A shield tunnel construction safety risk prevention and control method and system

Abstract

The application provides a shield tunnel construction safety risk prevention and control method, and relates to the technical field of tunnel construction, comprising the following steps: extracting risk factors from construction data by persistent homology technology, and obtaining corresponding risk coefficients; constructing a causal network based on historical risk information and prevention and control measures, inputting the construction data and the risk factors into the causal network, and obtaining the activation probability of the risk factors and the corresponding prevention and control measures; dividing the tunnel along the line into a cell grid, simulating the risk diffusion process, and obtaining first risk information; obtaining the minimum activation probability of the risk factors after implementing the prevention and control measures based on the causal network, simulating the risk diffusion process based on the minimum activation probability, and obtaining second risk information; and recommending the prevention and control measures by comparing the first risk information and the second risk information. The application realizes accurate identification, dynamic evaluation and scientific prevention and control of the shield tunnel construction risk, effectively reduces the construction risk, and ensures construction safety.

Description

Technical Field

[0001] This invention relates to the field of tunnel construction technology, and more specifically, to a method and system for preventing and controlling safety risks in shield tunnel construction. Background Technology

[0002] Shield tunneling is one of the core technologies for urban underground space development. With the acceleration of urbanization, its application scope has expanded from traditional subway construction to multiple fields such as integrated utility tunnels, river-crossing tunnels, and deep-buried underground engineering. However, due to the characteristics of shield tunneling, such as its high degree of concealment, variable geological conditions, and complex surrounding environment, it faces multiple risks during construction, including ground instability, equipment failure, surface subsidence, and sudden water and sand inrush. Traditional risk prevention and control mainly rely on manual experience judgment and decentralized monitoring, which makes it difficult to achieve real-time risk perception, accurate prediction, and rapid response, resulting in delayed accident warnings and low handling efficiency, seriously affecting project safety and progress. In recent years, with the rapid development of new-generation information technology, building an intelligent and digital shield tunneling risk prevention and control system to achieve dynamic perception, intelligent analysis, and proactive prevention and control of construction risks has become an inevitable trend for improving tunnel construction safety and a key technical challenge that the industry urgently needs to overcome. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for preventing and controlling safety risks in shield tunnel construction, so as to improve the above-mentioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0004] In a first aspect, the present invention provides a method for preventing and controlling safety risks in shield tunnel construction, including:

[0005] Acquire construction data, extract risk factors from the construction data using persistent coherence technology, and obtain the corresponding risk coefficients;

[0006] A causal network is constructed based on historical risk information and prevention and control measures. Construction data and risk factors are input into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures.

[0007] The tunnel route is divided into cellular grids, and the risk diffusion process is simulated based on risk factors, risk coefficients and activation probabilities to obtain the first risk information.

[0008] The lowest activation probability of risk factors after the implementation of prevention and control measures is obtained based on the causal network. The risk diffusion process is simulated based on the lowest activation probability to obtain the second risk information.

[0009] Prevention and control measures are recommended by comparing the first and second risk information.

[0010] Secondly, this application also provides a safety risk prevention and control system for shield tunnel construction, including:

[0011] The extraction module is used to acquire construction data, extract risk factors from the construction data through persistent cohomology technology, and obtain the corresponding risk coefficients;

[0012] The module is used to construct a causal network based on historical risk information and prevention and control measures. Construction data and risk factors are input into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures.

[0013] The first simulation module is used to divide the tunnel into cellular grids and simulate the risk diffusion process based on risk factors, risk coefficients and activation probabilities to obtain the first risk information.

[0014] The second simulation module is used to obtain the minimum activation probability of risk factors after the implementation of prevention and control measures based on the causal network, and to simulate the risk diffusion process based on the minimum activation probability to obtain the second risk information.

[0015] The recommendation module is used to recommend prevention and control measures by comparing the first risk information and the second risk information.

[0016] The beneficial effects of this invention are as follows:

[0017] This invention accurately extracts risk factors and quantifies risk coefficients during the construction process using persistent coherence technology, dynamically assesses risk activation probabilities using causal networks, and simulates the risk diffusion process using cellular grids. By comparing the risk diffusion before and after implementing control measures, the optimal control measures are intelligently recommended. This method achieves accurate identification, dynamic assessment, and scientific control of risks in shield tunnel construction, effectively reducing construction risks and ensuring construction safety.

