Graded early warning management system for risk-causing factor risk of tunnel group

Abstract

The invention discloses a graded early warning management system for risk-causing factor risks of a tunnel group, and relates to the technical field of tunnel safety monitoring and risk assessment, and the system comprises the following modules: a data collection module which collects risk-causing factor data of each tunnel in real time through a sensor network deployed in the tunnel group, the risk-causing factor data comprises tunnel structure risk-causing factor data, traffic safety risk-causing factor data, operation management risk-causing factor data and surrounding environment risk-causing factor data. According to the graded early warning management system for the risk-causing factor risk of the tunnel group, by introducing global risk analysis and graded early warning logic, the reasonable allocation of emergency resources and the consistency of a global safety target are ensured, and meanwhile, the intelligence and the coping capacity of the whole system are improved by combining advanced machine learning and a graph theory model; the problem that a local response strategy is disjointed from global risk evolution is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel safety monitoring and risk assessment, and particularly to a hierarchical early warning management system for the risks of risk-causing factors in a tunnel group. Background Art

[0002] In the field of tunnel group safety risk early warning, existing technologies mostly rely on single-tunnel independent analysis models. By deploying a sensor network, data on risk-causing factors such as tunnel structure risk-causing factors (lining cracks, lining deformation displacement, lining deterioration, back voids, water seepage, etc.), traffic safety risk-causing factors (traffic volume, tunnel alignment, traffic safety facilities, driver factors, vehicle factors, etc.), operation and management risk-causing factors (ventilation, lighting, monitoring, power supply and distribution, fire detection and alarm, fire protection, risk control measures, etc.), and surrounding environment risk-causing factors (falling rocks, bad weather, geological disasters, road landscape, etc.) of a single tunnel are monitored, and static thresholds or probability statistical methods are used to evaluate their risk levels. However, existing early warning strategies often trigger response measures based on the highest risk level of a single tunnel, but ignore the overall risk entropy change of the tunnel group. When implementing plugging and emergency rescue for high-risk tunnels, it may change the traffic load distribution of surrounding tunnels and induce secondary risks. Due to the lack of a global analysis of the tunnel group risk propagation network, existing systems are prone to falling into the "local optimal solution trap", resulting in defects such as misallocation of emergency resources and decline in overall safety. Summary of the Invention

[0003] (I) Technical Problems to be Solved

[0004] In view of the deficiencies of the existing technology, the present invention provides a hierarchical early warning management system for the risks of risk-causing factors in a tunnel group, which solves the problem of the disconnection between local response strategies and global risk evolution.

[0005] (II) Technical Solutions

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A hierarchical early warning management system for the risks of risk-causing factors in a tunnel group includes the following modules:

[0007] A data acquisition module that real-time collects data on risk-causing factors of each tunnel through a sensor network deployed in the tunnel group. The data on risk-causing factors includes tunnel structure risk-causing factors (lining cracks, lining deformation displacement, lining deterioration, back voids, water seepage, etc.), traffic safety risk-causing factors (traffic volume, tunnel alignment, traffic safety facilities, driver factors, vehicle factors, etc.), operation and management risk-causing factors (ventilation, lighting, monitoring, power supply and distribution, fire detection and alarm, fire protection, risk control measures, etc.), and surrounding environment risk-causing factors (falling rocks, bad weather, geological disasters, road landscape, etc.);

[0008] A risk assessment module that, based on the collected data on risk-causing factors, uses a multi-dimensional assessment model to calculate the safety risk level of each tunnel and dynamically adjusts risk assessment parameters according to the overall safety condition of the tunnel group;

[0009] The global risk analysis module, based on the global safety status of the tunnel group, simulates the risk propagation path and the change of risk entropy, evaluates the mutual influence among tunnels, and generates the global risk change trend;

[0010] The hierarchical early warning module, according to the output of the global risk analysis module, automatically generates a hierarchical early warning response strategy through hierarchical logic, and reasonably allocates emergency resources according to the overall safety condition of the tunnel group;

[0011] The response decision-making module, according to the early warning level of the hierarchical early warning module and the emergency management goal, optimizes the allocation plan of emergency resources, adjusts the response strategy, and ensures the rational use of emergency resources and the reduction of risks.

[0012] Preferably, the risk assessment module uses machine learning algorithms to train and predict various types of data of the tunnel group. Combining historical accident records and sensor data, it generates a safety risk model of the tunnel group. Among them, various types of data in the tunnel group are collected in real time through a sensor network. In addition to real-time sensor data, historical accident records are also collected and organized into a standardized data format, including accident type, occurrence time, accident location, and influence scope. All the collected data needs to go through the data preprocessing module for cleaning, removing noise data, handling missing values, and unifying the data format and normalizing before entering the risk assessment module;

[0013] The risk assessment module first selects key features affecting tunnel safety from the processed data set, including tunnel structure risk factor data, traffic safety risk factor data, operation management risk factor data, and surrounding environment risk factor data. Combining historical accident records and real-time sensor data, it uses machine learning algorithms such as decision trees, random forests, and support vector machines to train the safety risk of the tunnel group. Among them, through historical accident data, the model can learn the influence of different risk factors on the safety risk of the tunnel group; the trained machine learning model will be able to predict the risk level of the tunnel group based on the real-time collected data. This kind of prediction model can identify potential high-risk areas or hidden dangers and generate a risk assessment report according to different risk levels;

[0014] According to the trained machine learning model, the risk assessment module can generate a safety risk model of the tunnel group. This model can be updated regularly and adjust risk assessment parameters according to real-time data. The generated risk model not only includes the risk assessment results of each individual tunnel, but also takes into account the overall risk of the tunnel group, simulates the mutual influence among tunnels and their evolution under different risk scenarios. Finally, the model will be used as the basis for subsequent global risk analysis, hierarchical early warning, and emergency decision-making.

[0015] Preferably, the global risk analysis module constructs a risk propagation network for the tunnel group using a graph theory model to comprehensively analyze the mutual influence of risks within the tunnel group and dynamically adjust the risk assessment criteria for the tunnel group. During the use of the global risk analysis module, first, the real-time data and historical data from each tunnel are integrated. These data include tunnel structure risk factors data, traffic safety risk factors data, operation management risk factors data, and surrounding environment risk factors data. Among them, the tunnel structure risk factors include parameters such as lining cracks, lining deformation displacement, lining deterioration, voids behind the lining, leakage, etc.; the traffic safety risk factors include parameters such as traffic volume, tunnel alignment, traffic safety facilities, driver factors, vehicle factors, etc.; the operation management risk factors include parameters such as ventilation, lighting, monitoring, power supply and distribution, fire detection and alarm, fire protection, risk control measures, etc.; the surrounding environment risk factors include parameters such as rockfall, bad weather, geological disasters, road landscape, etc. At the same time, it also includes the layout information of the tunnel group, the connection relationship between tunnels, and historical accident data. The integrated data is cleaned, denoised, and standardized to ensure the consistency and accuracy of the data. This step ensures that the subsequent analysis can be carried out on the basis of clear and controllable data;

