A method and system for monitoring deformation of a tunnel based on TBM construction

Abstract

The application discloses a kind of based on TBM construction tunnel deformation monitoring method and system, and the application relates to construction tunnel monitoring technical field, the method includes obtaining the multi-source geological data after TBM construction tunnel construction;Determine the multiple high-sensitive deformation point information of TBM construction tunnel based on the multi-source geological data after TBM construction tunnel construction;Based on the multiple high-sensitive deformation point information of TBM construction tunnel, clustering is obtained K cluster;Determine multiple key monitoring point information, multiple general monitoring point information based on the multiple high-sensitive deformation region information, the multiple medium-sensitive deformation region information;Determine target monitoring point based on the multiple key monitoring point information, the multiple general monitoring point information;Based on the target monitoring point, sensor placement is carried out and construction tunnel deformation monitoring is carried out, the method can accurately according to multi-source geological data Intelligent optimization arrangement of monitoring point is carried out to improve the accuracy of tunnel deformation monitoring.

Description

Technical Field

[0001] This invention relates to the field of construction tunnel monitoring technology, specifically to a method and system for monitoring the deformation of construction tunnels based on TBM (Tunnel Boring Machine). Background Technology

[0002] Full-face tunnel boring machines (TBMs) have become core equipment in modern long-distance tunnel construction, and their construction safety and efficiency directly affect the success or failure of the project. During and after TBM excavation, tunnel structures are prone to deformation and even collapse due to complex geological conditions, construction disturbances, and stress redistribution in the rock mass. Therefore, real-time and accurate deformation monitoring is crucial. Current technologies mainly rely on manual experience to deploy sensors for monitoring, but this method lacks scientific analysis of the overall deformation trend of the tunnel in selecting monitoring points. This experience-based approach can easily lead to the omission of critical high-risk deformation areas and may result in excessive investment of monitoring resources in low-risk areas. Faced with multi-source geological data generated by TBM construction, traditional methods struggle to achieve deep data fusion and intelligent analysis, making it difficult to dynamically and accurately identify highly sensitive and moderately sensitive deformation areas, thus hindering the development of optimal monitoring point layout schemes.

[0003] Therefore, how to accurately optimize the layout of monitoring points based on multi-source geological data to improve the accuracy of tunnel deformation monitoring is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem this invention addresses is how to accurately optimize the layout of monitoring points based on multi-source geological data to improve the accuracy of tunnel deformation monitoring.

[0005] According to a first aspect, the present invention provides a method for monitoring the deformation of a TBM-based construction tunnel, comprising: acquiring multi-source geological data after the construction of the TBM-based construction tunnel; determining multiple highly sensitive deformation points of the TBM-based construction tunnel based on the multi-source geological data after the construction of the TBM-based construction tunnel; clustering the multiple highly sensitive deformation points of the TBM-based construction tunnel to obtain K clusters; processing the K clusters and the multi-source geological data after the construction of the TBM-based construction tunnel using a deformation region determination model to obtain multiple highly sensitive deformation region information and multiple moderately sensitive deformation region information; determining multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation region information and the multiple moderately sensitive deformation region information; determining target monitoring points based on the multiple key monitoring point information and the multiple general monitoring point information; and placing sensors based on the target monitoring points and performing deformation monitoring of the construction tunnel.

[0006] In one possible implementation, determining multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation region information and the multiple moderately sensitive deformation region information includes: constructing a deformation analysis map, which includes multiple highly sensitive deformation region nodes, multiple moderately sensitive deformation region nodes, and edges between nodes; the node features of the highly sensitive deformation region nodes are highly sensitive deformation region information, and the node features of the moderately sensitive deformation region nodes are moderately sensitive deformation region information; and processing the deformation analysis map based on a graph neural network to obtain the multiple key monitoring point information and multiple general monitoring point information.

[0007] In one possible implementation, determining the target monitoring point based on the multiple key monitoring point information and the multiple general monitoring point information includes: determining multiple candidate monitoring schemes based on the multiple key monitoring point information and the multiple general monitoring point information, each candidate monitoring scheme including multiple candidate monitoring points; generating simulated monitoring information for each candidate monitoring scheme based on the multiple candidate monitoring schemes; determining mandatory monitoring point information based on the simulated monitoring information for each candidate monitoring scheme; and determining the remaining target monitoring points based on the mandatory monitoring point information and the simulated monitoring information for each candidate monitoring scheme.

[0008] In one possible implementation, the deformation region determination model is a deep neural network.

[0009] According to a second aspect, the present invention provides a TBM-based tunnel deformation monitoring system, comprising:

[0010] The acquisition module is used to acquire multi-source geological data after the construction of the TBM tunnel.

[0011] The high-sensitivity deformation point determination module is used to determine multiple high-sensitivity deformation point information of the TBM construction tunnel based on multi-source geological data after the construction of the TBM construction tunnel;

[0012] The clustering module is used to cluster K clusters based on the information of multiple highly sensitive deformation points of the TBM construction tunnel;

[0013] The deformation region determination module is used to process the K clusters and the multi-source geological data after the construction of the TBM tunnel using the deformation region determination model to obtain information on multiple highly sensitive deformation regions and multiple medium-sensitive deformation regions.

[0014] The monitoring point information determination module is used to determine multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation area information and the multiple medium sensitive deformation area information.

[0015] The target monitoring point determination module is used to determine the target monitoring point based on the information of the multiple key monitoring points and the information of the multiple general monitoring points.

[0016] The monitoring module is used to place sensors based on the target monitoring points and monitor the deformation of the construction tunnel.