[0018] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the method for preventing and controlling safety risks in shield tunnel construction according to an embodiment of this application.

[0021] Figure 2 This is a structural schematic diagram of the shield tunnel construction safety risk prevention and control system in this embodiment of the application.

[0022] In the diagram, the following modules are labeled: 100 - Extraction module; 110 - First mapping unit; 120 - Analysis unit; 130 - Second mapping unit; 200 - Construction module; 210 - First construction unit; 220 - Second construction unit; 230 - Third construction unit; 240 - Fourth construction unit; 250 - Update unit; 260 - Traversal unit; 270 - Optimization unit; 300 - First simulation module; 400 - Second simulation module; 500 - Recommendation module. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0024] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0025] Example 1

[0026] This embodiment provides a method for preventing and controlling safety risks in shield tunnel construction, including steps S100, S200, S300, S400 and S500.

[0027] S100. Obtain construction data, extract risk factors from the construction data through persistent coherence technology, and obtain the corresponding risk coefficients;

[0028] The construction data includes geological parameters, sensor parameters, and equipment status parameters. Geological parameters include surrounding rock grade, permeability coefficient, aquifer thickness, and fault zone width. The various sensors deployed monitor each location in real time and obtain sensor parameters such as seepage flow, earth pressure, water pressure, and settlement. Equipment status parameters include cutterhead torque, cutterhead rotation speed, propulsion speed, grouting pressure, and grouting volume.

[0029] This step specifically includes:

[0030] S110. Map the construction data into a point cloud structure or a graph structure;

[0031] For example, the time-series data of each sensor (such as earth pressure value and temperature) can be used to construct a point in a multi-dimensional space;

[0032] To construct a graph structure, data points from two spatially adjacent sensors can be connected to form edges, with the weight of the edges represented by physical distance or signal correlation.

[0033] S120. Based on the mapped construction data, the evolution of the data topology structure is analyzed by filtering complexes to obtain the topological features and the lifetime of the topological features.

[0034] This application employs persistent cohomology, a core tool in Topological Data Analysis (TDA), used to extract multi-scale topological features from data. Persistent cohomology reveals the topological structure of data by calculating the topological features of the data at different scales and recording the generation and decay processes of these features.

[0035] First, the continuous monitoring data is divided into sliding windows (e.g., one window every 10 minutes), and the topological characteristics of each window, such as connectivity, loops, and holes, are calculated.

[0036] Vietoris-Rips complexes are constructed based on pairwise distances from sensor point clouds. The distance threshold ε is gradually increased, and the "birth" and "death" times of topological features are observed. A persistent barcode or persistence diagram is output, and the lifetime of the topological features (ε_death - ε_birth) is recorded.

[0037] S130. Based on a preset lookup table, topological features are mapped to risk factors, and their lifetimes are mapped to risk coefficients.

[0038] The reference table can be pre-constructed based on expert knowledge. For example, high-dimensional holes in a pressure sensor network may correspond to abnormal pressure distribution (such as the formation of a "ring" topology in a local high-pressure area); persistent barcode splitting of 0-dimensional homology (H0) may indicate the formation of new seepage paths (such as a previously connected subset of sensors suddenly breaking apart); short-lived but dense H1 features in the persistence graph may reflect the instantaneous appearance of tiny seepage channels.

[0039] For equipment parameters, such as constructing a multi-dimensional point cloud from the multi-axis vibration signals of the cutterhead, wear can lead to changes in the topology of the vibration modes. For example, changes in the Betti number (the number of "holes" in different dimensions of space) may reflect changes in the distribution of vibration energy, indicating accelerated wear.

[0040] In addition to constructing a reference table based on expert knowledge, an analytical and predictive model can also be constructed based on historical known risk factors and topological characteristics, thereby mapping the topological characteristics of the data to possible risk factors.

[0041] The existence time of a topological feature is its lifetime, which can be considered to be positively correlated with the risk intensity. It can be mapped to a risk coefficient e through normalization.

[0042] S200. Construct a causal network based on historical risk information and prevention and control measures. Input construction data and risk factors into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures; specifically including:

[0043] S210. Construct risk factor nodes, trigger nodes, and prevention and control measure nodes based on historical risk information and prevention and control measures;

[0044] Risk factors include water leakage and land subsidence;

[0045] Triggering factors include abnormal earth pressure and abnormal water pressure;

[0046] Control measures include adjusting the progress speed and grouting pressure.

