteristics of each historical construction risk control instance;

[0040] Generate entity triples based on historical regional construction risk types, historical regional geological risk characteristics, and historical construction risk control plan characteristics, and construct a risk control knowledge graph based on the entity triples;

[0041] MetaPath random walks are used to represent and learn the risk management knowledge graph and obtain entity nodes corresponding to historical regional geological risk characteristics. Metapaths are generated based on the connections between entity nodes, and the historical risk management timeliness characteristics are used as subsidiary features of the metapaths to form a heterogeneous information network.

[0042] Deep learning and training the heterogeneous information network through a graph neural network to obtain risk assessment information of the regional construction risk analysis information carrier, and input it into the trained graph neural network to generate a target node;

[0043] In the graph neural network, third-order neighborhood sampling is performed on the target node to obtain neighbor nodes and secondary neighbor nodes, the attention coefficient between the neighbor nodes and the secondary neighbor nodes is obtained through the attention mechanism, and the feature vectors corresponding to the neighbor nodes are weighted updated to obtain the updated neighbor nodes;

[0044] The updated neighbor nodes and the target node are vector-concatenated through a shared attention parameter mechanism, and the target's feature vector is updated using a preset activation function. The candidate risk control solutions for the current construction scenario are output based on the updated target node's feature vector.

[0045] Extract the risk control timeliness characteristics corresponding to each candidate risk control plan, compare them with the expected timeliness, screen the candidate risk control plans that meet the expected timeliness and sort them, and provide construction risk control assistance based on the sorting results.

[0046] A second aspect of the present invention provides a full life cycle underground engineering geological safety risk collaborative management and control system, the system comprising: a memory and a processor, the memory containing a full life cycle underground engineering geological safety risk collaborative management and control method program, the full life cycle underground engineering geological safety risk collaborative management and control method program when executed by the processor to implement the following steps:

[0047] Conduct geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model;

[0048] The construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information. Construction risk analysis and cavity prediction are carried out in combination with the regional geological digital model to obtain regional construction risk analysis information.

[0049] Constructing a load-bearing body risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining load-bearing body risk assessment information;

[0050] Based on the regional construction risk analysis information and the carrier risk assessment information, it is determined whether construction risk control is needed. If necessary, a risk control plan is generated to carry out collaborative control of regional construction safety risks.

[0051] The present invention discloses a method and system for collaboratively managing underground engineering geological safety risks throughout the entire life cycle, including: identifying the regional geological structure and geological risk markers in the target construction area and constructing a regional geological digital model; obtaining regional construction monitoring information, combining it with the regional geological digital model to perform construction risk analysis and cavity prediction, and obtaining regional construction risk analysis information; constructing a carrier risk assessment model, analyzing whether the current construction risk affects ground buildings based on the regional construction risk analysis information and regional construction monitoring information, and obtaining carrier risk assessment information; determining whether construction risk management is necessary based on the regional construction risk analysis information and carrier risk assessment information, and if so, generating a risk management plan for collaboratively managing regional construction safety risks. This method improves the timeliness, comprehensiveness, and decision-making reliability requirements of underground engineering construction management in complex geological environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or exemplary descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to these drawings without paying any creative work.

[0053] Figure 1 A flowchart of a method for collaborative management and control of underground engineering geological safety risks throughout the entire life cycle provided by one embodiment of the present invention;

[0054] Figure 2 A flowchart of a method for collaborative management and control of underground engineering construction safety risks provided by one embodiment of the present invention;

[0055] Figure 3 A block diagram of a collaborative management and control system for underground engineering geological safety risks throughout the entire life cycle provided by one embodiment of the present invention;

[0056] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0057] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0058] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0059] Figure 1 A flowchart of a method for collaborative management and control of underground engineering geological safety risks throughout the entire life cycle provided by one embodiment of the present invention;

[0060] like Figure 1 As shown, the present invention provides a flow chart of a method for collaborative management and control of underground engineering geological safety risks throughout the entire life cycle, including:

[0061] S102, conducting geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model;

[0062] S104, monitoring the construction area in real time through a set heterogeneous sensor network to obtain regional construction monitoring information, and performing construction risk analysis and cavity prediction in combination with a regional geological digital model to obtain regional construction risk analysis information;

[0063] S106, constructing a load-bearing body risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining load-bearing body risk assessment information;

[0064] S108: Based on the regional construction risk analysis information and the carrier risk assessment information, determine whether construction risk control is needed. If necessary, generate a risk control plan to carry out collaborative control of regional construction safety risks.

[0065] Furthermore, in a preferred embodiment of the present invention, the geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, perform regional geological structure identification and geological risk identification, and construct a regional geological digital model specifically includes:

[0066] Surface scanning equipment is used to scan the target area at multiple angles according to a preset scanning plan to obtain surface terrain point cloud data. An iterative closest point algorithm is introduced to align the regional point cloud scan data of the target area to a unified engineering coordinate system and eliminate interference points to construct an initial digital surface model.