[0017] In one possible implementation, the monitoring point information determination module is further configured to: construct a deformation analysis map, which includes multiple highly sensitive deformation region nodes, multiple moderately sensitive deformation region nodes, and edges between nodes; the node features of the highly sensitive deformation region nodes are highly sensitive deformation region information, and the node features of the moderately sensitive deformation region nodes are moderately sensitive deformation region information; and process the deformation analysis map based on a graph neural network to obtain multiple key monitoring point information and multiple general monitoring point information.

[0018] In one possible implementation, the target monitoring point determination module is further configured to: determine multiple candidate monitoring schemes based on the multiple key monitoring point information and the multiple general monitoring point information, each candidate monitoring scheme including multiple candidate monitoring points; generate simulated monitoring information for each candidate monitoring scheme based on the multiple candidate monitoring schemes; determine mandatory monitoring point information based on the simulated monitoring information for each candidate monitoring scheme; and determine the remaining target monitoring points based on the mandatory monitoring point information and the simulated monitoring information for each candidate monitoring scheme.

[0019] In one possible implementation, the deformation region determination model is a deep neural network.

[0020] According to a third aspect, embodiments of the present invention provide an electronic device, including: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method described above, the method including: acquiring multi-source geological data after the construction of a TBM tunnel; determining multiple highly sensitive deformation point information of the TBM tunnel based on the multi-source geological data after the construction of the TBM tunnel; clustering the multiple highly sensitive deformation point information of the TBM tunnel to obtain K clusters; processing the K clusters and the multi-source geological data after the construction of the TBM tunnel using a deformation region determination model to obtain multiple highly sensitive deformation region information and multiple moderately sensitive deformation region information; determining multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation region information and the multiple moderately sensitive deformation region information; determining a target monitoring point based on the multiple key monitoring point information and the multiple general monitoring point information; and placing sensors based on the target monitoring point and performing deformation monitoring of the construction tunnel.

[0021] According to the fourth aspect, this embodiment provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the aforementioned TBM-based tunnel deformation monitoring method. The method includes: acquiring multi-source geological data after the construction of the TBM tunnel; determining multiple highly sensitive deformation points of the TBM tunnel based on the multi-source geological data after the construction of the TBM tunnel; clustering the multiple highly sensitive deformation points of the TBM tunnel to obtain K clusters; processing the K clusters and the multi-source geological data after the construction of the TBM tunnel using a deformation region determination model to obtain multiple highly sensitive deformation region information and multiple moderately sensitive deformation region information; determining multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation region information and the multiple moderately sensitive deformation region information; determining a target monitoring point based on the multiple key monitoring point information and the multiple general monitoring point information; and placing sensors based on the target monitoring point and performing tunnel deformation monitoring.

[0022] This invention provides a method and system for monitoring tunnel deformation during TBM construction. The method includes acquiring multi-source geological data after TBM tunnel construction; determining multiple highly sensitive deformation points of the TBM tunnel based on the multi-source geological data; clustering the multiple highly sensitive deformation points to obtain K clusters; processing the K clusters and the multi-source geological data using a deformation region determination model to obtain multiple highly sensitive deformation regions and multiple moderately sensitive deformation regions; determining multiple key monitoring points and multiple general monitoring points based on the multiple highly sensitive and moderately sensitive deformation regions; determining target monitoring points based on the multiple key and general monitoring points; and placing sensors based on the target monitoring points to monitor tunnel deformation. This method can accurately optimize the intelligent layout of monitoring points based on multi-source geological data to improve the accuracy of tunnel deformation monitoring. Attached Figure Description

[0023] Figure 1 A schematic flowchart of a TBM-based tunnel deformation monitoring method provided in an embodiment of the present invention;

[0024] Figure 2 A schematic diagram of a tunnel boring machine provided in an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of a construction tunnel provided for an embodiment of the present invention;

[0026] Figure 4This is a flowchart illustrating the process of determining information on multiple key monitoring points and multiple general monitoring points, provided in an embodiment of the present invention.

[0027] Figure 5 This is a schematic diagram of a process for determining target monitoring points provided in an embodiment of the present invention;

[0028] Figure 6 This is a schematic diagram of a TBM-based tunnel deformation monitoring system provided in an embodiment of the present invention.

[0029] Figure 7 This is a schematic diagram of the structure of an electronic device provided in one embodiment of this specification. Detailed Implementation

[0030] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0031] As used herein, the term "comprising" and its variations are open terms meaning "including but not limited to". The term "based on" means "at least partially based on". The terms "one embodiment" and "an embodiment" mean "at least one embodiment". The term "another embodiment" means "at least one other embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other definitions, whether explicit or implicit, may be included below. Unless explicitly indicated by the context, the definition of a term shall remain consistent throughout the specification.

[0032] In this embodiment of the invention, the following are provided: Figure 1 The method for monitoring the deformation of a TBM-based tunnel includes steps S1 to S7:

[0033] Step S1: Obtain multi-source geological data after the TBM tunnel construction.

[0034] A tunnel boring machine (TBM) is a large-scale tunnel construction machine that integrates functions such as excavation, support, and muck removal. TBMs can continuously excavate to form tunnels by rotating a cutterhead. Figure 2 This is a schematic diagram of a tunnel boring machine provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of a construction tunnel provided in an embodiment of the present invention.

[0035] Multi-source geological data after TBM tunnel construction refers to data collected after TBM excavation that reflects the surrounding rock and geological conditions of the tunnel. This multi-source geological data includes geological sketch data of the tunnel face, ground-penetrating radar data, core sampling analysis results from representative locations, and rock strength data from representative locations.

[0036] The geological sketch data of the tunnel face is information obtained through on-site geological logging of the tunnel excavation face. The geological sketch data of the tunnel face includes lithological boundary coordinates, joint occurrence parameters, fault trace location information, and the degree of rock weathering.

[0037] Ground-penetrating radar (GPR) data is information obtained by scanning and detecting the surrounding rock and strata of a tunnel using GPR equipment.