[0016] Based on the layout of the tunnel group and the connection relationship between tunnels, a risk propagation network for the tunnel group is constructed using a graph theory model. Each tunnel is regarded as a node, and the connections between tunnels (such as traffic flow or potential risk transfer paths) are regarded as edges. The weight of each edge can be set according to the possibility and influence of risk propagation. A risk factor based on sensor data and historical accident records is assigned to each tunnel node, and the weight of each connection edge takes into account the risk transfer probability between different tunnels, including traffic flow changes and the conduction effect of structural damage. The shortest path algorithm or diffusion model in graph theory is used to simulate the risk propagation path in the tunnel group. Through simulation, the mutual influence between different tunnels is evaluated. As the risk propagation path changes and the entropy fluctuates, the global risk analysis module will dynamically adjust the risk assessment criteria for the tunnel group. Specifically, when the risk state of a certain tunnel affects the surrounding tunnels, the system will update the model parameters based on the new risk assessment information, thereby affecting the subsequent risk assessment results. After each assessment, the generated global risk change trend will be transmitted as feedback to the hierarchical early warning module and the response decision-making module to ensure that the global risk evolution is accurately reflected in the early warning logic and emergency resource allocation strategy;

[0017] According to the above steps, the global risk analysis module generates a global risk change trend graph of the tunnel group in real time. This trend graph will show the risk states of each tunnel and their contributions to the risk level of the entire tunnel group. This trend will provide a scientific basis for subsequent early warning strategies and emergency management.

[0018] Preferably, based on the global risk change trend, the hierarchical early warning module adopts a hierarchical early warning model to issue different levels of early warnings for tunnels with different risk levels, and adjusts corresponding response measures according to the early warning levels. During the use of the hierarchical early warning module, first, it receives the output data from the global risk analysis module. The hierarchical early warning module integrates the risk levels and entropy change data of each tunnel provided by the global risk analysis module to provide a basis for subsequent early warning and emergency resource allocation. According to the global risk change trend, the hierarchical early warning module evaluates the overall safety status of the tunnel group through a preset hierarchical early warning model, and the early warning model generates corresponding early warning response strategies according to different risk levels;

[0019] When the risk of a tunnel is in the low-risk area, the early warning module may only issue routine monitoring prompts or status reports and does not take emergency response measures; when the risk of a tunnel is in the medium-risk area, the system will trigger an alarm and recommend regular inspections or targeted reinforcement work to ensure that the risk will not increase further; when the risk of a tunnel reaches the high-risk area, the system will automatically trigger an emergency early warning and recommend immediate measures, such as closing the tunnel, dispatching a rescue team or implementing traffic control;

[0020] According to the hierarchical early warning model, the hierarchical early warning module automatically analyzes the risk status of each tunnel and determines its belonging early warning level; according to the risk level and early warning response strategy determined by the hierarchical early warning module, the system will intelligently allocate emergency resources, including: in the low-risk situation, the emergency resources maintain routine allocation without special mobilization, in the medium-risk situation, the system may dispatch some emergency resources for inspection, evaluation or advance deployment of emergency personnel, in the high-risk situation, the system will give priority to mobilizing emergency resources, including rescue teams, traffic control equipment, monitoring equipment, and reasonably arrange the priority and location of resources according to the regional distribution of risks. The hierarchical early warning module ensures that emergency resources are not overly concentrated or wasted by analyzing the global risk situation and the actual needs of each tunnel in the tunnel group, optimizes the use of resources, and avoids the "local optimal solution trap";

[0021] The hierarchical early warning module will generate an allocation plan for emergency resources and output a report to the emergency management system to ensure the smooth execution of emergency response measures. When the safety status of the tunnel group changes, the hierarchical early warning module will continuously update the allocation plan for emergency resources to ensure that the resources are always matched with the risk status;

[0022] After the emergency response measures start to be implemented, the hierarchical early warning module will monitor its effect in real time and make adjustments through a feedback mechanism. If the original strategy fails to effectively mitigate the risk or generates new risks, the module will automatically adjust the response measures and optimize the allocation of emergency resources again.

[0023] Preferably, the response decision-making module adjusts the priority and allocation plan of emergency resources by monitoring the safety status of each tunnel in the tunnel group in real time, avoiding the trap of local optimal solutions, and ensuring the consistency between emergency response and the overall risk assessment goal. The response decision-making module first receives the warning levels from the hierarchical warning module, which are based on the overall safety status of the tunnel group and the risk status (low risk, medium risk, high risk) of each tunnel. In addition, it also receives the emergency management goals, that is, to ensure the overall safety of the tunnel group, the reasonable allocation of emergency resources, and the minimization of losses caused by risks. The response decision-making module analyzes the relationship between the warning levels and the emergency management goals, determines the priority of the emergency response, and determines the specific emergency measures to be taken for each tunnel or area;

[0024] The response decision-making module evaluates the emergency resources by integrating the real-time data, warning levels, and historical data of the tunnel group, including maintenance teams, emergency equipment, and traffic control. In the case of low risk, the system will give priority to ensuring regular monitoring and daily maintenance to ensure that resources are not over-consumed; when the risk reaches a medium level, the system will allocate a certain amount of emergency resources for inspection and regular assessment to prevent the risk from escalating further; when the risk of a tunnel reaches a high risk level, the response decision-making module will immediately mobilize a large number of emergency resources, including rescue teams, traffic control personnel, and key equipment, to quickly respond to potential emergencies; according to the changes in the overall risk assessment, the response decision-making module will adjust the response strategy in real time. Based on the risk level of each tunnel, the availability of emergency resources, and the overall safety status, the response decision-making module will generate a reasonable allocation plan of emergency resources in real time. The system takes into account the timeliness and spatial distribution of resources to ensure that emergency resources can reach the required places in the shortest time, improving the efficiency and effectiveness of emergency response. During the emergency response process, the system will monitor the implementation effect of each emergency measure. If some measures fail to effectively reduce the risk, the system will automatically adjust the resource allocation plan to optimize the emergency response process;

[0025] The response decision-making module generates a resource allocation report according to the optimized emergency resource allocation plan and outputs it to relevant emergency management personnel to ensure the implementation of resource allocation. The report details the allocation of emergency resources, priorities, and the specific implementation schedule and action plan. As the safety status of the tunnel group changes, the response decision-making module will continuously adjust the resource allocation plan to ensure that the emergency measures are always matched with the overall safety status of the tunnel group.

[0026] Preferably, the emergency resources include maintenance teams, traffic control resources, and emergency equipment, which are automatically allocated by the system to deal with tunnels of different risk levels.