[0047] S220. By connecting nodes with causal relationships, directed edges are obtained;

[0048] For example, if abnormal earth pressure can lead to surface settlement, then a directed edge is constructed from the earth pressure anomaly node to the surface settlement node; if adjusting the grouting pressure can improve surface settlement, then a directed edge is constructed from the grouting pressure adjustment node to the surface settlement node.

[0049] S230. Obtain the weight of each directed edge based on the degree of influence between nodes, and obtain the delay step size for weight update based on node type.

[0050] The weight of the directed edge can be defined based on expert knowledge, such as directly defining the weight of "earth pressure anomaly → surface settlement" as 0.7;

[0051] It can also be defined by analyzing and statistically analyzing historical data. For example, if there are 10 earth pressure anomalies in history, and 6 of them caused surface subsidence, then the causal weight between the two can be defined by the ratio of the number of times they occurred, which is 0.6.

[0052] In addition, the influence between some nodes may be delayed, so it is necessary to assign a delay step size to some causal edges (such as "grouting pressure adjustment → settlement improvement" delayed by 3 time steps) and realize historical state memory through a buffer queue.

[0053] S240. Construct the activation probability function for each risk factor node based on the risk coefficient, edge weight, and delay step size to obtain the causal network.

[0054] The activation probability function is as follows:

[0055] ;

[0056] in, Let i be the activation probability of the risk factor node i at step t+1. The activation probability function is, for example, Sigmoid; j represents the relevant nodes involved in the update that connect to risk factor node i, and n is the total number of relevant nodes; Let be the edge weight from node j to node i; Let e ​​be the state characteristic of node j at step t; e is the risk coefficient. The risk coefficient weight.

[0057] in It is obtained by mapping the current parameters of the node. For example, for the "abnormal earth pressure" causative node, the current earth pressure parameter is directly normalized as the node state feature, or the difference between the abnormal earth pressure parameter and the normal earth pressure parameter is used as the node state feature.

[0058] After constructing the causal network, the construction data and risk factors are input into the causal network, as follows:

[0059] S250. Input the current construction data and risk factors into the causal network for propagation, update the node status, and obtain the activation probability of the risk factors.

[0060] The state of the risk factor node is jointly determined by the trigger node and the prevention and control measure node, but this step calculates the risk under the condition that no prevention and control measures are implemented, so the prevention and control measure node does not participate in the update;

[0061] S260. Traverse the causal network in reverse from the risk factor nodes to obtain all prevention and control measure nodes that point to the risk factor nodes.

[0062] The causal network contains all the risk factor nodes that have appeared in history, but this step focuses on the risk factors obtained in step S100, and only uses these risk factor nodes to find the associated prevention and control measure nodes.

[0063] S270. Construct an objective function by summing the activation probabilities of all risk factor nodes. Minimize the objective function by adjusting the parameters of the prevention and control measures to obtain recommended prevention and control measures.

[0064] It is necessary to construct the relationship equation between each prevention and control measure and risk factor based on historical data in advance. For example, the increase in grouting pressure has a significant inhibitory effect on leakage, so a relationship equation between the two needs to be constructed.

[0065] There may be multiple prevention and control measures, and these measures may affect the same risk factor. Therefore, the optimization of parameters for multiple prevention and control measures can be achieved by using methods such as gradient descent, heuristic search, and Bayesian optimization to find the optimal combination of parameters.

[0066] Once the parameter optimization reaches the maximum number of iterations or converges, the sum of the activation probabilities of the risk factor nodes reaches its minimum value. The parameters for the prevention and control measures at this point are then obtained and used as recommended prevention and control measures.

[0067] As an alternative implementation method, since the control measure that minimizes the sum of activation probabilities may not be the most suitable measure at present, when optimizing the parameters of the control measures, the top few parameter combinations that minimize the objective function value can be recorded to generate multiple recommended control measures for staff to refer to and select. Subsequently, the risk diffusion trend can be simulated based on multiple recommended control measures, thereby enabling a more intuitive judgment on the advantages and disadvantages of each measure in the current environment.

[0068] S300. Divide the tunnel into a cellular grid, and simulate the risk diffusion process based on risk factors, risk coefficients and activation probabilities to obtain the first risk information;

[0069] This application employs cellular automata to simulate the spatiotemporal evolution of complex systems. It consists of a regularly arranged grid of cells, each of which can be in one of a finite number of discrete states. The state of a cell is updated in discrete time steps according to local rules. The state update of a cell depends only on its own state and the states of its neighbors. This application uses a three-dimensional cellular automata, where the state of each cell is updated by its six neighboring cells (upper, lower, left, right, front, and back).