[0038] Ground-penetrating radar data can reflect the distribution of geological structures at a certain depth underground, including the location of stratigraphic interfaces, potential weak interlayers, and the distribution range and morphology of karst caves or fissure zones.

[0039] The core sampling analysis results are obtained by analyzing the characteristics of the core samples taken after drilling at representative locations in the tunnel. The results include information such as rock type, mineral composition, and water content of the core samples.

[0040] Rock strength data are quantitative data used to represent the rock mass's ability to resist external forces, determined through mechanical testing after drilling and sampling at representative locations in the tunnel. Rock strength data includes the rock's uniaxial compressive strength, tensile strength, and shear strength.

[0041] Step S2: Based on the multi-source geological data after the TBM construction tunnel is constructed, determine the information of multiple highly sensitive deformation points of the TBM construction tunnel.

[0042] In some embodiments, a high-sensitivity deformation point determination model can be used to determine information on multiple high-sensitivity deformation points in the TBM construction tunnel. The high-sensitivity deformation point determination model is a Transformer model. The input to the high-sensitivity deformation point determination model is multi-source geological data after the TBM construction tunnel is completed, and the output of the high-sensitivity deformation point determination model is information on multiple high-sensitivity deformation points in the TBM construction tunnel.

[0043] The Transformer model is a deep learning model based on a self-attention mechanism. Its core structure includes an encoder and a decoder. The encoder extracts features from the input sequence through multiple self-attention layers and a feedforward neural network, while the decoder generates the target sequence using the encoder output and a masked self-attention mechanism. The Transformer model effectively captures long-range dependencies and is suitable for sequence data modeling tasks.

[0044] Information on multiple highly sensitive deformation points in TBM-constructed tunnels is specific information about suspected locations within the tunnel with a high risk of future deformation, determined by a highly sensitive deformation point identification model. This information includes the predicted three-dimensional spatial coordinates of the high-risk locations, the predicted stress concentration values ​​for the high-sensitive deformation points, and the potential rock mass slip direction.

[0045] Multi-source geological data collected after TBM tunnel construction includes the physical and mechanical properties of the rock mass and the spatial distribution characteristics of structural planes. This data exhibits a continuous spatial distribution, and geological parameters at different locations interact non-linearly. For example, areas where low-strength rock overlaps with high-density joints are often prone to deformation. The Transformer model can utilize the spatial sequence characteristics of this multi-source data to identify potential weak points.

[0046] The Transformer model utilizes a self-attention mechanism to deeply encode multi-source geological data after TBM tunnel construction. During encoding, the model calculates attention weights between geological parameters at different locations, thereby identifying global geological structural relationships, such as the potential impact of fault traces at one location on the stability of distant rock masses. The decoder, based on the encoded geological feature context, predicts stress distribution trends and stability states along the tunnel. Through fully connected layers, the model maps high-dimensional features to specific deformation risk probability distributions, thereby filtering out spatial locations with risk probabilities exceeding a preset threshold and outputting the coordinates and related physical parameters of these locations as multiple highly sensitive deformation points for the TBM tunnel construction.

[0047] In some embodiments, determining the information of multiple highly sensitive deformation points of the TBM construction tunnel based on multi-source geological data after the TBM construction tunnel includes steps S21 to S23:

[0048] Step S21: Based on the multi-source geological data after the TBM construction tunnel is completed, determine the rock strength distribution map, joint density distribution map, and geological anomaly index sequence.

[0049] In some embodiments, the Transformer model can be used to determine rock intensity distribution maps, joint density distribution maps, and geological anomaly index sequences.

[0050] The rock strength distribution map is a two-dimensional image output by the Transformer model, showing the spatial variation of the hardness and compressive strength of the surrounding rock along the tunnel. Each pixel on the rock strength distribution map corresponds to a uniaxial compressive strength value of the rock at a specific location within the tunnel.

[0051] The joint density distribution map is an image output by the Transformer model showing the degree of development and density of fractures at various locations inside the tunnel surrounding rock.

[0052] Joint density distribution maps can reflect the spatial distribution pattern of rock mass fragmentation.

[0053] The geological anomaly index sequence is a set of values ​​output by the Transformer model, arranged in order of tunnel excavation mileage, that reflects the degree to which geological conditions deviate from the normal range.

[0054] Each indicator value can comprehensively quantify the degree of anomaly in the surrounding rock geological conditions at the corresponding mileage. The higher the value, the greater the possibility of the presence of unfavorable geological bodies such as faults or fracture zones at that location.

[0055] The Transformer model utilizes a self-attention mechanism to deeply explore the spatial variation patterns of multi-source geological data after TBM tunnel construction, mapping these one-dimensional sequence data into spatial distribution features. By learning the intrinsic correlations between different geological parameters, the Transformer model can accurately deduce the spatial distribution of rock strength and joint density, and identify anomalous fluctuations in the sequence, thereby determining rock strength distribution maps, joint density distribution maps, and geological anomaly index sequences.

[0056] Step S22: Based on the rock strength distribution map, the joint density distribution map, and the geological anomaly index sequence, determine multiple potential risk points, the comprehensive risk score of each potential risk point, and the spatial correlation index between risk points.

[0057] In some embodiments, the Transformer model can be used to determine multiple potential risk points, a comprehensive risk score for each potential risk point, and a spatial correlation index between risk points.

[0058] Multiple potential risk points are a set of points within the tunnel identified by the Transformer model that contain unstable factors.

[0059] Potential risk points may include locations with low rock strength, high joint density, or abnormal geological indicators.

[0060] The overall risk score for each potential risk point is a quantitative assessment of the likelihood of deformation at each potential risk point, determined by the Transformer model.

[0061] The spatial correlation index between risk points is a numerical value determined by the Transformer model, which describes the strength of mutual influence between different potential risk points on deformation risk.

