A method and system for intelligent evaluation of structural safety during operation of a submarine tunnel

Abstract

The present invention provides a method and system for intelligent evaluation of structural safety during the operation of an undersea tunnel, which relates to the technical field of intelligent evaluation of structural safety of undersea tunnels. First, sensors are arranged at the vault, joints and support positions of the tunnel to collect strain, displacement, acceleration and temperature data; then, historical earthquake data of the area is collected and feature extracted for subsequent machine learning models, and the model is trained with historical earthquake feature vectors, with the maximum acceleration as the output, and a vibration damage index is assigned to each output data based on the expert scoring method; the stress data of the tunnel is obtained by analyzing the strain and temperature change data, and a stress data source is provided for subsequent model construction; the collected data is compared with a preset threshold, the health state probability is calculated and a damage assessment model is trained; finally, the relevant data is analyzed to generate a structural safety index of the undersea tunnel to judge the current operating state. This method aims to provide a scientific and effective evaluation basis for the safe operation of undersea tunnels.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent evaluation of the structural safety of submarine tunnels, and in particular to a method and system for intelligent evaluation of the structural safety of submarine tunnels during operation. Background Art

[0002] Undersea tunnels are important infrastructure that connects the two sides of the land. They are usually used for transportation, water supply, drainage or power transmission, and can effectively reduce obstacles to water transportation and improve transportation efficiency. These tunnels are usually built using complex engineering techniques and face challenges from the marine environment, such as water pressure, earthquakes, soil characteristics, etc. Therefore, their design and construction need to fully consider safety and durability. With the acceleration of urbanization and the promotion of economic globalization, undersea tunnels play an increasingly critical role in modern transportation networks and have become an important link in promoting regional economic development. At the same time, they also put forward higher requirements for engineering technology.

[0003] Traditional assessment methods often rely on manual inspections, which are inefficient and susceptible to subjective factors. The data sources for analysis are relatively one-sided and fail to comprehensively analyze the health status of the structure during the operation of the undersea tunnel from multiple aspects. In addition, the tunnel environment cannot be monitored in real time, which lags in discovering possible abnormalities.

[0004] The prior art has the following deficiencies:

[0005] There are many shortcomings in the existing safety assessment technology of submarine tunnel structures. First, traditional monitoring methods often rely on manual inspections and regular assessments, which makes it difficult to achieve real-time data collection and analysis, resulting in insufficient response capabilities to emergencies. In addition, manual assessments are easily affected by subjective factors and may lead to misjudgments. Secondly, the existing technology lacks flexibility and adaptability when dealing with complex environmental changes and variable load conditions, and cannot fully reflect the actual operating status of the tunnel. The lack of data analysis capabilities is also particularly prominent, and the failure to effectively use big data technology for deep mining and intelligent processing has resulted in potential risks and damages that cannot be identified and assessed in a timely manner. In addition, the lack of the ability to integrate multiple data sources often leads to one-sided evaluation results, thereby affecting the scientificity and accuracy of decision-making. In summary, the existing technology still has obvious deficiencies in real-time, accuracy and comprehensive analysis capabilities, and more advanced technical means are urgently needed to improve the safety management level of submarine tunnels.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for intelligently evaluating the structural safety of a submarine tunnel during operation, so as to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] A method for intelligently evaluating the structural safety of a submarine tunnel during operation, comprising the following specific steps:

[0010] Step 1: Mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data;

[0011] Step 2: Collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, focal mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on the historical earthquake data except the maximum acceleration to obtain the feature vector of the historical earthquake data;

[0012] Step 3: Based on the machine learning model, the feature vector of the historical earthquake data is input, and the maximum acceleration of the corresponding timestamp is used as the label to train the seismic motion model. Based on the expert scoring method, the maximum acceleration of each timestamp is assigned a vibration damage index;

[0013] Step 4: Combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel at the corresponding time stamp;

[0014] Step 5: Compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset thresholds, mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label;

