A tunnel monitoring data emergency response warning method and system

By conducting spatiotemporal correlation analysis and risk score in the tunnel monitoring system, and dynamically adjusting the early warning threshold in combination with traffic flow, the problem of inability to comprehensively consider multi-dimensional risk factors in the existing technology is solved, and the safety and efficiency of tunnel operations are improved.

CN118548114BActive Publication Date: 2024-11-19CHINA RAILWAY FIRST GROUP FIFTH ENGINEERING CO LTD +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411032677.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-19
Estimated Expiration
2044-07-30

AI Technical Summary

Technical Problem

The existing tunnel monitoring emergency warning system cannot fully consider the unique environmental and structural characteristics of different sections, ignores the interaction between various factors, and fails to comprehensively consider multi-dimensional risk factors, making it difficult for managers to make accurate and timely response decisions. , reducing the safety and efficiency of tunnel operations.

Method used

Using the emergency response warning method of tunnel monitoring data, the time series and spatial data analysis is carried out by obtaining structural, environmental and traffic flow monitoring data, establishing a spatio-temporal correlation model, extracting a comprehensive feature set, inputting a risk prediction model, calculating a comprehensive risk score, and accordingly Traffic flow dynamically adjusts the warning threshold and generates emergency warning prompts.

Benefits of technology

It improves the accuracy and timeliness of emergency warnings, optimizes resource allocation, ensures the safety and efficiency of tunnel operations, and provides strong technical support for tunnel emergency response management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118548114B_ABST
    Figure CN118548114B_ABST
Patent Text Reader

Abstract

The present application relates to a tunnel monitoring data emergency response warning method and system, which belongs to the field of monitoring and warning technology. The method includes: obtaining monitoring data information of each section in the tunnel and performing data preprocessing; performing time series analysis and spatial data analysis on the preprocessed monitoring data information, and establishing a time-space correlation model corresponding to each section; extracting a comprehensive feature set for each section according to the time-space correlation model; inputting the comprehensive feature set into the risk prediction model of each section respectively, and predicting the section feature prediction results of each section; calculating the comprehensive risk score of each section according to the section feature prediction results; dynamically adjusting the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow of each section; generating emergency warning prompt information in response to the comprehensive risk score of the section being greater than the corresponding warning threshold and sending it to the warning terminal of the section. The present application improves the safety and efficiency of tunnel operation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of monitoring and early warning technology, and in particular to a tunnel monitoring data emergency response early warning method and system. Background Art

[0002] In the modern transportation system, tunnels are key channels connecting two places, and their safe and stable operation is of vital importance. With the expansion of tunnel construction scale and the increase of service life, factors such as the health status of the tunnel's internal structure, environmental changes, and traffic flow fluctuations have jointly constituted complex operational safety challenges.

[0003] In the construction and maintenance of some extra-long tunnels, tunnels are usually divided into multiple sections for monitoring. Each section faces different environmental challenges and structural safety issues due to its unique geographical location, geological structure, construction materials and service life. For example, some sections may have high humidity and water seepage risks due to their proximity to water sources, while others may be more prone to rock displacement due to geological instability. This difference poses great challenges to tunnel safety monitoring and maintenance.

[0004] At present, most common emergency warning systems cannot fully consider the unique environment and structural characteristics of different sections. The warning of monitoring data often focuses on a single aspect, ignoring the interaction between various factors and the impact of dynamic factors such as traffic flow. In the process of risk assessment and emergency response decision-making, there is a lack of a scientific decision-making framework that comprehensively considers multi-dimensional risk factors, making it difficult for managers to make accurate and efficient response decisions in a short period of time, reducing the safety and efficiency of tunnel operations. Summary of the invention

[0005] In order to improve the safety and efficiency of tunnel operations, the present application provides a tunnel monitoring data emergency response warning method and system.

[0006] In the first aspect, the present application provides a tunnel monitoring data emergency response warning method, which adopts the following technical solution:

[0007] A tunnel monitoring data emergency response warning method, the method comprising:

[0008] Acquire monitoring data information of each section in the tunnel and perform data preprocessing; the monitoring data information includes structure monitoring data, environmental monitoring data and real-time traffic flow monitoring data;

[0009] Perform time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establish a spatiotemporal correlation model corresponding to each section;

[0010] Extracting a comprehensive feature set of each segment according to the spatiotemporal correlation model;

[0011] The comprehensive feature set is input into the pre-built risk prediction model of each section respectively to predict the section feature prediction results of each section; the section feature prediction results include the structural safety factor prediction interval of the section in the future period;

[0012] Calculate the comprehensive risk score of each segment according to the segment feature prediction results;

[0013] According to the real-time monitoring data of the traffic flow in each section, dynamically adjust the warning threshold corresponding to each section;

[0014] In response to the comprehensive risk score of the section being greater than the corresponding warning threshold, emergency warning prompt information is generated and sent to the warning terminal of the section.