[0062] Spatial correlation indices between risk points can reveal the chain reaction and patchy distribution characteristics of geological disasters in space.

[0063] The Transformer model, through its multi-head self-attention mechanism, can capture spatial local details in the rock strength distribution map and joint density distribution map, as well as long-range trend features in the geological anomaly index sequence, thereby achieving comprehensive and coordinated perception of geological risks along the tunnel. Based on this comprehensive perception, the Transformer model can accurately identify the spatial overlap between low-rock-strength areas and high-joint-density areas, and locate the core locations of these overlapping areas as potential risk points. The Transformer model can quantify the independent risk level of each potential risk point by analyzing the comprehensive geological conditions, thus generating a comprehensive risk score. By calculating the correlation between feature vectors of different points, the Transformer model can assess the influence of a point's risk status on its surrounding points, thereby determining the spatial correlation index between risk points.

[0064] Step S23: Based on the multiple potential risk points, the comprehensive risk score of each potential risk point, and the spatial correlation index between the risk points, determine the information of multiple highly sensitive deformation points in the TBM construction tunnel.

[0065] In some embodiments, deep neural networks can be used to determine information on multiple highly sensitive deformation points in a TBM construction tunnel.

[0066] A deep neural network (DNN) is an artificial neural network containing multiple hidden layers. Through hierarchical feature extraction and nonlinear transformations, DNNs can approximate complex functional relationships. A deep neural network consists of an input layer, multiple hidden layers, and an output layer. It can utilize the backpropagation algorithm to adjust weights to minimize the loss function.

[0067] Deep neural networks can aggregate the spatial coordinates of each potential risk point, its own comprehensive risk score, and its spatial correlation index with other points into a comprehensive feature vector. In the hidden layers, through multiple transformations and feature integrations using non-linear activation functions, the deep neural network can assess whether the independent risk of each point reaches a high-sensitivity threshold and further determine its importance in the entire risk network. For example, a potential risk point with a high self-score and strong correlation with multiple high-risk points will have a significantly higher priority in being classified as a high-sensitivity deformation point. Finally, the output layer of the deep neural network can classify each potential risk point according to the learned rules, selecting a subset of points that truly require high attention, thereby generating information on multiple high-sensitivity deformation points in TBM construction tunnels.

[0068] Step S3: Based on the information of multiple highly sensitive deformation points of the TBM construction tunnel, K clusters are obtained.

[0069] The clustering method used is K-means clustering. K-means clustering is an unsupervised learning algorithm that divides samples in a dataset into K distinct clusters, ensuring that each sample belongs to the cluster corresponding to its nearest cluster center. K-means clustering iteratively updates the cluster centers, minimizing the sum of squared errors among samples within a cluster, ultimately achieving data clustering. The value of K in K-means clustering can be pre-determined manually.

[0070] The K clusters are formed by using the K-means clustering algorithm to divide information from multiple highly sensitive deformation points based on feature similarity. Each cluster contains a group of highly sensitive deformation points that are similar in spatial location or geological features, and the highly sensitive deformation points within a cluster have a high degree of similarity.

[0071] Information on multiple highly sensitive deformation points includes their spatial coordinates and geological features, providing measurable attribute dimensions for clustering algorithms. K-means clustering can group points with similar characteristics into the same cluster, thus revealing the distribution patterns of deformation points.

[0072] The process of clustering multiple highly sensitive deformation points in TBM construction tunnels using the K-means clustering algorithm is as follows: First, K points are randomly selected from the dataset of highly sensitive deformation points as initial cluster centers. Next, for each highly sensitive deformation point in the dataset, Euclidean distance is used to measure and calculate its distance to these K initial cluster centers, and the highly sensitive deformation point is assigned to the corresponding cluster according to the principle of closest proximity. After all highly sensitive deformation points have been partitioned, the average values ​​of various features of the highly sensitive deformation points within each cluster are recalculated to update the cluster centers of each cluster. This process of partitioning and updating cluster centers is repeated until the changes in cluster centers are minimal. At this point, the clustering process is considered to have converged, and the clustering is complete.

[0073] Clustering highly sensitive deformation points into K clusters allows for grouping points with similar geological structural features and deformation risk patterns into one category, thus achieving structured grouping of dispersed deformation points. This grouping method can clearly distinguish deformation risk units with different characteristics and avoid indiscriminate analysis of all deformation points. Subsequently, based on the overall characteristics of the clusters, high- and medium-sensitive deformation areas can be accurately identified, improving the efficiency and accuracy of deformation area delineation.

[0074] Step S4: Based on the K clusters and the multi-source geological data after the TBM construction tunnel, the deformation region determination model is used to process and obtain information on multiple highly sensitive deformation regions and multiple moderately sensitive deformation regions.

[0075] The deformation region determination model is a deep neural network. The input to the deformation region determination model is the K clusters and the multi-source geological data after the TBM tunnel construction. The output of the deformation region determination model is information on multiple highly sensitive deformation regions and information on multiple moderately sensitive deformation regions.

[0076] Highly sensitive deformation area information is a detailed description of the spatial areas within the tunnel with a high risk of future deformation, determined by the deformation area determination model.

[0077] The information on medium-sensitive deformation areas is a specific description of the spatial areas within the tunnel where the risk of future deformation is at a medium level, as determined by the deformation area determination model.

[0078] The deformation area information includes the spatial boundary range of the area, the average rock mass strength grade within the area, the predicted maximum deformation, the distribution map of the geological risk impact degree of the deformation area, the distribution map of the structural stability of the deformation area, the distribution map of the rock strength of the deformation area, and the distribution map of the joint density of the deformation area.

[0079] The geological risk impact distribution map of the deformation area is an image that shows the distribution of the intensity of geological risk impact at different locations within a specific deformation area.