[0027] A method for hierarchical warning management of risk factors for tunnel groups includes the following steps:

[0028] S1. Collect the risk factor data of each tunnel in the tunnel group in real time, and clean and preprocess the data through the data processing module;

[0029] S2. Use the risk assessment module to calculate the risk level of each tunnel according to the preprocessed data, and dynamically adjust the assessment criteria;

[0030] S3. Analyze the risk propagation network of the tunnel group based on the global risk analysis module, simulate the risk evolution process, and calculate the mutual influence between tunnels;

[0031] S4. According to the global risk change trend, use the hierarchical warning module to generate a hierarchical response warning strategy and generate corresponding emergency management objectives;

[0032] S5. On the basis of hierarchical warning, optimize the emergency resource allocation plan through the response decision module and implement targeted risk management measures.

[0033] Preferably, in step S2, a machine learning algorithm is used to optimize the risk assessment model, and it is trained and verified in combination with the historical data in the tunnel group.

[0034] Preferably, in step S3, a risk propagation network of the tunnel group is constructed, and a graph theory analysis model is used to calculate the risk propagation path.

[0035] Preferably, in step S4, the warning level is optimized according to the global safety condition of the tunnel group through the hierarchical warning strategy, and the optimal allocation of resources is achieved by adjusting the response measures.

[0036] (III) Beneficial effects

[0037] The present invention provides a hierarchical warning management system for the risk of risk factors in a tunnel group. It has the following beneficial effects:

[0038] This hierarchical warning management system for the risk of risk factors in a tunnel group ensures the reasonable allocation of emergency resources and the consistency of global safety objectives by introducing global risk analysis and hierarchical warning logic. At the same time, it combines advanced machine learning and graph theory models to improve the intelligence and response ability of the overall system, and solves the problem of the disconnection between local response strategies and global risk evolution. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a structural schematic diagram / framework schematic diagram of the whole of the present invention;

[0040] Figure 2 It is a flowchart of a method for hierarchical warning management of the risk of risk factors in a tunnel group of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0041] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0042] Please refer to Figure 1 and Figure 2 , the present invention provides a technical solution: a hierarchical early warning management system for the risk of risk factors in a tunnel group, including the following modules:

[0043] The data acquisition module collects the risk factor data of each tunnel in real time through the sensor network deployed in the tunnel group. The risk factor data includes tunnel structure risk factor data, traffic safety risk factor data, operation management risk factor data, and surrounding environment risk factor data;

[0044] The risk assessment module, based on the collected risk factor data, uses a multi-dimensional assessment model to calculate the safety risk level of each tunnel, and dynamically adjusts the risk assessment parameters according to the overall safety status of the tunnel group;

[0045] The global risk analysis module, based on the global safety status of the tunnel group, simulates the risk propagation path and the change of risk entropy, evaluates the mutual influence between tunnels, and generates the global risk change trend;

[0046] The hierarchical early warning module, according to the output of the global risk analysis module, automatically generates a hierarchical early warning response strategy through hierarchical logic, and reasonably allocates emergency resources according to the overall safety status of the tunnel group;

[0047] The response decision module, according to the early warning level of the hierarchical early warning module and the emergency management goal, optimizes the allocation plan of emergency resources, adjusts the response strategy, and ensures the rational use of emergency resources and the reduction of risks.

[0048] Among them, the sensor network is composed of various types of sensors and communication devices, mainly including sensor nodes, data transmission and communication devices, data processing and storage systems, edge computing devices, and cloud computing platforms;

[0049] The sensor nodes are various sensors deployed inside and outside the tunnel, responsible for collecting real-time data of the tunnel; different types of sensors are deployed according to the types and functions of the collected data; the main sensor types include structural health sensors, environmental monitoring sensors, traffic monitoring sensors, leakage monitoring sensors, and vibration sensors;

[0050] Structural health monitoring sensors include crack sensors, deformation sensors, displacement sensors, and stress sensors, which are used to monitor the safety of tunnel structures in real time and detect whether there are cracks, deformations, displacements, or other structural problems; environmental monitoring sensors include temperature and humidity sensors, air quality sensors, and gas concentration sensors, which are used to detect the environmental conditions inside the tunnel, such as temperature, humidity, gas concentration, etc., in order to timely discover potential hazards, such as harmful gas leakage and excessive humidity; traffic monitoring sensors include vehicle counters, speed sensors, and traffic flow sensors, which are used to monitor information such as traffic flow, vehicle speed, and vehicle type inside the tunnel; this data helps to evaluate the impact of traffic load on the tunnel structure; leakage and water seepage monitoring sensors include water level sensors, humidity sensors, and leakage detection sensors, which are used to monitor whether there is water leakage or water accumulation inside the tunnel, in order to take measures to prevent the impact of water damage on the tunnel; vibration sensors are used to monitor the vibration conditions inside and outside the tunnel, especially when the tunnel passes through earthquake-prone areas or mining areas where vibrations are frequent, and detect the impact of underground vibration changes on tunnel safety;

[0051] Data transmission and communication equipment enables each sensor to transmit the collected data to the data centralized management system or real-time monitoring system through wireless or wired networks; these communication devices include wireless transmission equipment, wired networks, and data aggregation nodes;

[0052] Wireless transmission equipment includes Wi-Fi, ZigBee, LoRaWAN (Long Range Wide Area Network with Low Power Consumption), etc., which is suitable for transmissions that need to cover a large range and is suitable for wide deployment within a tunnel group; the wired network is fiber optic communication or Ethernet communication, which is used for stable connections between the data center and each sensor node; the data aggregation node is used to collect data from multiple sensor nodes and transmit it to the data processing center;

[0053] The data processing and storage system enables the data collected by the sensors to be transmitted to the data center through transmission equipment, and the data is aggregated, stored, preprocessed, and analyzed by the data processing and storage system; this system usually includes edge computing devices and cloud computing platforms;

[0054] Edge computing devices are used to perform preliminary processing and analysis of data near the sensors, reduce the burden on network bandwidth, and can respond to emergencies faster; the cloud computing platform is used to store a large amount of historical data and execute complex analysis tasks, such as training machine learning models and risk prediction;

[0055] Structural health monitoring sensors are arranged at key load-bearing structural positions of the tunnel, including the vault, support beams, and joints, with a focus on detecting areas prone to cracks or deformations; environmental monitoring sensors are arranged at different positions in the tunnel, especially where there is potential gas leakage or high humidity, and appropriate monitoring points are set according to the ventilation conditions of the tunnel; traffic monitoring sensors are arranged at the tunnel entrances and exits, key sections, etc., where the traffic flow is large, to record traffic flow and vehicle speed data in real time; leakage and water seepage monitoring sensors are arranged in areas where the tunnel is prone to water accumulation, including the tunnel bottom and water accumulation-prone areas; vibration sensors should be arranged on the ground around the tunnel and at key positions inside and outside the tunnel to monitor underground vibrations;

[0056] Sensor nodes regularly collect data and transmit it to the centralized data management system or real-time monitoring platform; the data collection frequency can be flexibly adjusted according to the safety status and risk assessment needs of the tunnel, including increasing the data collection frequency for tunnels with high risks; the raw data transmitted by the sensor network will go through a data preprocessing module for operations such as denoising, missing value handling, and standardization to ensure the accuracy and usability of the data; when the data collected by the sensors exceeds the preset safety threshold or abnormal conditions occur, the system will automatically trigger an alarm mechanism to notify the tunnel management personnel to take corresponding emergency measures.