[0015] Step 6: Obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

[0016] Furthermore, the specific logic for collecting data related to the undersea tunnel is as follows:

[0017] The vault, joints and support positions of the undersea tunnel are marked as key positions, and strain gauges, optical fiber sensors, accelerometers and temperature sensors are arranged at each key position to obtain the strain, displacement, acceleration and temperature data of each monitoring point. The missing data are replaced by the mean of the previous timestamp and the next timestamp. The data of all monitoring points are integrated together, and their mean is calculated. The mean of each type of data is calibrated as the strain, displacement, acceleration and temperature data of the undersea tunnel, and the data of the undersea tunnel are preprocessed. The preprocessing includes standardization and denoising, and the preprocessed data are integrated into a sample data set.

[0018] Furthermore, the specific logic for collecting historical seismic data of the submarine tunnel is as follows:

[0019] The longitude and latitude coordinates of the epicenter of historical earthquake data are collected according to the timestamp, and the distance between the epicenter and the nearest monitoring point is obtained. The distance is marked as the epicenter distance and used as a new eigenvalue. The main frequency and the corresponding amplitude in the spectral feature data are extracted as the eigenvalues ​​of the spectral feature. The time history includes the time history of acceleration and displacement. The corresponding mean and standard deviation are calculated as the eigenvalues, and all numerical data are normalized. The soil properties include soil type and soil layer thickness. The unique hot encoding is used to convert the soil properties into numerical features. The pre-processed epicenter distance, magnitude, focal mechanism, numerical characteristics of soil type, main frequency and corresponding amplitude, mean and standard deviation of acceleration time history, and mean and standard deviation of displacement time history are integrated into the eigenvector of historical earthquake data.

[0020] Furthermore, the specific logic for obtaining the vibration damage index is as follows:

[0021] The characteristic vector of historical earthquake data is input into the seismic motion model, and the maximum acceleration of the corresponding timestamp is used as the label. The seismic motion model is trained based on the random forest model. The robustness of the model is evaluated using cross-validation, and the mean square error is used as the performance indicator for each evaluation. The mean square error value of the local seismic model is The model training is completed within the range, and the maximum acceleration of each timestamp is assigned a vibration damage index based on the expert scoring method.

[0022] Furthermore, the specific logic for obtaining the stress data of the submarine tunnel is as follows:

[0023] The strain data and temperature variation of the submarine tunnel are analyzed, and the formula for obtaining the stress data of the submarine tunnel is as follows:

[0024] ;

[0025] in, for The stress of the moment, is the elastic modulus, for The strain difference between the moment and the previous moment, is the coefficient of thermal expansion, for The temperature difference between the current moment and the previous moment.

[0026] Furthermore, the specific logic for constructing the damage assessment model is as follows:

[0027] The stress, strain, displacement, acceleration, temperature and maximum acceleration data of the submarine tunnel are compared with the preset threshold. If the collected data exceeds the preset threshold, it is marked as a damaged state. If the collected data is lower than the preset threshold, it is marked as a healthy state. The data marked with safe states are integrated into feature vectors and input into the damage assessment model. The corresponding health state probability is used as a label to train the damage assessment model. The formula for obtaining the health state probability is as follows:

[0028] ;

[0029] in, for The probability of health status at any moment, for The amount of health status data at all times, for The amount of damage status data at any moment.

[0030] Furthermore, the specific logic for judging the current operating status of the undersea tunnel is as follows:

[0031] At the entrance of the tunnel, vehicles entering are weighed in real time, and the load and tire contact area of ​​each vehicle are recorded. The formula for obtaining the traffic load of the undersea tunnel is: ;

[0032] in, for The total traffic load of the undersea tunnel at that moment, for The number of vehicles passing through the undersea tunnel at any given time, for Moment The weight of the vehicle, for Moment The contact area of ​​a car's tires, for Moment The dynamic load factor of the vehicle;

[0033] The formula for obtaining the structural safety index by analyzing the vibration damage index, stress data and health state probability of the submarine tunnel is as follows:

[0034] ;

[0035] in, is the structural safety index of the undersea tunnel, is the vibration damage index, is stress, is the traffic load, is the health status probability, are the weight coefficients of the corresponding data, ,and ;

[0036] Collect relevant data of the undersea tunnel at the current moment in real time, obtain the structural safety index at the current moment, and compare it with the preset safety threshold. If the structural safety index at the current moment exceeds the preset safety threshold, it is judged that the current moment is in an abnormal state and an early warning is issued. If the structural safety index at the current moment is lower than the preset safety threshold, it is judged that the current moment is in a safe state and the undersea tunnel data will continue to be monitored.

[0037] The present invention further provides a system for intelligently evaluating the structural safety of a submarine tunnel during operation, which is used to implement the above-mentioned method for intelligently evaluating the structural safety of a submarine tunnel during operation, and comprises:

[0038] The data acquisition module is used to mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data;

[0039] The historical earthquake data feature extraction module is used to collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, source mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on historical earthquake data except for the maximum acceleration to obtain the characteristic direction of historical earthquake data;

[0040] The vibration damage assessment module is used to train the seismic motion model based on the machine learning model, input the feature vector of historical earthquake data, and use the maximum acceleration of the corresponding timestamp as a label. It also assigns a vibration damage index to the maximum acceleration of each timestamp according to the expert scoring method.

[0041] The physical model testing module is used to combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel at the corresponding timestamp. The health status assessment module is used to compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset threshold, and mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label;

[0042] The structural safety status analysis module is used to obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

[0043] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0044] This intelligent evaluation method for the structural safety of undersea tunnels has significantly improved the monitoring and evaluation capabilities of the tunnel's health status by integrating a variety of advanced technologies. First, the arrangement and data acquisition of sensors can capture the strain, displacement, acceleration and temperature changes at key locations in real time, making the evaluation of the structural status more comprehensive and accurate. Secondly, by combining the feature extraction of historical earthquake data with the training of machine learning models, it can effectively predict the impact of seismic motion on the tunnel and quantify the damage index, thereby improving the ability to identify potential risks. Finally, by comprehensively analyzing traffic loads and health status probabilities, the structural safety index formed can not only reflect the operating status of the tunnel in a timely manner, but also provide managers with a basis for scientific decision-making, ensuring the safety and operational efficiency of the tunnel. Therefore, this method has significant technical effects and advantages in improving the accuracy, real-time and scientific nature of structural safety monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0046] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0048] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0049] Example:

[0050] See also Figure 1 , the present invention provides a technical solution:

[0051] A method for intelligently evaluating the structural safety of a submarine tunnel during operation, comprising the following specific steps:

[0052] Step 1: Mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data;

[0053] In this embodiment, the specific logic for collecting the relevant data of the submarine tunnel is as follows:

[0054] The vault, joints and support positions of the undersea tunnel are marked as key positions, and strain gauges, optical fiber sensors, accelerometers and temperature sensors are arranged at each key position to obtain the strain, displacement, acceleration and temperature data of each monitoring point. The missing data are replaced by the mean of the previous timestamp and the next timestamp. The data of all monitoring points are integrated together, and their mean is calculated. The mean of each type of data is calibrated as the strain, displacement, acceleration and temperature data of the undersea tunnel, and the data of the undersea tunnel are preprocessed. The preprocessing includes standardization and denoising, and the preprocessed data are integrated into a sample data set.