[0080] The structural stability distribution map of the deformation region is an image that represents the spatial distribution of the stability and safety level of the tunnel structure at different locations in a specific deformation region.

[0081] A rock intensity distribution map of a deformation area is an image that represents the spatial distribution of rock intensity at different locations within a specific deformation area.

[0082] A joint density distribution map of a deformation region is an image that represents the spatial distribution of the density of rock joint development at different locations within a specific deformation region.

[0083] The K clusters provide an overview of the spatial distribution of potential risk points. The multi-source geological data after the construction of the TBM tunnel provides detailed geological background data around the clusters, which can help the model determine the extent of risk expansion and the degree of attenuation. This allows the model to not only identify the location of the risk points, but also infer the extent and level of the entire affected connected area based on the geological conditions of the surrounding rock mass.

[0084] Deep neural networks can fuse the spatial distribution characteristics of K clusters with multi-source geological data after TBM tunnel construction. The input layer of the deep neural network receives the coordinates of the cluster center, the density of points within the cluster, the corresponding rock strength, and the rock mass integrity characteristics reflected by the cluster information. The hidden layers can analyze the coupling effect between the cluster center and the surrounding geological environment through nonlinear activation functions, such as determining whether a high-density cluster of risk points is located in a fractured rock mass zone, thereby inferring the spatial radiation range of deformation risk. Through layer-by-layer feature abstraction, the deep neural network can expand discrete clusters into continuous spatial regions and calculate the comprehensive risk index for each region. The output layer can then classify the tunnel space into a grid based on the calculated risk index and geological boundary conditions, identifying grid sets with risk indices exceeding a high threshold as highly sensitive deformation areas and grid sets with risk indices between a medium threshold as moderately sensitive deformation areas.

[0085] Step S5: Based on the information of the multiple highly sensitive deformation regions and the information of the multiple moderately sensitive deformation regions, determine the information of multiple key monitoring points and multiple general monitoring points.

[0086] In some embodiments, Figure 4 This invention provides a flowchart illustrating the process of determining information on multiple key monitoring points and multiple general monitoring points, wherein determining the information on multiple key monitoring points and multiple general monitoring points includes steps S51-S52:

[0087] Step S51: Construct a deformation analysis map. The deformation analysis map includes multiple highly sensitive deformation region nodes, multiple medium sensitive deformation region nodes, and edges between nodes. The node features of highly sensitive deformation region nodes are information about highly sensitive deformation regions, and the node features of medium sensitive deformation region nodes are information about medium sensitive deformation regions.

[0088] A graph is a data structure composed of vertices and edges that can represent the relationships between objects. Deformation analysis graphs are a type of graph data that can represent the adjacency relationships and degree of influence between deformed areas of a tunnel.

[0089] In the deformation analysis map, each node corresponds to a deformation region. The deformation analysis map includes multiple highly sensitive deformation region nodes, multiple moderately sensitive deformation region nodes, and edges between nodes. The node characteristics of highly sensitive deformation region nodes are information about highly sensitive deformation regions, and the node characteristics of moderately sensitive deformation region nodes are information about moderately sensitive deformation regions. The edges between nodes are used to represent the association between nodes of different deformation regions, and the edges represent the positional relationship between the deformation region nodes.

[0090] The positional relationship between nodes in the deformation zone refers to the relative orientation and relative distance between nodes in different sensitive deformation zones in the three-dimensional space of the tunnel.

[0091] Step S52: Process the deformation analysis map based on the graph neural network to obtain information on multiple key monitoring points and multiple general monitoring points.

[0092] Graph Neural Networks (GNNs) are deep learning models capable of processing graph-structured data. GNNs learn a global representation of the graph by aggregating neighbor node information and updating node features. GNNs can capture dependencies between nodes in the graph through message passing mechanisms. The input to the GNN is the deformation analysis graph, and the output is information on multiple key monitoring points and multiple general monitoring points.

[0093] Information on multiple key monitoring points is determined through a graph neural network, identifying key locations that require high-frequency, high-precision monitoring.

[0094] The information of multiple general monitoring points is the routine location information determined by graph neural networks to assist in monitoring and improve the overall deformation field data.

[0095] The monitoring point information includes the specific location of the monitoring point, the rock mass strength grade, joint density, risk grade, stability grade, and monitoring priority.

[0096] By constructing a deformation analysis map, the positional relationship network between nodes in sensitive deformation areas can be clearly reflected. Since the risk propagation effect of sensitive deformation areas is directly related to the spatial correlation between regions, using the spatial boundary range, main geological risk types, average rock mass strength grade, and predicted maximum deformation of nodes in sensitive deformation areas as node features, while using orientation and distance as edges, allows for more comprehensive utilization of the core information of the deformation area. This helps graph neural networks better understand the risk correlation and transmission patterns between sensitive deformation areas, thereby improving the accuracy of monitoring point identification. Processing deformation analysis map data based on graph neural networks can effectively learn the complex relationships and information transmission between nodes, thus more accurately uncovering the distribution characteristics of key monitoring points and general monitoring points.

[0097] Graph neural networks (Graph Neural Networks) can perform graph convolution operations on deformation analysis maps. In each layer of graph convolution, nodes in highly sensitive and moderately sensitive deformation regions transmit their deformation risk characteristics to neighboring nodes through edges. The Graph Neural Network can aggregate the neighborhood information received by each node and update the node's hidden state, ensuring that each node not only contains its own local region information but also incorporates the stress environment characteristics of the surrounding region. After multiple layers of propagation, the Graph Neural Network can analyze the stress concentration and topological importance within each region, thereby identifying points located at the center of high-risk regions or at key topological locations connecting regions, and decoding them into multiple key monitoring point information. Simultaneously, the Graph Neural Network can also identify points used to cover the background deformation field, decoding them into multiple general monitoring point information.