[0067] Conduct geological monitoring of the target area to obtain regional geological monitoring information, and perform data preprocessing on the obtained regional geological monitoring information to eliminate abnormal data and supplement missing values;

[0068] Perform feature extraction on the pre-processed regional geological monitoring information, extract the regional geological features of the target area and convert them into a vector point set associated with geographic coordinates, and use the Kriging interpolation algorithm to perform spatial distribution calculation;

[0069] Taking the sensor coordinates as control points, the spatial autocorrelation of lithologic parameters is analyzed based on the semivariogram, and a physical property distribution matrix of each cubic meter of rock mass is established to obtain a rock mass property matrix. The rock mass property matrix is ​​then linearly weighted fused with the initial digital surface model in the depth direction to generate a three-dimensional geological grid model.

[0070] It should be noted that, based on a pre-set scanning scheme, a multi-rotor drone equipped with a lidar system performed multi-angle surface scans of the target area, acquiring a raw topographic point cloud dataset with centimeter-level accuracy. The scanning process adhered to a crisscross flight strip design, with single-point positioning accuracy maintained within ±2 cm and a point cloud density of no less than 200 points per square meter. The multi-strip scan data was automatically registered using the Iterative Closest Point (ICP) algorithm: Using the coordinates of the engineering control network base station as the conversion target, the point clouds of each flight strip were uniformly converted to the engineering coordinate system by minimizing the root mean square error (RMSE <5 cm) of the Euclidean distance between corresponding point clouds. The registered point clouds were then filtered using an improved moving surface filtering algorithm to remove non-topographic point interference, such as vegetation cover and temporary facilities. A high-fidelity digital surface model was generated using Poisson surface reconstruction technology, clearly depicting the surface undulations and the outlines of man-made structures. A multi-type geological monitoring sensor network, including in-hole strain gauges (buried at depths of 0-50m), interwell seismic velocimeter arrays, and groundwater pressure sensors, is deployed simultaneously to continuously collect parameters such as rock deformation rate, elastic wave velocity, and seepage pressure. The raw monitoring data undergoes two stages of preprocessing: first, a boxplot statistical analysis is used to identify and remove anomalous jump points caused by temperature drift and electrical interference. Subsequently, a spatiotemporal kriging interpolation algorithm is used to supplement signal interruptions or missing values, ensuring data temporal and spatial continuity. Wavelet packet decomposition is used to extract seven-dimensional geological characteristic parameters from the processed structured data, including triaxial strain energy density, wave velocity gradient modulus, and seepage loss instability coefficient. Each characteristic parameter is bound to its spatial geographic coordinates (longitude, latitude, and elevation), forming a vector feature point set with location labels.

[0071] The vector point set was then input into a spatial interpolation engine, and a kriging spatial distribution calculation was performed using the three-dimensional coordinates of the monitoring points as control points. The spatial autocorrelation of lithologic parameters was quantified using the semivariogram. The spatial variation of velocity and strain parameters was analyzed, and standard model parameters were determined: a nugget effect value of 0.12, a range of 15 meters, and a sill value of 1.8. Based on this model, interpolation calculations were performed on a regular 0.5m×0.5m×0.5m grid in three-dimensional space to generate a physical property distribution matrix per cubic meter of rock mass (including mechanical indices such as dynamic elastic modulus, Poisson's ratio, permeability, and joint density). The spatial resolution and error distribution closely matched the borehole calibration results (vertical error <5%, horizontal error <8%). Ultimately, a three-dimensional integration of surface topography and rock mass properties is achieved through depth-wise linear weighted fusion. For shallow depths (0-5 m), the digital surface model weight coefficient is set to 0.8 (emphasizing topographic constraints), while the rock mass attribute matrix weight is 0.2. For intermediate depths (5-20 m), a balanced weighting of 0.5:0.5 is applied. For depths greater than 20 m, rock mass attributes dominate (weighting 0.9). The fusion process utilizes an octree spatial index structure to perform multi-layer Level of Dimension (LOD) processing, generating a three-dimensional geological raster model. This model is stored as a tensor, with each voxel containing 12-dimensional attribute fields, including terrain elevation, lithology code, and mechanical parameters. This provides a fully parameterized digital foundation for subsequent intelligent identification of geological risks.

[0072] Furthermore, in a preferred embodiment of the present invention, the geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, perform regional geological structure identification and geological risk identification, and construct a regional geological digital model further includes:

[0073] Obtaining a three-dimensional geological grid model, inputting the three-dimensional geological grid model into a pre-trained three-dimensional convolutional neural network for structure recognition, traversing the entire three-dimensional geological grid model based on a preset sliding window through the first convolutional layer, performing feature scanning on local rock mass voxel units, and obtaining local rock mass feature vectors;

[0074] The local rock mass feature vector is input into the second layer for compression through the maximum pooling operation, and the maximum eigenvalue is taken for output. Finally, the high-dimensional feature is mapped into a lithology boundary probability map using a fully connected layer;