[0015] By adopting the above-mentioned technical solutions, integrating advanced monitoring technologies, data analysis methods and intelligent algorithms, a comprehensive predictive analysis of the safety status of each section in the tunnel is carried out, which fully considers the unique environment and structural characteristics of each section, can proactively identify the potential risks of each section, and form a scientific decision-making framework that comprehensively considers multi-dimensional risk factors. It not only improves the accuracy and timeliness of emergency warnings, but also optimizes resource allocation through the adjustment of dynamic thresholds through real-time traffic flow, ensures the safety and efficiency of tunnel operations, and provides strong technical support for the tunnel's emergency response management.

[0016] Optionally, the steps of performing time series analysis and spatial data analysis on the preprocessed monitoring data information of each section and establishing a spatiotemporal correlation model corresponding to each section include:

[0017] Perform time series analysis based on the preprocessed monitoring data information, identify trend characteristics and periodic characteristics, and obtain time series analysis results;

[0018] Respectively mapping the preprocessed environmental monitoring data and structural monitoring data in the spatial dimension and performing dimensionality reduction processing to obtain the spatial distribution characteristics of the environmental monitoring data and structural monitoring data;

[0019] According to the spatial distribution characteristics, the spatial correlation characteristics of the environmental monitoring data and the structural monitoring data are identified through cluster analysis to obtain spatial data analysis results;

[0020] Constructing a space-time cube data structure according to the time series analysis results and the spatial data analysis results;

[0021] Extracting spatiotemporal correlation features based on the spatiotemporal cube data structure;

[0022] Based on the multivariate analysis algorithm, a spatiotemporal correlation model is established according to the spatiotemporal correlation characteristics.

[0023] By adopting the above technical solution, the spatiotemporal correlation of the monitoring data of each section of the tunnel is identified and quantified, and an independent spatiotemporal correlation model is constructed for a single section. This model can describe and predict how specific environmental factors affect the response of the structure at a specific spatial location, providing a more accurate basis for risk assessment and early warning, effectively improving the intelligent level of tunnel monitoring, and enhancing the predictability and response capabilities to potential risks.

[0024] Optionally, according to the spatiotemporal correlation model, the step of extracting a comprehensive feature set for each segment includes:

[0025] Based on the spatiotemporal association model, a feature weight list is obtained;

[0026] Based on a preset weight threshold, the feature weight list is screened;

[0027] Perform correlation analysis and recursive feature elimination on the filtered feature set to obtain feature subsets and corresponding weights;

[0028] Calculate the weighted average score of the features according to the feature subsets and the corresponding weights;

[0029] Feature selection is performed based on the weighted average score to obtain a comprehensive feature set.

[0030] By adopting the above technical solution, the feature weights output by the model are screened to ensure that the subsequent prediction analysis can focus on the most explanatory features for the safety and environmental sensitivity of the section. Through further correlation analysis and recursive feature elimination process, redundant information is effectively removed, the dimension is reduced, and the diversity and independence of the data are maintained. The weighted average score is calculated and the feature selection is performed accordingly, which not only scientifically quantifies the relative importance of each feature, but also constructs a highly concentrated and most predictive comprehensive feature set. This comprehensive feature set not only comprehensively reflects the time series dynamics and spatial distribution characteristics of the section, but also deeply reflects the complex interaction between structural risks and environmental impacts, providing a strong and accurate data foundation for the safety monitoring, risk assessment and intelligent management of the tunnel, greatly improving the efficiency and accuracy of the decision support system.

[0031] Optionally, the method further comprises the step of training the risk prediction model of the segment, and the training step comprises:

[0032] Acquire a sample data set; the sample data set includes historical monitoring data of the section and corresponding historical structural safety factors;

[0033] Performing data preprocessing on the historical monitoring data;

[0034] Perform time series analysis and spatial data analysis on the preprocessed historical monitoring data to extract a comprehensive sample feature set;

[0035] Dividing the comprehensive sample feature set into training samples and test samples;

[0036] Inputting the training samples into a pre-built machine learning model for training, optimizing model parameters, and obtaining a pre-trained risk prediction model;

[0037] The risk prediction model is tested and model parameters are corrected based on the test sample until the loss function of the model meets the preset conditions, thereby obtaining the trained risk prediction model.