[0055] The vaults, joints and support locations are marked as critical locations because these locations are usually subjected to the greatest stress and deformation and are prone to structural damage or failure. By focusing on monitoring these locations, potential problems can be discovered in a timely manner to ensure the safety and stability of the tunnel. Strain gauges, fiber optic sensors, accelerometers and temperature sensors are arranged at each critical location to obtain comprehensive information on the status of the subsea tunnel. Different types of sensors can provide multi-dimensional data, which helps to fully understand the structural health of the tunnel. Integrating the data of all monitoring points together, calculating their mean, and calibrating the mean of each type of data as the strain, displacement, acceleration and temperature data of the subsea tunnel helps to provide a unified and comparable data set and reduce errors caused by differences between different monitoring points. By standardizing the data, the comparability between different types of data can be improved, which is conducive to subsequent data analysis and model building. De-noising can eliminate external interference and measurement errors, and improve the accuracy and reliability of the data. Real-time monitoring and data analysis can detect structural anomalies in a timely manner, prevent potential safety hazards, and ensure the safe operation of the subsea tunnel.

[0056] Step 2: Collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, focal mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on the historical earthquake data except the maximum acceleration to obtain the feature vector of the historical earthquake data;

[0057] In this embodiment, the specific logic for collecting historical seismic data of the submarine tunnel is as follows:

[0058] The longitude and latitude coordinates of the epicenter of historical earthquake data are collected according to the timestamp, and the distance between the epicenter and the nearest monitoring point is obtained. The distance is marked as the epicenter distance and used as a new eigenvalue. The main frequency and the corresponding amplitude in the spectral feature data are extracted as the eigenvalues ​​of the spectral feature. The time history includes the time history of acceleration and displacement. The corresponding mean and standard deviation are calculated as the eigenvalues, and all numerical data are normalized. The soil properties include soil type and soil layer thickness. The unique hot encoding is used to convert the soil properties into numerical features. The pre-processed epicenter distance, magnitude, focal mechanism, numerical characteristics of soil type, main frequency and corresponding amplitude, mean and standard deviation of acceleration time history, and mean and standard deviation of displacement time history are integrated into the eigenvector of historical earthquake data.

[0059] The structural safety of undersea tunnels is closely related to seismic activity. Collecting historical seismic data can provide background information on the behavior of tunnels under the influence of earthquakes and help evaluate their seismic resistance and structural safety. The distance between the epicenter and the nearest monitoring point (epicenter distance) is an important feature because the intensity of seismic waves decreases with increasing distance. This feature helps to evaluate the actual impact of earthquakes on tunnels. Extracting the main frequency and amplitude in the spectrum characteristics can help analyze the characteristics of seismic waves and identify the frequencies that may have a greater impact on the tunnel, which is an important basis for structural dynamics analysis. The mean and standard deviation of acceleration and displacement time history are important indicators for measuring the characteristics of seismic waves. These features can provide the degree of impact of seismic waves on tunnels in various time periods. By extracting and integrating multi-dimensional features, the model input can be enriched, which is conducive to improving the accuracy and prediction ability of the earthquake impact analysis model on tunnels. Comprehensive seismic feature data can help to conduct more scientific risk assessments, identify possible risk factors, and provide a basis for tunnel safety management.

[0060] Step 3: Based on the machine learning model, the feature vector of the historical earthquake data is input, and the maximum acceleration of the corresponding timestamp is used as the label to train the seismic motion model. Based on the expert scoring method, the maximum acceleration of each timestamp is assigned a vibration damage index;

[0061] In this embodiment, the specific logic for obtaining the vibration damage index is:

[0062] The characteristic vector of historical earthquake data is input into the seismic motion model, and the maximum acceleration of the corresponding timestamp is used as the label. The seismic motion model is trained based on the random forest model. The robustness of the model is evaluated using cross-validation, and the mean square error is used as the performance indicator for each evaluation. The mean square error value of the local seismic model is The model training is completed within the range, and the maximum acceleration of each timestamp is assigned a vibration damage index based on the expert scoring method.

[0063] When the maximum acceleration is less than 0.05, the vibration damage index is 0. Within the range, the vibration damage index is 0.5, and the maximum acceleration is Within the range, the vibration damage index is 1, and the maximum acceleration is Within the range, the vibration damage index is 2, and the maximum acceleration When the vibration damage index is 3.