[0075] Obtaining the velocity difference and RQD value of each voxel unit through the lithologic boundary probability map, and identifying the adjacent voxel units as candidate points of geological discontinuity when the velocity difference or RQD value of adjacent voxel units is greater than a preset threshold;

[0076] For each candidate point on the geological discontinuity, the wave velocity and rock quality index values ​​within the preset voxel range around the point are extracted, and the parameter change rate is calculated in the three dimensions of the X, Y, and Z of the spatial rectangular coordinate system to generate a parameter gradient vector representing the direction of the geological parameter mutation;

[0077] The marching cube algorithm is introduced to connect candidate points of geological discontinuities to generate a fault surface, and a spatial triangulated grid is generated by combining the parameter gradient vector. The engineering fault distance value of each section is calculated using the spatial triangulated grid, and sections with engineering fault distance values ​​less than a preset threshold are eliminated to construct a fault grid model.

[0078] Based on the fault grid model, taking the endpoints where the fault lines intersect as seed source points, a three-dimensional spatial extension calculation is performed along the dominant direction of the rock mass joint surface, and a risk score for each voxel unit is generated by traversing each voxel and extracting risk assessment parameters for weighted calculation, wherein the risk assessment parameters include wave velocity anomaly, RQD attenuation rate, crack density, and seepage convergence intensity;

[0079] The geological risk level is identified based on the calculated risk score, and the final regional geological digital model is output after the geological risk identification is completed.

[0080] It should be noted that a 3D geological grid model was used as the input data source for a pre-trained 3D convolutional neural network (3D-CNN). The first convolutional layer uses a 3×3×3 cubic convolution kernel, which traverses the entire model space in a sliding window format. As the kernel moves, it performs a deep feature scan of local rock mass voxels, focusing on extracting nine-dimensional feature parameters such as velocity gradient, joint density, and stress concentration. The output is a local rock mass feature vector matrix consisting of 128-channel feature maps. This feature matrix is ​​input to the second processing layer, where a 2×2×2 maximum pooling operation is performed to reduce the data dimension. The most significant geological anomaly signal within each cubic region is extracted (retaining the maximum value), reducing the data size to 1 / 8 of its original size. This pooled feature map is then input to a fully connected layer, where a hyperbolic tangent activation function is used to map the high-dimensional features into a lithologic boundary probability map. This probability map quantifies the probability of a lithologic abrupt change at each spatial location (ranging from 0 to 1), thereby clearly identifying geological interfaces such as sandstone-mudstone boundaries and fault fracture zones.

[0081] Discontinuity detection is performed based on lithologic boundary probability maps. The absolute value of the velocity difference between adjacent voxels (measured across cell boundaries) and the amplitude of the rock quality indicator (RQD) jump are calculated. When a velocity difference >500 m / s or an RQD decay rate >30% within three consecutive voxels is detected, these voxels are identified as candidate geological discontinuity points, and their 3D coordinates and abnormal parameter values ​​are recorded. A parameter gradient field analysis is performed for each candidate point. The velocity and RQD values ​​within a 3×3×3 voxel range surrounding the candidate point are extracted, and the parameter change rate is calculated along the X, Y, and Z axes of the rectangular coordinate system. The X-direction gradient is calculated as the velocity difference between the left and right adjacent voxels divided by the voxel spacing. The Y-direction calculates the difference between the preceding and following voxels. The Z-direction analyzes vertical changes, resulting in a composite 3D gradient vector. This vector points in the direction of the most significant geological parameter degradation, and its modulus indicates the severity of the change. Subsequently, the marching cube algorithm is used to construct the fault space surface. A cubic computational domain is generated with the candidate discontinuity points as vertices, and isosurface intersection points are located along the cube edges using linear interpolation. Adjacent intersection points are connected based on the spatial continuity of the gradient vector G, and triangular facets are generated according to the right-hand rule. The engineering fault throw value is automatically calculated for each triangular facet: matching pairs of lithologic features are identified on the fault plane, and vertical fault throws are calculated based on the unit normal vector N. Horizontal fault throws are calculated using the strike vector T, resulting in a three-dimensional composite fault throw value. Microfracture surfaces with fault throws less than a preset threshold are automatically eliminated, retaining the main fault zone to form a fault grid model.

[0082] Risk zoning is initiated based on the fault grid model: Using the intersection of fault lines as the seed source, a three-dimensional region growing algorithm is applied along the dominant direction of the joint plane (within the strike range of ±15° and the dip range of ±10°). The growing process traverses the covered voxel units and simultaneously calculates four key parameters: velocity anomaly (the standard deviation relative to the regional background value), RQD attenuation (the slope of the linear degradation along the depth direction), fracture density (the number of microcracks identified in borehole CT images per cubic meter), and seepage convergence intensity (the divergence of the groundwater flow vector field). The risk assessment formula is: risk score = 0.4 × velocity anomaly + 0.3 × RQD attenuation + 0.2 × fracture density + 0.1 × seepage convergence intensity. Voxels with a risk score greater than 0.7 are colored red (high-risk areas), those between 0.5 and 0.7 are colored yellow (warning areas), and the remaining areas retain the base color. The calculated risk scores are used to assign geological risk levels. After completing the geological risk identification, the final regional geological digital model is output.