[0098] Step S6: Determine the target monitoring point based on the information of the multiple key monitoring points and the information of the multiple general monitoring points.

[0099] Target monitoring points refer to the locations selected from multiple key monitoring points and multiple general monitoring points to place sensors for monitoring, in order to achieve comprehensive and effective monitoring of tunnel deformation during TBM construction. Target monitoring points include mandatory monitoring points and other target monitoring points.

[0100] The mandatory monitoring points are the core points that are indispensable in deformation monitoring.

[0101] The remaining target monitoring points are supplementary points to the mandatory monitoring points to achieve comprehensive monitoring.

[0102] In some embodiments, Figure 5 This is a schematic diagram of a process for determining a target monitoring point according to an embodiment of the present invention. The determination of the target monitoring point includes steps S61 to S64:

[0103] Step S61: Based on the information of the multiple key monitoring points and the information of the multiple general monitoring points, determine multiple candidate monitoring schemes, each candidate monitoring scheme including multiple candidate monitoring points.

[0104] In some embodiments, a monitoring scheme determination model, which is a deep neural network, can be used to determine multiple candidate monitoring schemes. The input of the monitoring scheme determination model is the information of the multiple key monitoring points and the information of the multiple general monitoring points, and the output of the monitoring scheme determination model is the multiple candidate monitoring schemes.

[0105] Multiple candidate monitoring schemes are a set of alternative schemes for tunnel deformation monitoring generated by the monitoring scheme determination model based on information from multiple key monitoring points and multiple general monitoring points. Each candidate monitoring scheme includes multiple candidate monitoring points.

[0106] The multiple candidate monitoring points in each candidate monitoring scheme are different subsets of key monitoring point information and general monitoring point information, and can represent different monitoring focuses and coverage strategies.

[0107] Key monitoring point information clarifies the key monitoring targets in high-risk areas of the tunnel, while general monitoring point information supplements the auxiliary monitoring targets in medium-risk areas. The attribute differences and spatial distribution characteristics of the two types of point information provide the preconditions for differentiated subset combinations of monitoring schemes, enabling the monitoring scheme determination model to select and combine multiple candidate monitoring schemes with differentiated characteristics based on different monitoring focuses and coverage strategies.

[0108] Deep neural networks can analyze the synergistic effects of different monitoring point combinations by learning the strengths and weaknesses of historical monitoring layouts. Under the premise of meeting pre-set monitoring coverage and cost constraints, deep neural networks can calculate the probability distribution of various monitoring point combinations that satisfy the constraints. The output layer can then sample and generate multiple differentiated monitoring point lists based on these probability distributions, with each list constituting a candidate monitoring scheme. For example, some schemes focus on densely surrounding highly sensitive areas, while others focus on uniform, discrete coverage throughout the tunnel.

[0109] Step S62: Generate simulated monitoring information for each candidate monitoring scheme based on the multiple candidate monitoring schemes.

[0110] In some embodiments, a generative adversarial network (GAN) can be used to generate simulated monitoring information for each candidate monitoring scheme. The input to the GAN is the plurality of candidate monitoring schemes, and the output of the GAN is simulated monitoring information for each candidate monitoring scheme.

[0111] Generative Adversarial Networks (GANs) consist of two parts: a generator and a discriminator. The generator is responsible for generating data that is as realistic as possible, while the discriminator is responsible for distinguishing between real and generated data. Through adversarial training, both continuously improve the quality of the generated data, ultimately enabling the generator to produce data that conforms to the true distribution.

[0112] The simulated monitoring information for each candidate monitoring scheme is generated by a generative adversarial network (GAN) and obtained from sensor monitoring data acquired under a simulated actual monitoring scenario corresponding to the combination of monitoring points for that candidate scheme. The simulated monitoring information includes the predicted displacement sequence, stress change curve, and complementary information on monitoring blind zones between monitoring points for each monitoring point in the scheme over a future period.

[0113] Complementary information on monitoring blind spots between monitoring points includes the percentage overlap of monitoring ranges of adjacent points, the percentage of uncovered areas in highly sensitive regions, and the cross-validation rate of deformation anomalies at different points.

[0114] Multiple candidate monitoring schemes identify different combinations of monitoring points. Each scheme has different characteristics such as point distribution and risk coverage. These differences provide diverse input for generative adversarial networks, enabling the model to generate simulated data that matches the actual monitoring effect of different schemes.

[0115] The generator in a Generative Adversarial Network (GAN) first extracts features from multiple candidate monitoring schemes, including the location of candidate monitoring points, their corresponding risk levels, and monitoring priorities. These features are then mapped into high-dimensional vectors. Based on the distribution patterns of historical real monitoring data and the extracted features of candidate monitoring schemes, the generator produces preliminary simulated monitoring information. A discriminator receives the simulated monitoring information generated by the generator and historical real monitoring data, distinguishes between them, and feeds the judgment back to the generator. The generator adjusts its generation strategy based on the feedback to continuously optimize the realism of the simulated monitoring information. After multiple rounds of adversarial training, the generator can output simulated monitoring information for each candidate monitoring scheme that is consistent with the distribution of real monitoring data.

[0116] Step S63: Determine the required monitoring point information based on the simulated monitoring information of each candidate monitoring scheme.

[0117] In some embodiments, a monitoring point determination model can be used to determine the mandatory monitoring point information. The monitoring point determination model is a deep neural network. The input to the monitoring point determination model is the simulated monitoring information of each candidate monitoring scheme, and the output of the monitoring point determination model is the mandatory monitoring point information.

[0118] The mandatory monitoring point information consists of monitoring point data selected from numerous candidate points by the monitoring point determination model. These monitoring points play a decisive role in capturing tunnel deformation characteristics. The mandatory monitoring point information includes the specific location of each mandatory monitoring point, its monitoring priority, and its importance weight within the monitoring network.