[0057] The risk assessment module uses machine learning algorithms to train and predict various types of data of the tunnel group. Combining historical accident records and sensor data, it generates a safety risk model for the tunnel group. Among them, various types of data in the tunnel group are collected in real time through the sensor network. In addition to real-time sensor data, historical accident records are also collected and organized into a standardized data format, including accident type, occurrence time, accident location, and impact range. All the collected data needs to go through a data preprocessing module for cleaning, removing noise data, handling missing values, and unifying the data format and normalization before entering the risk assessment module;

[0058] The risk assessment module first selects key features affecting tunnel safety from the processed dataset, including tunnel structure risk factors data, traffic safety risk factors data, operation management risk factors data, and surrounding environment risk factors data. Combining historical accident records and real-time sensor data, it uses decision trees, random forests, and support vector machines in machine learning algorithms to train the safety risks of the tunnel group. Among them, through historical accident data, the model can learn the impact of different risk factors on the safety risks of the tunnel group; the trained machine learning model will be able to predict the risk level of the tunnel group based on the real-time collected data. This prediction model can identify potential high-risk areas or hidden dangers and generate a risk assessment report according to different risk levels, where the risk levels include low risk, medium risk, and high risk;

[0059] Based on the trained machine learning model, the risk assessment module can generate a safety risk model for the tunnel group. This model can be updated regularly and adjust risk assessment parameters according to real-time data. The generated risk model not only includes the risk assessment results of each individual tunnel but also takes into account the overall risk of the tunnel group, simulates the mutual influence between tunnels and their evolution under different risk scenarios. Finally, the model will serve as the basis for subsequent global risk analysis, hierarchical early warning, and emergency decision-making. As new sensor data and accident records are continuously updated, the machine learning algorithm will retrain the risk model regularly to ensure its adaptability and predictive ability in a dynamic environment. The update feedback mechanism of the risk model ensures that the emergency management system can respond to the risk changes of the tunnel group in real time and adjust the hierarchical early warning strategy and emergency resource allocation in a timely manner.

[0060] Through the above steps, the risk assessment module uses machine learning algorithms to identify valuable patterns in various data of the tunnel group and generates a safety risk model for the tunnel group that can continuously learn and update, thereby providing a scientific and accurate basis for subsequent risk assessment, early warning, and emergency decision-making.

[0061] The global risk analysis module uses a graph theory model to construct a risk propagation network for the tunnel group to comprehensively analyze the mutual influence of risks within the tunnel group and dynamically adjust the risk assessment criteria for the tunnel group. During the use of the global risk analysis module, first, integrate the real-time data and historical data from each tunnel. These data include tunnel structure risk factors data, traffic safety risk factors data, operation management risk factors data, and surrounding environment risk factors data. At the same time, it also includes the layout information of the tunnel group, the connection relationship between tunnels, and historical accident data. Clean, denoise, and standardize the integrated data to ensure the consistency and accuracy of the data. This step ensures that the subsequent analysis can be carried out on the basis of clear and controllable data.

[0062] Based on the layout of the tunnel group and the connection relationship between tunnels, use a graph theory model to construct a risk propagation network for the tunnel group. Each tunnel is regarded as a node, and the connection between tunnels is regarded as an edge, such as traffic flow or potential risk transmission path. The weight of each edge can be set according to the possibility and influence of risk propagation. Assign a risk factor to each tunnel node based on sensor data and historical accident records, and the weight of each connection edge takes into account the risk transmission probability between different tunnels, including the conduction effect of traffic flow changes and structural damage.

[0063] The shortest path algorithm or diffusion model in graph theory is used to simulate the risk propagation path in the tunnel group. Through simulation, the mutual influence between different tunnels is evaluated. For example, when a high-risk event occurs in one tunnel, how the risk states of the surrounding tunnels change. Based on the risk factors of each node (tunnel) and the propagation probability of the connecting edges, the risk entropy of the tunnel group system is calculated. The risk entropy is used to describe the risk uncertainty of the entire tunnel group. Through the change of entropy, the degree of risk diffusion can be dynamically understood. When the risk entropy increases, it indicates that the overall risk of the system is rising and attention needs to be paid;

[0064] With the change of the risk propagation path and the fluctuation of entropy, the global risk analysis module will dynamically adjust the risk assessment criteria of the tunnel group. Specifically, when the risk state of a certain tunnel affects the surrounding tunnels, the system will update the model parameters based on the new risk assessment information, thus affecting the subsequent risk assessment results. After each assessment, the generated global risk change trend will be transmitted as feedback to the hierarchical early warning module and the response decision-making module to ensure that the global risk evolution is accurately reflected in the early warning logic and the emergency resource allocation strategy;

[0065] According to the above steps, the global risk analysis module generates a global risk change trend graph of the tunnel group in real time. This trend graph will show the risk states of each tunnel and their contributions to the risk level of the entire tunnel group. This trend will provide a scientific basis for the subsequent early warning strategy and emergency management. The global risk change trend will be presented through a visual interface, enabling decision-makers to intuitively see the safety states of each tunnel and the risk propagation process from local to global;

[0066] The global risk analysis module can effectively simulate the risk propagation path in the tunnel group, analyze the mutual influence between tunnels, and dynamically adjust the risk assessment criteria. The application of the graph theory model and the change of risk entropy enables the risk situation of the entire tunnel group to be accurately captured and updated in real time, thus providing strong support for hierarchical early warning and emergency management.

[0067] Based on the global risk change trend, the hierarchical early warning module adopts a hierarchical early warning model to issue different levels of warnings to tunnels with different risk levels and adjust the corresponding response measures according to the warning levels. During the use of the hierarchical early warning module, it first receives the output data from the global risk analysis module, including the global risk change trend of the tunnel group. These data describe the risk states of each tunnel, the mutual influence between tunnels, and the change of the global risk entropy. The hierarchical early warning module integrates the risk level and entropy change data of each tunnel provided by the global risk analysis module to provide a basis for subsequent early warning and emergency resource allocation. The hierarchical early warning module evaluates the overall safety status of the tunnel group according to the global risk change trend through a preset hierarchical early warning model. The early warning model generates corresponding early warning response strategies according to different risk levels;

[0068] Low risk level: When the risk of the tunnel is in the low-risk area, the warning module may only issue routine monitoring prompts or status reports and does not take emergency response measures;

[0069] Medium risk level: When the risk of the tunnel is in the medium-risk area, the system will trigger an alarm and recommend regular inspections or targeted reinforcement work to ensure that the risk does not increase further;