[0064] When the maximum acceleration is less than 0.05, it is considered low acceleration. The acceleration in this range is usually not enough to cause structural damage to the undersea tunnel. The vibration damage index is 0. The range is low to medium acceleration, which may have a slight impact on some non-structural components, so the vibration damage index is set to 0.5, and the maximum acceleration is The range is medium acceleration. In this range, the building may suffer a certain degree of damage, so the vibration damage index is 1. The maximum acceleration is The vibration damage index is 2, which is a high acceleration. The structure of the building may be seriously damaged or even cracked. When the acceleration is extremely high, the building will face large-scale structural damage, so the vibration damage index is set to 3.

[0065] The impact of earthquake motion on the structure can be quantified by calculating the maximum acceleration through a formula. The maximum acceleration is an important indicator for assessing structural damage and can reflect the intensity of earthquake motion within a specific time period. By assigning the maximum acceleration of each timestamp through the expert scoring method, a vibration damage index can be assigned to each acceleration value based on professional knowledge and experience. This method can effectively integrate quantitative analysis and qualitative judgment to enhance the reliability of the model. The machine learning model can accurately map the relationship between multidimensional features and vibration damage, provide more accurate damage prediction, and help evaluate the safety of tunnels under earthquakes. By obtaining the vibration damage index, time periods with higher damage risks can be identified in a timely manner, so that corresponding maintenance measures can be taken to reduce the probability of accidents.

[0066] Step 4: Combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel corresponding to the timestamp; in this embodiment, the specific logic for obtaining the stress data of the submarine tunnel is:

[0067] The strain data and temperature variation of the submarine tunnel are analyzed, and the formula for obtaining the stress data of the submarine tunnel is as follows:

[0068] ;

[0069] in, for The stress of the moment, is the elastic modulus, for The strain difference between the moment and the previous moment, is the coefficient of thermal expansion, for The temperature difference between the current moment and the previous moment.

[0070] Using this formula to calculate stress can comprehensively consider the elastic properties and temperature effects of the material. Here, the elastic modulus is a material property, deformation Reflects the actual strain and temperature changes The influence of thermal effects on stress is taken into account. The obtained stress data can be used to evaluate the rationality of the existing design, promote the optimization and improvement of the structure, enhance the seismic resistance of the overall structure, conduct risk assessment more accurately, and formulate emergency plans for extreme events such as earthquakes that may occur in the future.

[0071] Step 5: Compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset thresholds, mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label;

[0072] In this embodiment, the specific logic for constructing the damage assessment model is:

[0073] The stress, strain, displacement, acceleration, temperature and maximum acceleration data of the submarine tunnel are compared with the preset threshold. If the collected data exceeds the preset threshold, it is marked as a damaged state. If the collected data is lower than the preset threshold, it is marked as a healthy state. The data marked with safe states are integrated into feature vectors and input into the damage assessment model. The corresponding health state probability is used as a label to train the damage assessment model. The formula for obtaining the health state probability is as follows:

[0074] ;

[0075] in, for The probability of health status at any moment, for The amount of health status data at all times, for The amount of damage status data at any moment.

[0076] By comparing the collected stress, strain, displacement, acceleration, temperature and maximum acceleration data with the preset threshold, real-time health monitoring of the undersea tunnel can be achieved. This monitoring can detect potential damage in time and ensure the safety of the structure. Marking the health status and damage status helps to construct feature vectors, provide basic data for subsequent machine learning models, and provide clear standards for judging the health status. Inputting feature vectors into the damage assessment model can improve the ability to predict future states. Through machine learning, the model can identify complex patterns and relationships, providing a more reliable basis for health monitoring. The damage assessment model based on machine learning helps to realize the automated monitoring of undersea tunnels, reduce manual intervention, and improve efficiency and accuracy. By analyzing the probability of health status, a more reasonable maintenance and overhaul plan can be formulated, giving priority to the parts with higher risks, thereby effectively reducing potential safety hazards.