[0038] By adopting the above technical solutions, a risk prediction model was established that can predict the range of structural safety factors in future periods. This not only provides a scientific basis for the safety management of infrastructure, but also helps decision makers take preventive measures in advance through quantitative risk assessment, effectively avoid potential risks, and greatly improve the foresight and effectiveness of safety management.

[0039] Optionally, the step of calculating a comprehensive risk score for each segment according to the segment feature prediction result includes:

[0040] According to the prediction results of the section characteristics, a prediction interval of the structural safety factor of each section in the future period is extracted;

[0041] According to the structural safety factor prediction interval, the safety factor prediction interval mean and interval width of each section are obtained;

[0042] Based on the preset risk level standard, according to the mean of the safety factor prediction interval, a risk score based on the mean is obtained;

[0043] According to the interval width, a risk gain based on the interval width is obtained;

[0044] The comprehensive risk score is calculated based on the risk score and the risk gain based on the preset calculation weights.

[0045] By adopting the above technical scheme, combined with the mean and interval width of the structural safety factor prediction interval, a quantitative assessment of the future safety risks of the tunnel section is carried out, providing an accurate, comprehensive and adjustable risk assessment method for tunnel safety management, enhancing the timeliness and accuracy of risk warning, and providing solid decision-making support for the formulation of preventive measures and emergency response.

[0046] Optionally, the step of dynamically adjusting the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section includes:

[0047] Obtain historical traffic flow monitoring data for each section;

[0048] According to the historical traffic flow monitoring data, identifying the periodic pattern of each section; the periodic pattern includes peak period, off-peak period and valley period;

[0049] According to the periodic pattern corresponding to the current time period, based on the real-time monitoring data of the traffic flow in each section, a dynamic adjustment factor for each section is determined; the dynamic adjustment factor includes a time factor and a real-time flow ratio;

[0050] Based on the basic warning threshold of each section, the basic warning threshold of each section is dynamically adjusted according to the dynamic adjustment factor.

[0051] By adopting the above technical solution, dynamic adjustment of the warning threshold is achieved, which not only takes into account the real-time changes in traffic flow, but also incorporates intelligent analysis of time periodicity and traffic fluctuations, ensuring the flexibility, pertinence and accuracy of the warning system, effectively reducing false alarms and missed alarms, and improving the intelligence level of road safety management.

[0052] In the second aspect, the present application provides a tunnel monitoring data emergency response warning system, which adopts the following technical solution:

[0053] A tunnel monitoring data emergency response warning system, the system comprising:

[0054] An acquisition module is used to acquire monitoring data information of each section in the tunnel; the monitoring data information includes structure monitoring data, environmental monitoring data and real-time traffic flow monitoring data;

[0055] A preprocessing module, used for performing data preprocessing on the monitoring data information;

[0056] The spatiotemporal correlation model building module is used to perform time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establish the spatiotemporal correlation model corresponding to each section;

[0057] A comprehensive feature set extraction module, used to extract a comprehensive feature set of each section according to the spatiotemporal correlation model;

[0058] A section feature prediction module is used to input the comprehensive feature set into a pre-built risk prediction model for each section to predict the section feature prediction results of each section; the section feature prediction results include a predicted interval of the structural safety factor of the section in the future period;

[0059] A comprehensive risk score calculation module, used to calculate the comprehensive risk score of each segment according to the segment feature prediction result;

[0060] A threshold dynamic adjustment module, used to dynamically adjust the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section;

[0061] The emergency warning prompt module is used to generate emergency warning prompt information and send it to the warning terminal of the section in response to the comprehensive risk score of the section being greater than the corresponding warning threshold.

[0062] Optionally, the spatiotemporal correlation model building module includes:

[0063] A time series analysis unit is used to perform time series analysis based on the preprocessed monitoring data information, identify trend characteristics and periodic characteristics, and obtain time series analysis results;

[0064] A spatial distribution feature generating unit is used to map the pre-processed environmental monitoring data and structural monitoring data in the spatial dimension and perform dimensionality reduction processing to obtain the spatial distribution features of the environmental monitoring data and structural monitoring data;

[0065] A spatial data analysis unit, configured to identify spatial correlation characteristics of the environmental monitoring data and the structural monitoring data through cluster analysis according to the spatial distribution characteristics, and obtain spatial data analysis results;

[0066] A space-time cube construction unit, used to construct a space-time cube data structure according to the time series analysis results and the spatial data analysis results;

[0067] A spatiotemporal correlation feature extraction unit, used for extracting spatiotemporal correlation features based on the spatiotemporal cube data structure;

[0068] The modeling unit is used to establish a spatiotemporal correlation model based on a multivariate analysis algorithm and the spatiotemporal correlation characteristics.