[0119] The simulated monitoring information for each candidate monitoring scheme simulates the monitoring effect under different combinations of points. Among them, the point information that can stably reflect the deformation of high-risk areas and play a decisive role in the monitoring results is the key basis for determining the mandatory monitoring points.

[0120] Deep neural networks can perform in-depth analysis of simulated monitoring information for each candidate monitoring scheme. The input layer receives features from the simulated monitoring information, such as monitoring data accuracy, coverage of risk areas, and ability to capture deformation trends. The hidden layers use non-linear activation functions to abstract and integrate these features layer by layer, and mine the monitoring performance patterns of different monitoring points in various candidate schemes, thereby evaluating the coreity and irreplaceability of each monitoring point. By calculating the contribution value of each monitoring point to the overall monitoring effect in simulated monitoring, the model can screen out monitoring points whose contribution value is consistently high and whose absence would lead to a significant drop in monitoring effect. The output layer integrates the information of these core monitoring points to form the information of essential monitoring points.

[0121] Step S64: Determine the remaining target monitoring points based on the information of the required monitoring points and the simulated monitoring information of each candidate monitoring scheme.

[0122] In some embodiments, a second monitoring point determination model can be used to determine the remaining target monitoring points. The second monitoring point determination model is a deep neural network. The input to the second monitoring point determination model is the required monitoring point information and the simulated monitoring information of each candidate monitoring scheme; the output of the second monitoring point determination model is the remaining target monitoring points.

[0123] The remaining target monitoring points are supplementary points selected by combining the information of the mandatory monitoring points and the simulated monitoring information of each candidate monitoring scheme through the second monitoring point determination model to achieve comprehensive monitoring.

[0124] The remaining target monitoring points include the specific location, monitoring priority, and importance weight in the monitoring network for each remaining target monitoring point.

[0125] The mandatory monitoring point information clarifies the locations that must be prioritized for monitoring. The simulated monitoring information for each candidate monitoring scheme provides the overall potential data distribution. By analyzing the mandatory monitoring point information and the simulated monitoring information for each candidate monitoring scheme, the model can determine whether there are any monitoring blind spots at the current mandatory monitoring points, and then find the best supplementary locations to fill these blind spots from the simulated information.

[0126] Deep neural networks can use mandatory monitoring point information as baseline features and input simulated monitoring information from each candidate monitoring scheme as a pool of potential features. In the hidden layers, the model calculates the effective monitoring range covered by the mandatory monitoring points and compares it with the full-field deformation features shown by the simulated monitoring information. The model identifies spatial regions or deformation patterns not yet fully covered by the mandatory monitoring points by calculating residuals and unexplained variance. For example, the model might find that while the mandatory monitoring points control the main fault zone, they lack sufficient control points for the uniform settlement of the tunnel arch. The deep neural network then searches the simulated monitoring information for points that can minimize this monitoring blind spot effect. The model evaluates the spatial complementarity between the remaining points and the mandatory points, prioritizing points with low correlation to the mandatory point data but high representativeness in the uncovered area. Through this complementarity analysis and optimization calculation, the deep neural network ultimately selects a set of points that maximize the robustness and coverage of the overall monitoring network, identifying these as the remaining target monitoring points.

[0127] Step S7: Place sensors based on the target monitoring points and monitor the deformation of the construction tunnel.

[0128] Once the target monitoring points are identified, high-precision displacement sensors and stress sensors are installed on the mandatory monitoring points and other target monitoring points to accurately capture displacement changes and rock stress fluctuations in the core risk areas, thereby achieving comprehensive and effective monitoring of tunnel deformation during construction.

[0129] Please refer to the following. Figure 6 , Figure 6 A schematic diagram of a TBM-based tunnel deformation monitoring system provided in an embodiment of this specification is shown. It should be noted that... Figure 6 The TBM-based tunnel deformation monitoring system shown is used to execute this specification. Figure 1 The methods shown in the embodiments are illustrated for ease of explanation, showing only the parts related to the embodiments of this specification. For specific technical details not disclosed, please refer to this specification. Figure 1 The example shown.

[0130] Based on the same inventive concept Figure 6 This invention provides a schematic diagram of a TBM-based tunnel deformation monitoring system, which includes:

[0131] Module 81 is used to acquire multi-source geological data after the construction of the TBM tunnel.

[0132] The high-sensitivity deformation point determination module 82 is used to determine multiple high-sensitivity deformation point information of the TBM construction tunnel based on the multi-source geological data after the construction of the TBM construction tunnel.

[0133] Clustering module 83 is used to cluster K clusters based on the information of multiple highly sensitive deformation points of the TBM construction tunnel;

[0134] The deformation region determination module 84 is used to process the K clusters and the multi-source geological data after the construction of the TBM tunnel using the deformation region determination model to obtain information on multiple highly sensitive deformation regions and multiple medium-sensitive deformation regions.

[0135] The monitoring point information determination module 85 is used to determine multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation area information and the multiple medium sensitive deformation area information.

[0136] The target monitoring point determination module 86 is used to determine the target monitoring point based on the information of the multiple key monitoring points and the information of the multiple general monitoring points;

[0137] The monitoring module 87 is used to place sensors based on the target monitoring point and monitor the deformation of the construction tunnel.

[0138] See Figure 7 It shows a schematic diagram of the structure of an electronic device according to an embodiment of this specification, which can be used to implement... Figure 1 The method in the illustrated embodiment. (As shown) Figure 7 As shown, the electronic device 700 may include: at least one central processing unit 701, at least one network interface 704, user interface 703, memory 705, and at least one communication bus 702.

[0139] The communication bus 702 is used to enable communication between these components.

[0140] The user interface 703 may include a display screen and a camera. Optionally, the user interface 703 may also include a standard wired interface and a wireless interface.