[0070] High risk level: When the risk of the tunnel reaches the high-risk area, the system will automatically trigger an emergency warning and recommend immediate measures, such as closing the tunnel, dispatching rescue teams or implementing traffic control, etc.;

[0071] According to the hierarchical warning model, the hierarchical warning module automatically analyzes the risk status of each tunnel and determines its warning level. The warning level is not only based on the risk status of a single tunnel, but also takes into account the overall risk evolution and mutual influence within the tunnel group to ensure that the generated warning response can reflect the changes in the overall risk. As the risk status of the tunnel group changes, the hierarchical warning module can adjust the warning response strategy in real time. For example, if the risk of a certain tunnel suddenly deteriorates and affects the safety of the surrounding tunnels, the hierarchical warning module will adjust the intensity and content of the warning response according to the new risk assessment results;

[0072] According to the risk level and warning response strategy determined by the hierarchical warning module, the system will intelligently allocate emergency resources, including: in the case of low risk, the emergency resources maintain the normal allocation without special mobilization; in the case of medium risk, the system may dispatch some emergency resources for inspection, assessment or advance deployment of emergency personnel; in the case of high risk, the system will give priority to mobilizing emergency resources, including rescue teams, traffic control equipment, and monitoring equipment, and reasonably arrange the priority and location of resources according to the regional distribution of risks. The hierarchical warning module ensures that the emergency resources are not overly concentrated or wasted by analyzing the overall risk situation and the actual needs of each tunnel in the tunnel group, optimizes the use of resources, and avoids the "local optimal solution trap";

[0073] The hierarchical warning module will generate an allocation plan for emergency resources and output a report to the emergency management system to ensure the smooth implementation of emergency response measures. When the safety status of the tunnel group changes, the hierarchical warning module will continuously update the allocation plan for emergency resources to ensure that the resources are always matched with the risk status;

[0074] After the emergency response measures are implemented, the hierarchical warning module will monitor their effects in real time and make adjustments through the feedback mechanism. If the original strategy fails to effectively mitigate the risk or generates new risks, the module will automatically adjust the response measures and optimize the allocation of emergency resources again;

[0075] Through the above process, after obtaining the output of the global risk analysis module in real time, the hierarchical early warning module can combine the hierarchical early warning model to automatically generate early warning response strategies suitable for different risk levels. At the same time, the module also reasonably allocates emergency resources according to the overall safety status of the tunnel group to ensure the effective utilization of resources and adjusts the early warning response and emergency resource allocation strategies in real time. This technical process ensures that the system can cope with the changing safety risks of the tunnel group through dynamic adjustment and minimizes the possibility of accidents.

[0076] The response decision-making module adjusts the priority and allocation plan of emergency resources by monitoring the safety status of each tunnel in the tunnel group in real time, avoiding the trap of local optimal solutions and ensuring the consistency between emergency response and global risk assessment goals. The response decision-making module first receives the early warning levels from the hierarchical early warning module, which are based on the overall safety status of the tunnel group and the risk status (low risk, medium risk, high risk) of each tunnel. In addition, it also receives the emergency management goals, that is, to ensure the overall safety of the tunnel group, the reasonable allocation of emergency resources, and the minimization of losses caused by risks. The response decision-making module analyzes the relationship between the early warning levels and the emergency management goals, determines the priority of emergency response, and determines the specific emergency measures that need to be taken for each tunnel or area.

[0077] The response decision-making module evaluates emergency resources by integrating the real-time data, early warning levels, and historical data of the tunnel group, including maintenance teams, emergency equipment, and traffic control. In the case of low risks, the system will give priority to ensuring routine monitoring and daily maintenance to prevent excessive consumption of resources. When the risk reaches a medium level, the system will allocate a certain amount of emergency resources for inspection and regular assessment to prevent the risk from escalating further. When the risk of a tunnel reaches a high risk level, the response decision-making module will immediately mobilize a large amount of emergency resources, including rescue teams, traffic control personnel, and key equipment, to quickly respond to potential emergencies.

[0078] According to the changes in the global risk assessment, the response decision-making module will adjust the response strategy in real time. For example, if the traffic load of surrounding tunnels is too large due to emergency response in a certain tunnel, which may affect the safety of other tunnels, the response decision-making module will automatically adjust the response strategy and allocate resources to the new high-risk area. Through the global risk assessment results, the response decision-making module ensures that the safety of all tunnels and areas is comprehensively considered, avoiding the situation where the resources in other areas are insufficient due to excessive concentration of resources in a local high-risk area, falling into the "local optimal solution trap". When formulating emergency response measures, the system not only focuses on the local risk assessment results but also conducts global optimization to ensure that the resources and emergency measures in each area can be coordinated and consistent, avoiding resource waste and uneven distribution.

[0079] Based on the risk level of each tunnel, the availability of emergency resources, and the overall safety situation, the response decision-making module will generate a reasonable allocation plan of emergency resources in real time. Considering the timeliness and spatial distribution of resources, the system ensures that emergency resources can reach the required places in the shortest time, improving the efficiency and effectiveness of emergency response. During the emergency response process, the system will monitor the implementation effects of various emergency measures. If some measures fail to effectively reduce risks, the system will automatically adjust the resource allocation plan and optimize the emergency response process;

[0080] The response decision-making module generates a resource allocation report according to the optimized emergency resource allocation plan and outputs it to relevant emergency management personnel to ensure the implementation of resource allocation. The report details the allocation of emergency resources, priorities, as well as the specific implementation schedule and action plan. As the safety situation of the tunnel group changes, the response decision-making module will continuously adjust the resource allocation plan to ensure that emergency measures are always matched with the overall safety situation of the tunnel group;

[0081] The response decision-making module can coordinate multiple emergency response tasks, ensuring that the execution order of each task and resource allocation can cooperate with each other, avoiding affecting the overall efficiency due to task conflicts or resource reuse. The module will also prioritize the mobilization of key resources according to the risk levels of different tunnels and coordinate the emergency response work in each region in real time to ensure the optimal use of resources and the maximization of overall safety;

[0082] Through the above process, the response decision-making module can automatically optimize the emergency resource allocation plan according to the warning level provided by the hierarchical warning module and the emergency management objectives, adjust the response strategy, and ensure the reasonable use of resources. The global optimization and dynamic adjustment mechanism of the module can avoid the trap of local optimal solutions, ensuring that resources in each region can be evenly and efficiently allocated during the emergency response process, minimizing the overall safety risks of the tunnel group.

[0083] Emergency resources include maintenance teams, traffic control resources, and emergency equipment, which are automatically allocated by the system to respond to tunnels with different risk levels.