[0077] Step 6: Obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

[0078] In this embodiment, the specific logic for determining the current operating status of the undersea tunnel is:

[0079] At the entrance of the tunnel, vehicles entering are weighed in real time, and the load and tire contact area of ​​each vehicle are recorded. The formula for obtaining the traffic load of the undersea tunnel is:

[0080] in, for The total traffic load of the undersea tunnel at that moment, for The number of vehicles passing through the undersea tunnel at any given time, for Moment The weight of the vehicle, for Moment The contact area of ​​a car's tires, for Moment The dynamic load factor of the vehicle;

[0081] Here, the contact area of ​​the vehicle tire can be directly obtained by querying the tire specifications, and the dynamic load factor can be calculated through the vehicle acceleration data based on the formula:

[0082] ;

[0083] in, for Moment The dynamic load factor of the vehicle, for Moment The acceleration of the car, is the empirical coefficient, which is affected by vehicle type, driving speed and road conditions, and its value range is The specific value is determined by combining relevant industry standards and real-time data.

[0084] The formula for obtaining the structural safety index by analyzing the vibration damage index, stress data and health state probability of the submarine tunnel is as follows:

[0085] ;

[0086] in, is the structural safety index of the undersea tunnel, is the vibration damage index, is stress, is the traffic load, is the health status probability, are the weight coefficients of the corresponding data, ,and .

[0087] Collect relevant data of the undersea tunnel at the current moment in real time, obtain the structural safety index at the current moment, and compare it with the preset safety threshold. If the structural safety index at the current moment exceeds the preset safety threshold, it is judged that the current moment is in an abnormal state and an early warning is issued. If the structural safety index at the current moment is lower than the preset safety threshold, it is judged that the current moment is in a safe state and the undersea tunnel data will continue to be monitored.

[0088] Real-time recording of vehicle loads and contact areas entering the undersea tunnel helps to accurately assess the tunnel's load-bearing capacity under different traffic conditions, which is an important part of ensuring tunnel structural safety. By analyzing the vibration damage index, stress data, health state probability, and traffic load, we can fully understand the operating status of the undersea tunnel and ensure that all influencing factors are taken into account. As the vibration damage index increases, it means that the tunnel is more damaged, so the structural safety index will decrease. Increase, The greater the stress on the structure, the closer the tunnel's bearing capacity is to its limit, and the lower its safety is, resulting in The greater the traffic load, the greater the pressure on the tunnel, which may cause The higher the health status probability, the healthier the tunnel structure is. Increase, if Increase, The maximum acceleration directly affects the dynamic response of the tunnel, especially when the vehicle accelerates and decelerates, which may cause the structure to be subjected to instantaneous impact. Therefore, the corresponding vibration damage index The weight is usually the highest. Stress is a key indicator for evaluating the bearing capacity of materials and structures. Excessive stress may cause the material to yield or fail. In tunnel design and safety assessment, the influence of stress cannot be ignored. The weight of the tunnel should also be relatively high. The traffic load of the tunnel is another important influencing factor, especially for the design and daily safety assessment of the tunnel. The change of traffic load can directly affect the structural stability of the tunnel, so The weight of is lower than that of vibration damage index and stress data. The health state probability reflects the actual health status of the structure and is of great significance for safety assessment, but it is usually a qualitative indicator and may be relatively weak to some extent. The weight may be smaller. This constraint ensures that the leakage probability index is a normalized indicator, ensuring that the total contribution of the weighted combination of the four factors is 100%.

[0089] See also Figure 2 The present invention also provides a method system for intelligently evaluating the structural safety of a submarine tunnel during operation. The system is used to implement the above-mentioned method for intelligently evaluating the structural safety of a submarine tunnel during operation, and includes:

[0090] The data acquisition module is used to mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data;

[0091] The historical earthquake data feature extraction module is used to collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, source mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on historical earthquake data except for the maximum acceleration to obtain the characteristic direction of historical earthquake data;

[0092] The vibration damage assessment module is used to train the seismic motion model based on the machine learning model, input the feature vector of historical earthquake data, and use the maximum acceleration of the corresponding timestamp as a label. It also assigns a vibration damage index to the maximum acceleration of each timestamp according to the expert scoring method.