[0070] This step specifically includes:

[0071] S310. Initialize the state information of each cell according to the risk coefficient and activation probability;

[0072] After obtaining the risk factor in step S100 above, the actual location of the risk factor can be determined based on the sensor location, that is, the cell at the corresponding location is found, and the initial state of the risk cell is updated with the corresponding risk coefficient and activation probability. At the same time, the risk type of each cell is marked, such as water leakage risk and soil loosening risk (the same cell may have multiple risk types at the same time).

[0073] If the cell is not the corresponding risk factor, then initialize the cell's risk coefficient and activation probability to 0;

[0074] In addition to the information mentioned above, the state information of a cell may also include the geological information corresponding to the cell, such as the surrounding rock grade.

[0075] S320. Based on the state information of each cell and its neighboring cells, a state transition function is obtained by constructing a conditional probability model; wherein, the weights of each item in the state transition function are dynamically adjusted according to the construction stage.

[0076] The state transition function is represented as a conditional probability model, where the state of each cell at the next time step is determined by its current state, multi-source risk inputs, and the influence of its neighbors.

[0077] As an optional implementation, the state transition rule can be designed as follows:

[0078] P = α × Risk coefficient + β × Activation probability + γ × Proportion of high-risk neighboring cells + × Risk intensity of neighboring cells;

[0079] P represents the risk intensity of a cell, which needs to be normalized to ensure that P∈[0,1]. The weighting coefficients α, β, γ, and... By fitting historical data or assigning values ​​by experts.

[0080] When updating the risk intensity of a cell, the risk type of its neighboring cells is also updated in the cell's information set.

[0081] When the P-value is greater than a preset threshold, the cell is identified as a high-risk cell. This threshold can be adjusted according to different geological types, such as setting different thresholds for sand layers and rock layers; or setting different thresholds for different surrounding rock grades.

[0082] Additionally, the weights of input elements can be adjusted according to the construction stage, for example:

[0083] In the early stages of tunneling: the focus is on geological risk factors, therefore the risk coefficient weight and / or activation probability weight of geological risk factors can be increased;

[0084] Alarm period: Primarily driven by activation probability, the weight of activation probability can be increased.

[0085] When cells with different types of risks meet during diffusion, the risk coefficient weight and activation probability weight can be adjusted accordingly.

[0086] S330. Determine the initial risk cell based on the region where the risk factor is located. Starting from the initial risk cell, update the state of adjacent cells sequentially to simulate the risk diffusion process and obtain the risk cell.

[0087] When determining the initial risk cell based on risk factors, the scope of the risk can be determined according to the corresponding risk coefficient. That is, the initial risk cell can be a single cell or a cluster of cells within a region. Alternatively, when there are many initial risk cells, they can be sorted in descending order according to their risk coefficients, and the top few cells in the sorted order can be selected as the starting point for updates. This method is suitable for simulating environmental risks such as water seepage and land subsidence.

[0088] When there are risk factors near the cutterhead of the tunnel boring machine, the cell corresponding to the current position of the cutterhead can be used as the starting point to simulate the risk diffusion process along the axial direction (excavation direction) or radial direction (such as seepage path). This is suitable for simulating equipment-related risks such as excavation face instability and cutter wear.

[0089] Ultimately, we can obtain the risk cell after risk diffusion.

[0090] S400: Based on the causal network, obtain the minimum activation probability of the risk factor after the implementation of prevention and control measures, simulate the risk diffusion process based on the minimum activation probability, and obtain the second risk information;

[0091] In step S270, recommended prevention and control measures have been obtained, and the activation probability of each risk factor can be obtained. Substituting the activation probability into the state transition rule, the risk cell after the implementation of the prevention and control measures can be obtained.

[0092] S500 recommends prevention and control measures by comparing the first risk information and the second risk information.

[0093] The first and second risk information refer to the status of risk cells before and after the implementation of prevention and control measures, including information such as the distribution and number of risk cells. From this, the percentage reduction in the number of risk cells after the implementation of prevention and control measures can be calculated, for example, a reduction of 57%, thereby determining whether the recommended prevention and control measures are effective.