[0083] Furthermore, in a preferred embodiment of the present invention, the construction area is monitored in real time by a set heterogeneous sensor network to obtain regional construction monitoring information, and construction risk analysis and cavity prediction are performed in combination with a regional geological digital model to obtain regional construction risk analysis information, specifically including:

[0084] Based on a preset construction plan, the target area is constructed and the construction data of the construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information, and data preprocessing is performed on the regional construction monitoring information;

[0085] Obtaining a regional geological digital model, inputting preprocessed regional construction monitoring information into the regional geological digital model, and using a spatial hashing algorithm to match the sensor's three-dimensional coordinates with each risk unit in the regional geological model to generate a sequence of construction disturbance vectors with geological attribute labels;

[0086] Performing coupled simulation analysis of geological responses based on the construction disturbance vector sequence, constructing an energy transfer model based on a discrete element-finite element hybrid calculation framework and inputting the construction disturbance vector sequence, wherein the energy transfer model includes a discrete element module and a finite element module;

[0087] In the discrete element module, the rock mass risk unit is used as a basic particle group to simulate the propagation and attenuation process of drilling impact according to the Hertz contact theory, where the energy attenuation rate is dynamically controlled by the rock mass damping coefficient.

[0088] In the finite element model, the interface between the support structure and the surrounding rock is discretized into shell-body coupling elements. The contact stress redistribution under the action of the perturbation load is dynamically solved based on the Lagrangian algorithm. When the peak shear stress of the coupling interface exceeds the rock mass shear strength stored in the geological model, the corresponding interface is calibrated as an abnormal interface.

[0089] Obtaining geological response coupling simulation results, combining them with regional construction monitoring information to input a gradient boosting decision tree model for cavity prediction, analyzing the probability of cavities occurring in the current construction area, and outputting regional construction cavity prediction information;

[0090] Based on the results of geological response coupling simulation, a stress redistribution cloud map is generated, and the area where the disturbance stress increase exceeds the preset value is marked as a risk area. Based on the construction void prediction information of the said area, the risk level is calibrated according to the void probability to obtain regional construction risk analysis information.

[0091] It should be noted that a heterogeneous sensor network is used to monitor construction data in the construction area in real time, generating regional construction monitoring information. A temperature drift compensation algorithm is used to eliminate environmental interference, and wavelet threshold noise reduction technology is used to filter out vibration noise from construction machinery. The final output is preprocessed regional construction monitoring information. This preprocessed monitoring data is then fed into the regional geological digital model. A spatial hash indexing engine precisely matches the sensor physical coordinates with risk cells in the geological model. Using a three-dimensional hash bucketing scheme (bucket size 0.5 m³), ​​risk voxels within a 1.2 m radius of the target cell are retrieved at a speed of 0.1 ms. This generates a construction disturbance vector sequence V = {timestamp, spatial coordinates, vibration acceleration, support stress, pore water pressure, and risk level code}, labeled with geological attributes.

[0092] An energy transfer model is constructed based on a discrete element (DEM)-finite element (FEM) hybrid computational framework. The construction disturbance vector sequence is input, and energy propagation modeling is performed using a hybrid DEM-FEM computational architecture. Within the DEM computational domain, the risk units of the rock mass are treated as dynamic particle swarms, and the propagation attenuation process is simulated based on contact mechanics principles, where the energy attenuation rate is dynamically controlled by the rock mass damping coefficient. Within the FEM computational domain, the contact surface between the support structure and the surrounding rock is discretized into coupled units. The Lagrangian algorithm dynamically solves for the contact stress redistribution under the action of the disturbance load. When the peak shear stress at the interface exceeds the rock mass shear strength threshold stored in the geological model, the interface is automatically marked as a slip risk zone and warning coordinates are output. Furthermore, the coupled geological response simulation results are obtained and combined with regional construction monitoring information to input into a gradient boosting decision tree model for cavity prediction. Using a gradient boosting decision tree algorithm, six-dimensional characteristic parameters were extracted, including key indicators such as relative disturbance stress increase, cumulative energy of microseismic events, dynamic attenuation rate of rock mass integrity, and abnormal seepage direction deflection. A fault space constraint factor was specifically introduced to strengthen geological correlations, resulting in a gridded void probability distribution map that clearly displays the potential location and risk level of future cavity development. Based on the results of the coupled geological response simulation, a stress redistribution cloud map was generated. Areas exceeding a preset disturbance stress increase were designated as risk areas. Based on the construction void prediction information in these areas, the risk level was calibrated according to the void probability, resulting in regional construction risk analysis information.

[0093] The rock mass damping coefficient is calibrated in real time by inverting the cross-hole seismic velocity field, and the calculation formula is:

[0094] ,

[0095] in, is the rock mass damping coefficient, is the longitudinal wave velocity, is the stress difference between adjacent rock layers.