[0093] The physical model testing module is used to combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel at the corresponding timestamp. The health status assessment module is used to compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset threshold, and mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label;

[0094] The structural safety status analysis module is used to obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

[0095] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0096] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0097] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0098] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for intelligent evaluation of structural safety during operation of a submarine tunnel, characterized in that: The specific steps include: Step 1: Mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data; Step 2: Collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, focal mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on the historical earthquake data except the maximum acceleration to obtain the feature vector of the historical earthquake data; Step 3: Based on the machine learning model, the feature vector of the historical earthquake data is input, and the maximum acceleration of the corresponding timestamp is used as the label to train the seismic motion model. Based on the expert scoring method, the maximum acceleration of each timestamp is assigned a vibration damage index; Step 4: Combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel at the corresponding time stamp; Step 5: Compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset thresholds, mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label; Step 6: Obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

2. The intelligent evaluation method for structural safety of a submarine tunnel during operation according to claim 1 is characterized in that: The specific logic for collecting data related to the undersea tunnel is: The vault, joints and support positions of the undersea tunnel are marked as key positions, and strain gauges, optical fiber sensors, accelerometers and temperature sensors are arranged at each key position to obtain the strain, displacement, acceleration and temperature data of each monitoring point. The missing data are replaced by the mean of the previous timestamp and the next timestamp. The data of all monitoring points are integrated together, and their mean is calculated. The mean of each type of data is calibrated as the strain, displacement, acceleration and temperature data of the undersea tunnel, and the data of the undersea tunnel are preprocessed. The preprocessing includes standardization and denoising, and the preprocessed data are integrated into a sample data set.

3. The method for intelligent evaluation of structural safety during operation of a submarine tunnel according to claim 2, characterized in that: The specific logic for collecting historical seismic data of the submarine tunnel is as follows: The longitude and latitude coordinates of the epicenter of historical earthquake data are collected according to the timestamp, and the distance between the epicenter and the nearest monitoring point is obtained. The distance is marked as the epicenter distance and used as a new eigenvalue. The main frequency and the corresponding amplitude in the spectral feature data are extracted as the eigenvalues ​​of the spectral feature. The time history includes the time history of acceleration and displacement. The corresponding mean and standard deviation are calculated as the eigenvalues, and all numerical data are normalized. The soil properties include soil type and soil layer thickness. The unique hot encoding is used to convert the soil properties into numerical features. The pre-processed epicenter distance, magnitude, focal mechanism, numerical characteristics of soil type, main frequency and corresponding amplitude, mean and standard deviation of acceleration time history, and mean and standard deviation of displacement time history are integrated into the eigenvector of historical earthquake data.

4. The intelligent evaluation method for structural safety during operation of a submarine tunnel according to claim 3 is characterized in that: The specific logic for obtaining the vibration damage index is: The characteristic vector of historical earthquake data is input into the seismic motion model, and the maximum acceleration of the corresponding timestamp is used as the label. The seismic motion model is trained based on the random forest model. The robustness of the model is evaluated using cross-validation, and the mean square error is used as the performance indicator for each evaluation. The mean square error value of the local seismic model is The model training is completed within the range, and the maximum acceleration of each timestamp is assigned a vibration damage index based on the expert scoring method.

5. The method for intelligent evaluation of structural safety during operation of a submarine tunnel according to claim 4, characterized in that: The specific logic for obtaining stress data of the submarine tunnel is: The strain data and temperature variation of the submarine tunnel are analyzed, and the formula for obtaining the stress data of the submarine tunnel is as follows: ； in, for The stress of the moment, is the elastic modulus, for The strain difference between the moment and the previous moment, is the coefficient of thermal expansion, for The temperature difference between the current moment and the previous moment.