[0084] A method for hierarchical warning management of risk factors in a tunnel group includes the following steps:

[0085] S1. Real-time collect the data of risk factors of each tunnel in the tunnel group, and clean and preprocess the data through the data processing module;

[0086] S2. Use the risk assessment module to calculate the risk level of each tunnel according to the preprocessed data and dynamically adjust the assessment criteria;

[0087] S3. Analyze the risk propagation network of the tunnel group based on the global risk analysis module, simulate the risk evolution process, and calculate the mutual influence between tunnels;

[0088] S4. Generate a hierarchical response warning strategy using the hierarchical warning module according to the global risk change trend, and generate corresponding emergency management objectives;

[0089] S5. On the basis of hierarchical warning, optimize the emergency resource allocation plan through the response decision-making module and implement targeted risk management measures.

[0090] In step S2, a machine learning algorithm is used to optimize the risk assessment model, which is trained and verified in combination with the historical data in the tunnel group. It should be further noted that in step S21, collect the historical data in the tunnel group, including but not limited to the structural health monitoring data, traffic flow data, leakage data, and historical accident records of each tunnel; these data can reflect the performance of the tunnel group in different risk states; preprocess the collected historical data, remove noise data, process missing values, and standardize various data formats; this step ensures the consistency and accuracy of the data and provides reliable input data for subsequent model training; according to the safety assessment objectives of the tunnel group, use domain knowledge and statistical analysis methods to extract key features from the historical data; common features may include structural deformation, environmental conditions, traffic flow, and historical accident incidence;

[0091] Step S22. According to the task requirements, select suitable machine learning algorithms, including decision trees, random forests, support vector machines (SVMs), and neural networks; these algorithms can process complex multi-dimensional data and learn the potential patterns that affect the safety of the tunnel group;

[0092] Among them, use supervised learning algorithms to train the model through the labeled historical data, such as known accident data and safety assessment results, and learn the relationship between the input features and the safety risks of the tunnel group; use the historical data to train the model, adjust the algorithm parameters, so that the model can better fit the historical data and accurately predict the safety risks of the tunnel group; use the built-in feature selection mechanism in the machine learning algorithm, that is, the feature importance score in the random forest, to evaluate which features have the greatest impact on the prediction results of the risk assessment model; optimize the feature set to improve the prediction ability of the model;

[0093] Step S23. Divide the historical data into a training set and a validation set, and use cross-validation techniques to improve the generalization ability of the model; through this method, ensure that the model can show good stability on different data subsets; use the validation set to evaluate the trained model and check the performance of the model on unknown data, including prediction accuracy, recall rate, and F1 score; if the performance of the model on the validation set is not ideal, then return to step 2 for parameter adjustment or select other more suitable machine learning algorithms;

[0094] During the training process, pay attention to avoiding overfitting, that is, the model overly relies on the training set, resulting in poor performance on new data, and also pay attention to avoiding underfitting, that is, the model fails to capture important patterns in the data; solve these problems by regularization, adjusting hyperparameters or using more data;

[0095] Step S24. According to the verification results, adjust the hyperparameters in the machine learning algorithm, including the learning rate, the depth of the tree, and the regularization term, to optimize the prediction effect of the model; use the grid search or random search method for parameter tuning; if the effect of a single model is not ideal, then adopt the ensemble learning method, including random forest or gradient boosting decision tree GBDT, to improve the prediction accuracy through the combination of multiple weak models;

[0096] The risk assessment model after training and verification will be deployed into the system to be used for real-time assessment of the safety risks of the tunnel group; according to the real-time data input, the system uses the trained machine learning model to predict the risks of the tunnel group and timely identify potential high-risk areas; as new data is continuously input and historical data is updated, the risk assessment model will be retrained and optimized regularly to ensure its good prediction effect in a changing environment;

[0097] The trained and optimized model can provide an accurate risk assessment report for the tunnel group. The report includes the safety levels of each tunnel, including low, medium, and high risks, as well as possible risk factors and countermeasures; the output results of the model will be used to support subsequent hierarchical early warning, emergency response decision-making, and resource allocation to ensure that emergency management personnel can take timely and effective measures based on accurate risk assessment results;

[0098] In step S2, optimizing the risk assessment model through the machine learning algorithm and training and verifying it in combination with the historical data within the tunnel group can improve the accuracy and reliability of the tunnel group safety risk prediction; this process ensures through data preprocessing, feature extraction, model training and verification, and model optimization that the risk assessment model can adapt to the data changes and complexities in actual applications, thereby providing strong support for subsequent early warning and emergency management.

[0099] In step S3, construct the risk propagation network of the tunnel group and use the graph theory analysis model to calculate the risk propagation path. It should be further noted that S31. First, collect the geographical layout data of the tunnel group and the connection relationships between the tunnels. Each tunnel is regarded as a node in the network, and the mutual relationships between the tunnels, such as traffic flow and risk propagation paths, are regarded as the connecting edges between the nodes; these connections may represent actual physical connections, the influence of traffic flow, or potential risk transmission paths;

[0100] Collect risk factor data for each tunnel. For example, structural deformation, leakage, and traffic flow. These data will be used as inputs for the risk status of each node (tunnel).

[0101] S32. Consider the tunnel group as a graph, where each tunnel is a node, and the connections between tunnels, such as traffic flow, structural influence, and shared traffic paths, are the edges of the graph. Based on the layout of the tunnel group and the mutual influence relationship between tunnels, construct a weighted graph. The edge weights of the graph can reflect the intensity of risk propagation. The attributes of each tunnel node include the risk status of the tunnel and the value of the risk-causing factors. The risk status is at low, medium, and high risk levels. The edge weight between each pair of adjacent tunnels represents the possibility or intensity of risk propagation between tunnels. These weights can be set according to the geographical location, traffic flow, or historical data between tunnels. For example, if the traffic flow between two tunnels is large, then the edge weight between them can be large, indicating that when an accident occurs in one tunnel, the risk will spread to the other tunnel relatively quickly.

[0102] S33. Use the propagation models in graph theory to simulate the propagation path of risks in the tunnel group. The methods include diffusion models, including: breadth-first search, shortest path algorithm, Dijkstra algorithm, or flow network-based models, including: minimum cut algorithm, maximum flow algorithm. Through these models, it can be calculated how the risk spreads in the tunnel group when a high-risk event occurs in a certain tunnel, and potential high-risk areas can be identified.

[0103] Among them, breadth-first search is used to evaluate how the risk spreads layer by layer from a tunnel node to other tunnels. The shortest path algorithm is used to identify the shortest propagation path from one tunnel to another in the tunnel group, considering the edge weights of risk propagation. If it is necessary to evaluate the transfer capacity and propagation intensity of risks in the entire tunnel group, the maximum flow algorithm can be used to calculate the propagation ability of risks in the tunnel group.

[0104] S34. Calculate the risk propagation intensity of the entire tunnel group when a risk event occurs in one tunnel. This process uses a graph theory-based propagation intensity algorithm, combining the risk values of each node and the weights of the edges to calculate the propagation effect of risks in the entire network. Paths with higher propagation intensity may mean greater potential safety hazards and need to be monitored keyly. The calculation of the risk propagation network not only focuses on the risk paths between individual nodes but also needs to calculate the overall risk entropy of the tunnel group. Risk entropy represents the uncertainty of the entire system, and the change in risk entropy can reflect the change in the risk status of the system. When the risk entropy increases, it indicates that the overall risk of the system is increasing.