[0096] Furthermore, in a preferred embodiment of the present invention, the carrier risk assessment model is constructed to analyze whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, thereby obtaining the carrier risk assessment information, which specifically includes:

[0097] Introducing a data network, obtaining historical construction risk instances through the big data network, extracting features from the historical construction risk instances, classifying the extracted historical construction risk instance features according to construction risk categories, generating several clusters, and constructing a historical construction risk feature set;

[0098] A load-bearing risk assessment model is built based on a Bayesian inference network. During the model training process, a node conditional probability table is introduced to optimize the load-bearing risk assessment model. A training dataset is constructed using a historical construction risk feature set, and a variational Bayesian inference algorithm is used to train the node conditional probability table.

[0099] Establishing a validation data set based on the historical construction risk feature set to validate the load-bearing body risk assessment model; when the model output results meet expectations, retaining the model parameters and outputting the trained load-bearing body risk assessment model;

[0100] Obtain regional construction risk analysis information and regional construction monitoring information, and input them into the trained load-bearing body risk assessment model to analyze the risk probability of regional ground load-bearing bodies when construction risks occur;

[0101] Using the regional construction risk analysis information and regional construction monitoring information input into the model, a current construction risk characteristic map is generated, and the spatial relationship between construction risk and ground bearing bodies is calculated, and topological association is performed with the current construction risk characteristic map;

[0102] The risk nodes are initialized through the current construction risk characteristic graph after topological association, and the node transfer probability distribution is obtained from the node conditional probability table. The node transfer probability distribution is used to infer risk conduction and obtain the carrier risk assessment information.

[0103] It should be noted that a global historical construction risk case database was constructed based on an engineering big data platform. Distributed crawler technology was used to collect several underground engineering risk event cases. Multidimensional features were extracted from the original cases, focusing on core elements such as construction disturbance intensity, geological anomaly type, support structure response, and disaster evolution path, forming a structured dataset of multidimensional feature vectors. A clustering algorithm was used to classify the cases into six main clusters based on risk mechanisms (e.g., fault activation seepage, large soft rock deformation, and joint slip), establishing a labeled knowledge base of historical construction risk characteristics. A network model framework for bearing body risk assessment was constructed based on Bayesian theory. The network nodes comprise a four-layer topology structure consisting of construction disturbance factors, rock mass degradation indicators, support state parameters, and ground structure vulnerability. A three-layer optimization mechanism was introduced: first, node association rules were initialized based on geotechnical principles (e.g., a positive correlation between drilling and blasting vibration and fault activation, and a hysteresis effect between support stress and surface settlement). Second, prior conditional probabilities based on industry expert experience were embedded. Finally, a variational Bayesian inference algorithm was used for deep learning. By performing thousands of iterations of probabilistic reasoning on the risk evolution path, the parameters of the conditional probability table between nodes are gradually revised. Model validation is implemented through independent case set stress testing: 500 sets of risk scenario characteristics not included in the training are input. When the error rate between the model's output risk probability and the measured disaster results remains stable within a ±7% range, the model is deemed to meet engineering application standards and the trained model is output.

[0104] Subsequently, the model uses real-time input of regional construction risk analysis information and high-precision monitoring data. The model first integrates this multidimensional information to generate a spatiotemporal linkage feature map. This map, based on a geographic information system (GIS), maps dynamic parameters such as the current construction disturbance intensity, fault activation status, and shallow rock mass fracture density. The model then calculates the three-dimensional spatial relationship between construction risks (such as void areas) and ground hazard-bearing structures, and establishes a risk transmission topological chain based on joint surface attitude data from the geological model. The topological feature map initializes risk node states, and transfer rules are invoked from a trained conditional probability table to implement risk transmission deduction. This model simulates the hazard chain from construction vibration to fault movement to shallow fracture expansion to building foundation settlement. The model then outputs three indicators: the probability of ground building tilt, the risk level of road cracking, and the deformation threshold of underground pipelines. The resulting risk assessment report integrates spatial location labels, a hazard path map, and probability predictions. A red-orange-yellow risk zoning map visually identifies sensitive ground targets affected by construction.

[0105] Figure 2 A block diagram of a collaborative management and control system for underground engineering geological safety risks throughout the entire life cycle provided by one embodiment of the present invention;

[0106] like Figure 2 As shown, the present invention provides a block diagram of a collaborative management and control system for underground engineering geological safety risks throughout the entire life cycle, including:

[0107] S202, obtaining several historical construction risk control instances through historical data retrieval, extracting features of each historical construction risk control instance, and obtaining the historical regional construction risk type, historical regional geological risk characteristics, historical construction risk control plan characteristics, and historical risk control timeliness characteristics of each historical construction risk control instance;

[0108] S204, generating entity triples based on the historical regional construction risk type, historical regional geological risk characteristics, and historical construction risk control plan characteristics, and constructing a risk control knowledge graph based on the entity triples;