6. The method for intelligent evaluation of structural safety during operation of a submarine tunnel according to claim 5, characterized in that: The specific logic for constructing the damage assessment model is as follows: The stress, strain, displacement, acceleration, temperature and maximum acceleration data of the submarine tunnel are compared with the preset threshold. If the collected data exceeds the preset threshold, it is marked as a damaged state. If the collected data is lower than the preset threshold, it is marked as a healthy state. The data marked as safe states are integrated into feature vectors and input into the damage assessment model. The corresponding health state probability is used as a label to train the damage assessment model. The formula for obtaining the health state probability is as follows: ； in, for The probability of health status at any moment, for The amount of health status data at all times, for The amount of damage status data at any moment.

7. The method for intelligently evaluating the structural safety of a submarine tunnel during operation according to claim 6, characterized in that: The specific logic for judging the current operating status of the undersea tunnel is: At the entrance of the tunnel, vehicles entering are weighed in real time, and the load and tire contact area of ​​each vehicle are recorded. The formula for obtaining the traffic load of the undersea tunnel is: ； in, for The total traffic load of the undersea tunnel at that moment, for The number of vehicles passing through the undersea tunnel at any given time, for Moment The weight of the vehicle, for Moment The contact area of ​​a car's tires, for Moment The dynamic load factor of the vehicle; The formula for obtaining the structural safety index by analyzing the vibration damage index, stress data and health state probability of the submarine tunnel is as follows: ； in, is the structural safety index of the undersea tunnel, is the vibration damage index, is stress, is the traffic load, is the health status probability, are the weight coefficients of the corresponding data, ,and ; Collect relevant data of the undersea tunnel at the current moment in real time, obtain the structural safety index at the current moment, and compare it with the preset safety threshold. If the structural safety index at the current moment exceeds the preset safety threshold, it is judged that the current moment is in an abnormal state and an early warning is issued. If the structural safety index at the current moment is lower than the preset safety threshold, it is judged that the current moment is in a safe state and the undersea tunnel data will continue to be monitored.

8. An intelligent evaluation system for structural safety during the operation of a submarine tunnel, characterized in that: The intelligent evaluation system for structural safety during operation of a submarine tunnel is used to implement the intelligent evaluation method for structural safety during operation of a submarine tunnel according to any one of claims 1 to 7, comprising: The data acquisition module is used to mark the arch, joints and support positions of the submarine tunnel as key positions, arrange relevant sensors at the key positions, obtain the strain, displacement, acceleration and temperature data of the submarine tunnel, and perform preprocessing operations on the data; The historical earthquake data feature extraction module is used to collect historical earthquake data in the submarine tunnel area, including epicenter location, magnitude, source mechanism, soil characteristics and earthquake wave parameters, where the earthquake wave parameters include maximum acceleration, spectrum characteristics and time history, and perform feature extraction operations on historical earthquake data except for the maximum acceleration to obtain the characteristic direction of historical earthquake data; The vibration damage assessment module is used to train the seismic motion model based on the machine learning model, input the feature vector of historical earthquake data, and use the maximum acceleration of the corresponding timestamp as a label. It also assigns a vibration damage index to the maximum acceleration of each timestamp according to the expert scoring method. The physical model testing module is used to combine the strain data and temperature data of the submarine tunnel to obtain the stress data of the submarine tunnel corresponding to the timestamp; The health status assessment module is used to compare the collected stress, strain, displacement, acceleration, temperature, and maximum acceleration with the preset thresholds, mark the safety status of each data, generate feature vectors, calculate the health status probability of each timestamp data, and train the damage assessment model based on the machine learning model, taking the feature vector as input and the corresponding health status probability as a label; The structural safety status analysis module is used to obtain the traffic load of the undersea tunnel, analyze the vibration damage index, stress, health status probability and traffic load of the undersea tunnel, obtain the structural safety index, and compare it with the preset threshold to determine the current operating status of the undersea tunnel.

Images