[0105] S35. Optimize the risk propagation path by dynamically adjusting the edge weights of the graph, i.e., the risk propagation intensity and the node risk status; update the weights of nodes and edges according to real-time data to ensure that the calculated propagation path can adapt to the real-time safety condition of the tunnel group; optimize the emergency response path by calculating the shortest path or the maximum flow path; determine the priority of resource allocation according to these paths to ensure that emergency resources can reach the places in greatest need in a timely manner;

[0106] S36. Finally, the risk propagation path based on the graph theory model will provide a basis for subsequent global risk assessment, hierarchical early warning, and emergency response; this propagation network can clearly show the mutual influence relationship of risks among tunnels in the tunnel group and predict how potential risk events affect the safety of other tunnels; the risk propagation path and the change of risk entropy will be displayed to decision-makers in real time through a graphical interface to help quickly identify risk hotspots and key propagation paths, so as to allocate emergency resources more accurately in the decision-making process;

[0107] In step S3, constructing the risk propagation network of the tunnel group and using the graph theory analysis model to calculate the risk propagation path can comprehensively and dynamically evaluate the mutual influence among tunnels in the tunnel group and simulate the propagation process of risks in the tunnel group; this process provides a scientific basis for subsequent global risk assessment, hierarchical early warning, and emergency response strategies, ensuring that emergency management personnel can accurately identify high-risk areas and make timely responses.

[0108] In step S4, optimize the early warning level according to the global safety condition of the tunnel group through the hierarchical early warning strategy, and achieve the optimal allocation of resources by adjusting response measures. It should be further noted that S41. The hierarchical early warning module receives the output data from the global risk analysis module, and these data describe the overall safety state of the tunnel group, including the risk levels of each tunnel, the risk propagation path, and the change of the global risk entropy; the hierarchical early warning module analyzes the global safety condition of the tunnel group, evaluates the mutual influence among tunnels, and identifies the nodes that may pose significant risks to the entire tunnel group, i.e., high-risk tunnels and their surrounding environments; by comprehensively considering the local risk levels of each tunnel, the risk propagation path, and historical accident data, determine the overall risk situation of the tunnel group;

[0109] S42. According to the global safety condition of the tunnel group, the hierarchical early warning module adjusts the early warning levels of each tunnel; specifically:

[0110] Low-risk tunnels: For low-risk tunnels, maintain the low-risk early warning, do not trigger a response, and only conduct routine monitoring;

[0111] Medium-risk tunnels: For medium-risk tunnels, the system will strengthen monitoring and issue medium-level early warnings, which may trigger measures such as routine inspections and structural health monitoring;

[0112] High-risk tunnels: For high-risk tunnels, the system will issue a high-level warning and recommend immediate emergency response measures, such as dispatching rescue teams, implementing traffic control, or blocking the tunnel; depending on the overall safety situation, it may affect the risk assessment and response strategies of other tunnels;

[0113] In actual operation, the warning level will change dynamically, depending on the fluctuations in the overall safety situation; for example, if the risk level of a certain tunnel rises due to an accident or abnormal event, the hierarchical warning module will automatically raise the warning level of that tunnel and evaluate its impact on the safety of the entire tunnel group, and adjust the corresponding emergency response strategy in a timely manner;

[0114] S43. According to the optimized warning level, the hierarchical warning module will adjust the response measures for each tunnel; for high-risk tunnels, the system will recommend rapid deployment of emergency response measures, such as mobilizing rescue resources, strengthening traffic management, or implementing emergency evacuation; for medium-risk tunnels, monitoring and inspection may be strengthened; for low-risk tunnels, response measures may be delayed and routine monitoring maintained;

[0115] The response decision module adjusts the allocation plan of emergency resources according to the output of the hierarchical warning module to ensure that the emergency response of each tunnel can be effectively supported; the allocation of emergency resources not only considers the risk level of a single tunnel, but also takes into account the overall safety situation of the tunnel group to optimize the allocation of resources:

[0116] Priority allocation of rescue resources: For high-risk tunnels, the system will give priority to dispatching rescue teams, emergency equipment, traffic control personnel, etc. to ensure rapid response and risk reduction;

[0117] Resource sharing and coordinated scheduling: When different levels of risks occur in multiple tunnels, the system will coordinate the emergency responses of each tunnel, avoid duplicate use or conflict of resources, and ensure that resources are optimally allocated according to the overall safety needs of the tunnel group;

[0118] S44. When adjusting response measures, the hierarchical warning module not only considers the risk level of a single tunnel, but also takes into account the risk propagation effect between tunnels; the allocation of resources is optimized based on the overall safety situation to avoid local optimal solutions; for example, when a high-risk tunnel is closed for rescue, it may be necessary to re-evaluate the traffic flow changes in the surrounding tunnels and adjust the allocation of emergency resources to balance the safety risks of each tunnel in the tunnel group; the system uses genetic algorithms and simulated annealing algorithms to dynamically adjust the resource configuration to minimize the overall risk level of the tunnel group while improving the resource utilization efficiency; the system will also update the resource allocation plan in real time according to the change of the warning level to ensure the rapidity and effectiveness of the emergency response;

[0119] S45. Based on the optimized warning levels and response measures, the hierarchical warning module generates an emergency resource allocation plan and outputs it to the emergency management personnel in the form of a report; the report lists in detail the warning level, response measures, and resource allocation priority and schedule for each tunnel; the report will be updated in real time according to the risk status of the tunnel group, and provide feedback during the emergency response process to adjust the allocation of resources and ensure that emergency resources are always executed according to the optimal configuration;

[0120] S46. During the implementation of response measures, the graded warning module will monitor the effect of emergency response in real time, including risk reduction, resource usage and the implementation effect of emergency measures; if the original response measures fail to effectively reduce the risk, the system will adjust the warning level and response measures according to the feedback information; according to the evaluation results, the system will adjust the resource allocation plan and optimize the emergency response strategy to ensure the maximization of overall safety;

[0121] In step S4, the graded warning module optimizes the warning level according to the global safety status of the tunnel group and achieves the optimal allocation of resources by adjusting the response measures; this process ensures that emergency resources are allocated efficiently and reasonably through dynamic optimization, resource coordination, and response effect evaluation, so as to minimize the overall risk of the tunnel group and improve the effectiveness and efficiency of emergency response.

[0122] It should be further explained that, in the specific implementation process, by introducing global risk analysis and graded warning logic, we ensure the rational allocation of emergency resources and the consistency of global safety goals. At the same time, we combine advanced machine learning and graph theory models to enhance the intelligence and response capabilities of the overall system, thus solving the problem of disconnection between local response strategies and global risk evolution.