[0141] The network interface 704 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0142] The processor 701 may include one or more processing cores. The processor 701 connects to various parts within the electronic device 700 using various interfaces and lines, and performs various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 705, and by calling data stored in the memory 705. Optionally, the processor 701 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 701 may integrate one or a combination of several of the following: a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 701 and may be implemented as a separate chip.

[0143] The memory 705 may include random access memory (RAM) or read-only memory. Optionally, the memory 705 may include a non-transitory computer-readable storage medium. The memory 705 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 705 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 705 may also be at least one storage system located remotely from the aforementioned processor 701. Figure 7 As shown, the memory 705, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and program instructions.

[0144] exist Figure 7In the illustrated electronic device 700, the user interface 703 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 701 can be used to call the image-based interactive application stored in the memory 705.

[0145] It should be noted that, in order to simplify the descriptions disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments of this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0146] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A method for monitoring tunnel deformation during TBM construction, characterized in that, include: Obtain multi-source geological data after the construction of the TBM tunnel; Based on the multi-source geological data after the TBM construction tunnel was constructed, information on multiple highly sensitive deformation points of the TBM construction tunnel was determined. K clusters were obtained by clustering based on the information of multiple highly sensitive deformation points in the TBM construction tunnel; Based on the K clusters and the multi-source geological data after the TBM construction tunnel, the deformation area determination model is used to process and obtain information on multiple highly sensitive deformation areas and multiple medium-sensitive deformation areas. Based on the information of the multiple highly sensitive deformation regions and the information of the multiple medium-sensitive deformation regions, multiple key monitoring point information and multiple general monitoring point information are determined. The target monitoring point is determined based on the information of the multiple key monitoring points and the information of the multiple general monitoring points. Sensors are placed based on the target monitoring points to monitor the deformation of the construction tunnel.

2. The method for monitoring tunnel deformation during TBM construction as described in claim 1, characterized in that, The determination of multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation region information and the multiple moderately sensitive deformation region information includes: A deformation analysis map is constructed, which includes multiple highly sensitive deformation region nodes, multiple medium sensitive deformation region nodes, and edges between nodes. The node features of highly sensitive deformation region nodes are information about highly sensitive deformation regions, and the node features of medium sensitive deformation region nodes are information about medium sensitive deformation regions. The deformation analysis map is processed using a graph neural network to obtain information on multiple key monitoring points and multiple general monitoring points.

3. The method for monitoring tunnel deformation during TBM construction as described in claim 1, characterized in that, The determination of the target monitoring point based on the information of the multiple key monitoring points and the information of the multiple general monitoring points includes: Based on the information of the multiple key monitoring points and the information of the multiple general monitoring points, multiple candidate monitoring schemes are determined, and each candidate monitoring scheme includes multiple candidate monitoring points. Based on the multiple candidate monitoring schemes, simulated monitoring information for each candidate monitoring scheme is generated; The required monitoring point information is determined based on the simulated monitoring information of each candidate monitoring scheme; The remaining target monitoring points are determined based on the information of the required monitoring points and the simulated monitoring information of each candidate monitoring scheme.

4. The method for monitoring tunnel deformation during TBM construction as described in claim 1, characterized in that, The deformation region determination model is a deep neural network.

5. A Tunnel Deformation Monitoring System Based on TBM Construction, characterized in that, include: The acquisition module is used to acquire multi-source geological data after the construction of the TBM tunnel. The high-sensitivity deformation point determination module is used to determine multiple high-sensitivity deformation point information of the TBM construction tunnel based on multi-source geological data after the construction of the TBM construction tunnel; The clustering module is used to cluster K clusters based on the information of multiple highly sensitive deformation points of the TBM construction tunnel; The deformation region determination module is used to process the K clusters and the multi-source geological data after the construction of the TBM tunnel using the deformation region determination model to obtain information on multiple highly sensitive deformation regions and multiple medium-sensitive deformation regions. The monitoring point information determination module is used to determine multiple key monitoring point information and multiple general monitoring point information based on the multiple highly sensitive deformation area information and the multiple medium sensitive deformation area information. The target monitoring point determination module is used to determine the target monitoring point based on the information of the multiple key monitoring points and the information of the multiple general monitoring points. The monitoring module is used to place sensors based on the target monitoring points and monitor the deformation of the construction tunnel.

6. The TBM-based tunnel deformation monitoring system as described in claim 5, characterized in that, The monitoring point information determination module is also used for: A deformation analysis map is constructed, which includes multiple highly sensitive deformation region nodes, multiple medium sensitive deformation region nodes, and edges between nodes. The node features of highly sensitive deformation region nodes are information about highly sensitive deformation regions, and the node features of medium sensitive deformation region nodes are information about medium sensitive deformation regions. The deformation analysis map is processed using a graph neural network to obtain information on multiple key monitoring points and multiple general monitoring points.

7. The TBM-based tunnel deformation monitoring system as described in claim 5, characterized in that, The target monitoring point determination module is also used for: Based on the information of the multiple key monitoring points and the information of the multiple general monitoring points, multiple candidate monitoring schemes are determined, and each candidate monitoring scheme includes multiple candidate monitoring points. Based on the multiple candidate monitoring schemes, simulated monitoring information for each candidate monitoring scheme is generated; The required monitoring point information is determined based on the simulated monitoring information of each candidate monitoring scheme; The remaining target monitoring points are determined based on the information of the required monitoring points and the simulated monitoring information of each candidate monitoring scheme.

8. The TBM-based tunnel deformation monitoring system as described in claim 5, characterized in that, The deformation region determination model is a deep neural network.

9. An electronic device, characterized in that, include: processor; Memory; And a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the TBM-based tunnel deformation monitoring method as described in any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the TBM-based tunnel deformation monitoring method as described in any one of claims 1 to 4.

Images