[0094] Furthermore, the risk cell distribution can also be used to observe whether the tunnel risk will affect surrounding buildings or facilities, thus enabling timely implementation of corresponding prevention and control measures.

[0095] Finally, to visually demonstrate the risk diffusion process, this method also includes:

[0096] Construct a three-dimensional model along the tunnel, which includes a tunnel model and models of buildings and geographical landmarks within a preset range near the tunnel;

[0097] The cellular mesh is mapped onto a 3D model along the tunnel, and risk cells are displayed in different colors compared to other cells to obtain a risk warning model. Different types of risk cells can be displayed in different colors to allow technicians to accurately understand risk information.

[0098] Example 2

[0099] This embodiment provides a safety risk prevention and control system for shield tunnel construction, including:

[0100] The extraction module 100 is used to acquire construction data, extract risk factors from the construction data through persistent coherence technology, and obtain the corresponding risk coefficients.

[0101] Module 200 is used to construct a causal network based on historical risk information and prevention and control measures. Construction data and risk factors are input into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures.

[0102] The first simulation module 300 is used to divide the tunnel into cellular grids and simulate the risk diffusion process based on risk factors, risk coefficients and activation probabilities to obtain the first risk information.

[0103] The second simulation module 400 is used to obtain the minimum activation probability of risk factors after the implementation of prevention and control measures based on the causal network, and to simulate the risk diffusion process based on the minimum activation probability to obtain the second risk information.

[0104] The recommendation module 500 is used to recommend prevention and control measures by comparing the first risk information and the second risk information.

[0105] As an optional implementation, the extraction module includes:

[0106] The first mapping unit 110 is used to map the construction data into a point cloud structure or a graph structure.

[0107] Analysis unit 120 is used to obtain topological features and the lifetime of topological features by analyzing the evolution of the data topology structure through filtered complex analysis based on the mapped construction data.

[0108] The second mapping unit 130 is used to map topological features to risk factors based on a preset lookup table, and at the same time map their lifetimes to risk coefficients.

[0109] As an optional implementation, the building module 200 includes:

[0110] The first construction unit 210 is used to construct risk factor nodes, cause nodes, and prevention and control measure nodes based on historical risk information and prevention and control measures.

[0111] The second building unit 220 is used to obtain directed edges by connecting nodes with causal relationships;

[0112] The third building unit 230 is used to obtain the weight of each directed edge based on the degree of influence between nodes, and to obtain the delay step size for weight update based on the node type.

[0113] The fourth building unit 240 is used to construct the activation probability function of each risk factor node based on the risk coefficient, edge weight and delay step size, so as to obtain the causal network.

[0114] As an optional implementation, the building module 200 further includes:

[0115] Update unit 250 is used to input the current construction data and risk factors into the causal network for propagation, update the node status, and obtain the activation probability of the risk factors;

[0116] Traversal unit 260 is used to traverse the causal network in reverse from the risk factor node to obtain all prevention and control measure nodes that point to the risk factor node;

[0117] The optimization unit 270 is used to construct an objective function by summing the activation probabilities of all risk factor nodes, and to minimize the objective function by adjusting the parameters of the prevention and control measures to obtain recommended prevention and control measures.

[0118] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0119] 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 variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for preventing and controlling safety risks in shield tunnel construction, characterized in that, include: Acquire construction data, extract risk factors from the construction data using persistent coherence technology, and obtain the corresponding risk coefficients; A causal network is constructed based on historical risk information and prevention and control measures. Construction data and risk factors are input into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures, including: The current construction data and risk factors are input into the causal network for propagation, the node status is updated, and the activation probability of the risk factors is obtained. Traverse the causal network backward from the risk factor nodes to obtain all prevention and control measure nodes that point to the risk factor nodes; The objective function is constructed by summing the activation probabilities of all risk factor nodes. The recommended prevention and control measures are obtained by minimizing the objective function by adjusting the parameters of the prevention and control measures. The tunnel route is divided into cellular grids, and the risk diffusion process is simulated based on risk factors, risk coefficients and activation probabilities to obtain the first risk information. The lowest activation probability of risk factors after the implementation of prevention and control measures is obtained based on the causal network. The risk diffusion process is simulated based on the lowest activation probability to obtain the second risk information. Prevention and control measures are recommended by comparing the first and second risk information.