[0123] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0124] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A hierarchical early warning management system for risk factors of tunnel groups, characterized in that: Includes the following modules: A data collection module collects risk factor data of each tunnel in real time through a sensor network deployed in the tunnel group, wherein the risk factor data includes tunnel structure risk factor data, traffic safety risk factor data, operation management risk factor data and surrounding environment risk factor data; The risk assessment module uses a multi-dimensional assessment model based on the collected risk factor data to calculate the safety risk level of each tunnel and dynamically adjust the risk assessment parameters according to the overall safety status of the tunnel group; The global risk analysis module simulates the risk propagation path and risk entropy changes based on the global safety status of the tunnel group, evaluates the mutual impact between tunnels, and generates global risk change trends; The hierarchical warning module automatically generates hierarchical warning response strategies based on the output of the global risk analysis module through hierarchical logic, and reasonably allocates emergency resources according to the overall safety status of the tunnel group; The response decision module optimizes the deployment plan of emergency resources and adjusts the response strategy according to the warning level of the hierarchical warning module and the emergency management objectives to ensure the rational use of emergency resources and reduce risks.

2. A hierarchical early warning management system for risk factors of tunnel groups according to claim 1, characterized in that: The risk assessment module uses a machine learning algorithm to train and predict various data of the tunnel group, and combines historical accident records with sensor data to generate a safety risk model for the tunnel group, wherein various data in the tunnel group are collected in real time through a sensor network, and historical accident records are also collected and organized into a standardized data format, which is cleaned by a data preprocessing module and enters the risk assessment module; key features that affect tunnel safety are selected from the processed data set, and the safety risks of the tunnel group are trained using a machine learning algorithm in combination with historical accident records and real-time sensor data; a safety risk model for the tunnel group is generated based on the machine learning model obtained through training, and the model is updated regularly, and risk assessment parameters are adjusted according to real-time data.

3. A hierarchical early warning management system for risk factors of tunnel groups according to claim 2, characterized in that: The global risk analysis module uses a graph theory model to construct a risk propagation network of a tunnel group, analyzes the mutual impact of risks within the tunnel group, adjusts the risk assessment criteria of the tunnel group, integrates real-time and historical data from each tunnel, and performs cleaning, denoising and standardization processing; Based on the layout of the tunnel group and the connection relationship between tunnels, a risk propagation network of the tunnel group is constructed using a graph theory model. Each tunnel is regarded as a node, and the connection between tunnels is set according to the possibility and influence of risk propagation, and each tunnel node is given a risk factor. The shortest path algorithm or diffusion model in graph theory is used to simulate the risk propagation path in the tunnel group and evaluate the mutual influence between different tunnels. With the change of risk propagation path and the fluctuation of entropy, the risk assessment standard of the tunnel group is dynamically adjusted. When the risk status of the tunnel affects the surrounding tunnels, the model parameters are updated based on the new risk assessment information. After evaluation, the global risk change trend is generated and transmitted to the hierarchical warning module and the response decision module.

4. A hierarchical early warning management system for risk factors of tunnel groups according to claim 3, characterized in that: The hierarchical warning module adopts a hierarchical warning model based on the trend of global risk changes, issues different levels of warnings to tunnels of different risk levels, and adjusts corresponding response measures according to the warning levels, receives output data from the global risk analysis module, and evaluates the overall safety status of the tunnel group based on the global risk change trend through a preset hierarchical warning model, and generates corresponding warning response strategies according to different risk levels; When the risk of a tunnel is in the low risk zone, only routine monitoring reminders or status reports are issued, and no emergency response measures are taken; When the risk of a tunnel is in the medium risk zone, an alarm is triggered and regular inspection or reinforcement work is recommended. When the risk of a tunnel reaches the high risk zone, an emergency warning is triggered and immediate measures are recommended. According to the hierarchical warning model, the risk status of each tunnel is analyzed and the warning level to which it belongs is determined; Emergency resources are deployed according to the risk level and warning response strategy determined by the graded warning module, including: in low-risk situations, emergency resources maintain regular allocation; in medium-risk situations, some emergency resources are dispatched for inspection, assessment or advance deployment of emergency personnel; in high-risk situations, emergency resources are mobilized first, and the priority and location of resources are reasonably arranged according to the regional distribution of risks; the graded warning module generates an allocation plan for emergency resources and outputs a report to the emergency management system.

5. A hierarchical early warning management system for risk factors of tunnel groups according to claim 4, characterized in that: The response decision module monitors the safety status of each tunnel in the tunnel group in real time, adjusts the priority and deployment plan of emergency resources, receives the warning level from the graded warning module, receives the emergency management target, analyzes the relationship between the warning level and the emergency management target, determines the priority of the emergency response, and determines the emergency measures to be taken for each tunnel or area; evaluates the emergency resources by integrating the real-time data, warning level and historical data of the tunnel group; According to the changes in the global risk assessment, the response strategy is adjusted in real time. According to the risk level of each tunnel, the availability of emergency resources and the global safety status, a reasonable allocation plan for emergency resources is generated in real time. The resource allocation plan is automatically adjusted according to the timeliness and spatial distribution of resources to optimize the emergency response process. A resource allocation report is generated based on the optimized emergency resource allocation plan and output to relevant emergency management personnel, listing the allocation status, priority, and specific implementation schedule and action plan of emergency resources.

6. A hierarchical early warning management system for risk factors of tunnel groups according to claim 5, characterized in that: The emergency resources include maintenance teams, traffic control resources, and emergency equipment, which are automatically deployed through the system to deal with tunnels of different risk levels.

7. A hierarchical early warning management method for risk factors of tunnel groups, characterized in that: The following steps are involved: S1. Collect the risk factor data of each tunnel in the tunnel group in real time, and clean and pre-process the data through the data processing module; S2. Use the risk assessment module to calculate the risk level of each tunnel based on the preprocessed data and dynamically adjust the assessment criteria; S3. Analyze the risk propagation network of the tunnel group based on the global risk analysis module, simulate the risk evolution process, and calculate the mutual influence between tunnels; S4. Based on the global risk change trend, use the graded warning module to generate a graded response warning strategy and generate corresponding emergency management goals; S5. Based on the graded warning, the emergency resource allocation plan is optimized through the response decision module, and targeted risk management measures are implemented.

8. A hierarchical early warning management method for risk factors of tunnel groups according to claim 7, characterized in that: In step S2, a machine learning algorithm is used to optimize the risk assessment model, and training and verification are performed in combination with historical data within the tunnel group.

9. A hierarchical early warning management method for risk factors of tunnel groups according to claim 8, characterized in that: In step S3, a risk propagation network of the tunnel group is constructed, and a graph theory analysis model is used to calculate the risk propagation path.

10. A hierarchical early warning management method for risk factors of tunnel groups according to claim 9, characterized in that: In step S4, the warning level is optimized according to the global security status of the tunnel group through a hierarchical warning strategy, and the optimal configuration of resources is achieved by adjusting the response measures.

Images