[0109] S206: Using MetaPath random walks to perform representation learning on the risk control knowledge graph and obtain entity nodes corresponding to historical regional geological risk characteristics, generating meta-paths based on the connections between the entity nodes, and using the historical risk control timeliness characteristics as subsidiary features of the meta-paths to form a heterogeneous information network;

[0110] S208, performing deep learning and training on the heterogeneous information network through a graph neural network to obtain risk assessment information of the regional construction risk analysis information carrier, and inputting the information into the trained graph neural network to generate a target node;

[0111] S210, in the graph neural network, performing third-order neighborhood sampling on the target node to obtain neighbor nodes and secondary neighbor nodes, obtaining attention coefficients between the neighbor nodes and the secondary neighbor nodes through an attention mechanism, performing weighted update on feature vectors corresponding to the neighbor nodes, and obtaining updated neighbor nodes;

[0112] S212, concatenating the updated neighbor nodes and the target node through a shared attention parameter mechanism, updating the target's feature vector using a preset activation function, and outputting a candidate risk control solution for the current construction scenario based on the updated target node feature vector;

[0113] S214: extract the risk control timeliness characteristics corresponding to each candidate risk control scheme, compare them with the expected timeliness, select the candidate risk control schemes that meet the expected timeliness, and sort them, and provide construction risk control assistance based on the sorting results.

[0114] It should be noted that historical risk management examples of underground projects were retrieved from the Engineering Data Center for structured feature extraction. Each case was broken down into a four-dimensional feature vector: historical regional construction risk type (such as fault water inrush, large soft rock deformation, and regional building settlement), historical regional geological risk characteristics (including quantitative indicators such as fault throw / inclination / RQD decay rate), historical construction risk management plan characteristics, and historical risk management timeliness characteristics (such as response delay hours and duration of action). Using entity-relationship modeling techniques, the first three types of features were constructed as entity triplets consisting of risk type, geological characteristics, and management plan, constructing a risk management knowledge graph encompassing all three types of entities. A MetaPath random walk strategy was used for graph representation learning: starting from a geological risk characteristic node, 1000 rounds of walk sampling were performed along the path from geological characteristics to risk events to management plans. During this process, historical risk management timeliness characteristics were converted into attributes of the metapath, ultimately forming a heterogeneous information network with spatiotemporal constraints. This heterogeneous information network was then fed into a graph neural network for deep learning and training, resulting in a satisfactory graph neural network.

[0115] Subsequently, risk assessment information from the regional construction risk analysis information carrier is obtained and input into the trained graph neural network to generate a target node. Third-order neighborhood sampling is used to obtain directly associated neighbor nodes (e.g., cases with the same fault characteristics) and secondary neighbor nodes (cases with similar rock mass structures) associated with the target node. A multi-head attention mechanism is used to calculate the node association strength, or attention coefficient. After weighted aggregation and updating neighbor node features, these features are concatenated with the target node feature vector using a parameter sharing mechanism and activated using a ReLU activation function to generate a new target node feature that incorporates global knowledge. This feature vector is mapped through a fully connected layer, outputting candidate risk management solutions. The risk management timeliness characteristics corresponding to each candidate risk management solution are extracted and compared with the expected timeliness. Candidate risk management solutions that meet the expected timeliness are selected and ranked. Construction risk management assistance is provided based on the ranking results. This approach improves the timeliness, comprehensiveness, and decision-making reliability required for underground construction management in complex geological environments.

[0116] Figure 3 A full-life cycle underground engineering geological safety risk collaborative management and control system 3 is provided in one embodiment of the present invention. The system includes: a memory 31 and a processor 32. The memory 31 contains a full-life cycle underground engineering geological safety risk collaborative management and control method program. When the full-life cycle underground engineering geological safety risk collaborative management and control method program is executed by the processor 32, the following steps are implemented:

[0117] Conduct geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model;

[0118] The construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information. Construction risk analysis and cavity prediction are carried out in combination with the regional geological digital model to obtain regional construction risk analysis information.

[0119] Constructing a load-bearing body risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining load-bearing body risk assessment information;

[0120] Based on the regional construction risk analysis information and the carrier risk assessment information, it is determined whether construction risk control is needed. If necessary, a risk control plan is generated to carry out collaborative control of regional construction safety risks.

[0121] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0122] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0123] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0124] Those skilled in the art will appreciate that all or part of the steps of the above-mentioned method embodiments may be implemented by hardware associated with program instructions, and the aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the program executes the steps of the above-mentioned method embodiments. The aforementioned storage medium includes various media that can store program codes, such as mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0125] Alternatively, if the integrated units described above are implemented as software modules and sold or used as standalone products, they can also be stored on a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product, stored on a storage medium, includes instructions for enabling a computer device (such as a personal computer, server, or network device) to execute all or part of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as removable storage devices, ROM, RAM, magnetic disks, or optical disks.