2. The shield tunneling safety risk prevention and control method according to claim 1, characterized in that, Risk factors are extracted from construction data using persistent coherence techniques, and corresponding risk coefficients are obtained, including: The construction data is mapped to a point cloud structure or a graph structure; Based on the mapped construction data, the evolution of the data topology is analyzed by filtering complexes to obtain the topological features and the lifetime of the topological features; Based on a pre-defined lookup table, topological features are mapped to risk factors, and their lifetimes are mapped to risk coefficients.

3. The shield tunneling safety risk prevention and control method of claim 1, wherein, The historical risk information includes historical risk factors and risk triggers. The construction of a causal network based on historical risk information and prevention and control measures includes: Construct risk factor nodes, trigger nodes, and prevention and control measure nodes based on historical risk information and prevention and control measures; By connecting nodes with causal relationships, directed edges are obtained; The weight of each directed edge is obtained based on the degree of influence between nodes, and the delay step size for weight update is obtained based on the node type. The activation probability function of each risk factor node is constructed based on the risk coefficient, edge weight, and delay step size to obtain the causal network.

4. The shield tunneling safety risk prevention and control method of claim 1, wherein, The tunnel route is divided into cellular grids. Based on the output of persistent cohomology and dynamic cognitive networks, cellular state transition functions are constructed to simulate the risk diffusion process, yielding the first risk information, including: Initialize the state information of each cell based on the risk coefficient and activation probability; Based on the state information of each cell and its neighboring cells, a state transition function is obtained by constructing a conditional probability model; wherein, the weights of each item in the state transition function are dynamically adjusted according to the construction stage. The initial risk cell is determined by the region where the risk factor is located. Starting from the initial risk cell, the state of adjacent cells is updated sequentially to simulate the risk diffusion process and obtain the risk cell.

5. The shield tunneling safety risk prevention and control method according to claim 4, characterized in that, The method further includes: Construct a three-dimensional model along the tunnel, which includes a tunnel model and models of buildings and geographical landmarks within a preset range near the tunnel; The cell mesh is mapped onto the 3D model along the tunnel, and the risk cells are displayed in different colors from other cells to obtain a risk warning model.

6. A shield tunneling construction safety risk prevention and control system, characterized in that, include: The extraction module is used to acquire construction data, extract risk factors from the construction data through persistent cohomology technology, and obtain the corresponding risk coefficients; The module is used to construct a causal network based on historical risk information and prevention and control measures. Construction data and risk factors are input into the causal network to obtain the activation probability of risk factors and corresponding prevention and control measures. The building module includes: The update unit is used to input the current construction data and risk factors into the causal network for propagation, update the node state, and obtain the activation probability of the risk factors. The traversal unit is used to traverse the causal network in reverse from the risk factor node to obtain all prevention and control measure nodes that point to the risk factor node; The optimization unit is used to construct an objective function by summing the activation probabilities of all risk factor nodes. By adjusting the parameters of the prevention and control measures, the objective function is minimized to obtain the recommended prevention and control measures. The first simulation module is used to divide the tunnel into cellular grids and simulate the risk diffusion process based on risk factors, risk coefficients and activation probabilities to obtain the first risk information. The second simulation module is used to obtain the minimum activation probability of risk factors after the implementation of prevention and control measures based on the causal network, and to simulate the risk diffusion process based on the minimum activation probability to obtain the second risk information. The recommendation module is used to recommend prevention and control measures by comparing the first risk information and the second risk information.

7. The shield tunneling safety risk prevention and control system according to claim 6, characterized in that, The extraction module includes: The first mapping unit is used to map the construction data into a point cloud structure or a graph structure. The analysis unit is used to analyze the evolution of the data topology structure based on the mapped construction data through filtered complex analysis to obtain the topological features and the lifetime of the topological features; The second mapping unit is used to map topological features to risk factors based on a preset lookup table, and at the same time map their lifetimes to risk coefficients.

8. The shield tunneling safety risk prevention and control system according to claim 6, characterized in that, The building module includes: The first building unit is used to construct risk factor nodes, cause nodes, and prevention and control measure nodes based on historical risk information and prevention and control measures. The second building unit is used to obtain directed edges by connecting nodes with causal relationships; The third building unit is used to obtain the weight of each directed edge based on the degree of influence between nodes, and to obtain the delay step size for weight updates based on the node type. The fourth building unit is used to construct the activation probability function of each risk factor node based on the risk coefficient, edge weight, and delay step size, thus obtaining the causal network.

Images