[0126] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A collaborative management and control method for underground engineering geological safety risks throughout the entire life cycle, characterized in that: include: Conduct geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, identify regional geological structures and geological risk markers, and construct a regional geological digital model; The construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information. Construction risk analysis and cavity prediction are carried out in combination with the regional geological digital model to obtain regional construction risk analysis information. Constructing a load-bearing body risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining load-bearing body risk assessment information; Determine whether construction risk control is necessary based on regional construction risk analysis information and carrier risk assessment information. If necessary, generate a risk control plan to collaboratively control regional construction safety risks. The method involves real-time monitoring of the construction area through a set heterogeneous sensor network to obtain regional construction monitoring information, and combining the regional geological digital model to perform construction risk analysis and cavity prediction to obtain regional construction risk analysis information, specifically including: Based on a preset construction plan, the target area is constructed and the construction data of the construction area is monitored in real time through a set heterogeneous sensor network to obtain regional construction monitoring information, and data preprocessing is performed on the regional construction monitoring information; Obtaining a regional geological digital model, inputting preprocessed regional construction monitoring information into the regional geological digital model, and using a spatial hashing algorithm to match the sensor's three-dimensional coordinates with each risk unit in the regional geological digital model to generate a sequence of construction disturbance vectors with geological attribute labels; Performing coupled simulation analysis of geological responses based on the construction disturbance vector sequence, constructing an energy transfer model based on a discrete element-finite element hybrid calculation framework and inputting the construction disturbance vector sequence, wherein the energy transfer model includes a discrete element module and a finite element module; In the discrete element module, the rock mass risk unit is used as a basic particle group to simulate the propagation and attenuation process of drilling impact according to the Hertz contact theory, where the energy attenuation rate is dynamically controlled by the rock mass damping coefficient. In the finite element model, the interface between the support structure and the surrounding rock is discretized into shell-body coupling elements. The contact stress redistribution under the action of the perturbation load is dynamically solved based on the Lagrangian algorithm. When the peak shear stress of the coupling interface exceeds the rock mass shear strength stored in the regional geological digital model, the corresponding interface is calibrated as an abnormal interface. Obtaining geological response coupling simulation results, combining them with regional construction monitoring information to input a gradient boosting decision tree model for cavity prediction, analyzing the probability of cavities occurring in the current construction area, and outputting regional construction cavity prediction information; Based on the results of geological response coupling simulation, a stress redistribution cloud map is generated, and the area where the disturbance stress increase exceeds the preset value is marked as a risk area. Based on the construction void prediction information of the said area, the risk level is calibrated according to the void probability to obtain regional construction risk analysis information.

2. A collaborative management and control method for underground engineering geological safety risks throughout the entire life cycle according to claim 1, characterized in that: The geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, perform regional geological structure identification and geological risk identification, and construct a regional geological digital model specifically includes: Surface scanning equipment is used to scan the target area at multiple angles according to a preset scanning plan to obtain surface terrain point cloud data. An iterative closest point algorithm is introduced to align the surface terrain point cloud data of the target area to a unified engineering coordinate system and eliminate interference points to construct an initial digital surface model. Conduct geological monitoring of the target area to obtain regional geological monitoring information, and perform data preprocessing on the obtained regional geological monitoring information to eliminate abnormal data and supplement missing values; Perform feature extraction on the pre-processed regional geological monitoring information, extract the regional geological features of the target area and convert them into a vector point set associated with geographic coordinates, and use the Kriging interpolation algorithm to perform spatial distribution calculation; Taking the sensor coordinates as control points, the spatial autocorrelation of lithologic parameters is analyzed based on the semivariogram, and a physical property distribution matrix of each cubic meter of rock mass is established to obtain a rock mass property matrix. The rock mass property matrix is ​​then linearly weighted fused with the initial digital surface model in the depth direction to generate a three-dimensional geological grid model.

3. A collaborative management and control method for underground engineering geological safety risks throughout the entire life cycle according to claim 1, characterized in that: The geological monitoring and surface topography scanning of the target area to obtain regional geological monitoring information and surface topography point cloud data, regional geological structure identification and geological risk identification, and construction of a regional geological digital model also include: Obtaining a three-dimensional geological grid model, inputting the three-dimensional geological grid model into a pre-trained three-dimensional convolutional neural network for structure recognition, traversing the entire three-dimensional geological grid model based on a preset sliding window through the first convolutional layer, performing feature scanning on local rock mass voxel units, and obtaining local rock mass feature vectors; The local rock mass feature vector is input into the second layer for compression through the maximum pooling operation, and the maximum eigenvalue is taken for output. Finally, the high-dimensional feature is mapped into a lithology boundary probability map using a fully connected layer; Obtaining the velocity difference and RQD value of each voxel unit through the lithologic boundary probability map, and identifying the adjacent voxel units as candidate points of geological discontinuity when the velocity difference or RQD value of adjacent voxel units is greater than a preset threshold; For each candidate point of geological discontinuity, the wave velocity and RQD values ​​within the preset voxel range are extracted, and the parameter change rate is calculated in the three dimensions of X, Y, and Z of the spatial rectangular coordinate system to generate a parameter gradient vector representing the direction of the geological parameter mutation; The marching cube algorithm is introduced to connect candidate points of geological discontinuities to generate fault surfaces, and a spatial triangulated grid is generated by combining parameter gradient vectors. The engineering fault distance value of each fault surface is calculated using the spatial triangulated grid, and fault surfaces with engineering fault distance values ​​less than a preset threshold are eliminated to construct a fault grid model. Based on the fault grid model, taking the endpoints where the fault lines intersect as seed source points, a three-dimensional spatial extension calculation is performed along the dominant direction of the rock mass joint surface, and risk assessment parameters are extracted by traversing each voxel unit and weighted calculation is performed to generate a risk score for each voxel unit, wherein the risk assessment parameters include wave velocity anomaly, RQD attenuation rate, crack density, and seepage convergence intensity; The geological risk level is identified based on the calculated risk score, and the final regional geological digital model is output after the geological risk level identification is completed.

4. A method for collaborative management and control of underground engineering geological safety risks throughout the entire life cycle according to claim 1, characterized in that: The constructing of the carrier risk assessment model, analyzing whether the current construction risk affects the building objects on the ground through the regional construction risk analysis information and the regional construction monitoring information, and obtaining the carrier risk assessment information specifically includes: A big data network is introduced to obtain historical construction risk instances through the big data network, and features of the historical construction risk instances are extracted. The extracted features of the historical construction risk instances are divided according to construction risk categories to generate several clusters and construct a historical construction risk feature set. A load-bearing risk assessment model is built based on a Bayesian inference network. During the training process of the load-bearing risk assessment model, a node conditional probability table is introduced to optimize the load-bearing risk assessment model. A training dataset is constructed using a historical construction risk feature set, and a variational Bayesian inference algorithm is used to train the node conditional probability table. Establishing a validation data set based on the historical construction risk feature set to validate the load-bearing body risk assessment model; when the output result of the load-bearing body risk assessment model meets expectations, retaining the model parameters and outputting the trained load-bearing body risk assessment model; Obtain regional construction risk analysis information and regional construction monitoring information, and input them into the trained load-bearing body risk assessment model to analyze the risk probability of regional ground load-bearing bodies when construction risks occur; Using the regional construction risk analysis information and regional construction monitoring information input into the carrier risk assessment model, a current moment construction risk characteristic map is generated, and the spatial relationship between the construction risk and the ground carrier is calculated, and topologically associated with the current moment construction risk characteristic map; The risk nodes are initialized through the current construction risk characteristic graph after topological association, and the node transfer probability distribution is obtained from the node conditional probability table. The node transfer probability distribution is used to infer risk conduction and obtain the carrier risk assessment information.

5. A collaborative management and control method for underground engineering geological safety risks throughout the entire life cycle according to claim 1, characterized in that: The determination of whether construction risk control is required based on the regional construction risk analysis information and the carrier risk assessment information, and if necessary, generating a risk control plan for coordinated regional construction safety risk control, specifically including: Through historical data retrieval, several historical construction risk control instances are obtained, and features are extracted for each historical construction risk control instance to obtain the historical regional construction risk type, historical regional geological risk characteristics, historical construction risk control scheme characteristics, and historical risk control timeliness characteristics of each historical construction risk control instance; Generate entity triples based on historical regional construction risk types, historical regional geological risk characteristics, and historical construction risk control plan characteristics, and construct a risk control knowledge graph based on the entity triples; MetaPath random walks are used to represent and learn the risk management knowledge graph and obtain entity nodes corresponding to historical regional geological risk characteristics. Metapaths are generated based on the connections between entity nodes, and the historical risk management timeliness characteristics are used as subsidiary features of the metapaths to form a heterogeneous information network. Deep learning and training the heterogeneous information network through a graph neural network to obtain regional construction risk analysis information and carrier risk assessment information, which are input into the trained graph neural network to generate target nodes; In the graph neural network, third-order neighborhood sampling is performed on the target node to obtain neighbor nodes and secondary neighbor nodes, the attention coefficient between the neighbor nodes and the secondary neighbor nodes is obtained through the attention mechanism, and the feature vectors corresponding to the neighbor nodes are weighted updated to obtain the updated neighbor nodes; The updated neighbor nodes and the target node are vector-concatenated through a shared attention parameter mechanism, and the feature vector of the target node is updated using a preset activation function. The candidate risk control plan for the current construction scenario is output based on the updated feature vector of the target node. Extract the risk control timeliness characteristics corresponding to each candidate risk control plan, compare them with the expected timeliness, screen the candidate risk control plans that meet the expected timeliness and sort them, and provide construction risk control assistance based on the sorting results.

6. A full life cycle underground engineering geological safety risk collaborative management and control system, characterized by: The system includes: a memory and a processor, wherein the memory contains a method program for collaboratively controlling underground engineering geological safety risks throughout the entire life cycle, and when the method program for collaboratively controlling underground engineering geological safety risks throughout the entire life cycle is executed by the processor, any one of the steps of the method for collaboratively controlling underground engineering geological safety risks throughout the entire life cycle as described in claims 1-5 is implemented.

Images