[0069] In a third aspect, the present application provides a computer device, which adopts the following technical solution:

[0070] A computer device comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect.

[0071] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution:

[0072] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any one of the methods in the first aspect.

[0073] To summarize, the present application includes at least one of the following beneficial technical effects: integrating advanced monitoring technology, data analysis methods and intelligent algorithms, conducting comprehensive predictive analysis on the safety status of each section in the tunnel, fully considering the unique environment and structural characteristics of each section, and being able to proactively identify the potential risks of each section, forming a scientific decision-making framework that comprehensively considers multi-dimensional risk factors, which not only improves the accuracy and timeliness of emergency warnings, but also optimizes resource allocation by adjusting dynamic thresholds through real-time traffic flow, ensuring the safety and efficiency of tunnel operations, and providing strong technical support for tunnel emergency response management. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 It is a first flow chart of the tunnel monitoring data emergency response warning method according to one of the embodiments of the present application.

[0075] Figure 2 It is a second flow chart of the tunnel monitoring data emergency response warning method of one of the embodiments of the present application.

[0076] Figure 3 It is a third flow chart of the tunnel monitoring data emergency response warning method of one of the embodiments of the present application.

[0077] Figure 4 It is a fourth flow chart of the tunnel monitoring data emergency response warning method of one of the embodiments of the present application.

[0078] Figure 5 It is a fifth flow chart of the tunnel monitoring data emergency response warning method of one of the embodiments of the present application.

[0079] Figure 6 It is the sixth flow chart of the tunnel monitoring data emergency response warning method of one of the embodiments of the present application. DETAILED DESCRIPTION

[0080] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-6 It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0081] The embodiment of the present application discloses a tunnel monitoring data emergency response warning method.

[0082] Reference Figure 1 , a tunnel monitoring data emergency response warning method, the method comprising:

[0083] Step S101, obtaining monitoring data information of each section in the tunnel and performing data preprocessing;

[0084] Among them, the monitoring data information includes structural monitoring data, environmental monitoring data and real-time traffic flow monitoring data;

[0085] In some embodiments, high-precision structural monitoring sensors (such as stress gauges, displacement sensors, crack monitors, etc.), environmental monitoring sensors (such as thermometers and hygrometers, groundwater level meters, soil moisture sensors, etc.) and traffic flow monitoring sensors can be deployed in each section of the tunnel, so as to comprehensively monitor the stability of the tunnel structure, environmental parameter conditions (such as temperature and humidity, concentration of harmful gases) and vehicle flow in different sections. The collection of these data is the basis for subsequent analysis and early warning, ensuring the comprehensiveness and real-time nature of the information.

[0086] Specifically, the data preprocessing steps may include data cleaning, removing outliers, filling missing values, data smoothing and other operations to ensure data quality. By normalizing the data, the dimensional effect can be eliminated to facilitate subsequent analysis.

[0087] Step S102, performing time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establishing a spatiotemporal correlation model corresponding to each section;

[0088] Among them, time series analysis is mainly used to study the temporal dynamics of monitoring data, including trend analysis and periodic detection of monitoring data in this section over time; spatial data analysis aims to explore the spatial distribution patterns and relationships between environmental factors and structural responses, that is, to analyze the spatial distribution characteristics of environmental parameters and how these distributions are associated with the spatial patterns of structural responses.

[0089] Based on this, a spatiotemporal correlation model that reflects the characteristics of the section is constructed. The spatiotemporal correlation model can capture the temporal dynamics and spatial heterogeneity of the data, improve the depth of understanding of the complexity of the tunnel state, and provide possibilities for refined management.

[0090] Step S103, extracting a comprehensive feature set of each segment according to the spatiotemporal correlation model;

[0091] Specifically, the spatiotemporal correlation model is used to extract key indicators that can reflect the safety status of the section from the time series and spatial distribution data to form a comprehensive feature set that integrates multiple dimensions such as structural stability, environmental impact, and traffic load pressure.

[0092] It is understandable that by simplifying a large amount of raw data, the most critical information for safety assessment can be extracted, which facilitates the efficient application of subsequent risk prediction models.

[0093] Step S104, inputting the comprehensive feature set into the pre-built risk prediction model for each segment respectively, and predicting the segment feature prediction result of each segment;

[0094] Among them, the prediction results of section characteristics include the predicted interval of structural safety factor of the section in the future period;

[0095] Specifically, the pre-built risk prediction model learns the nonlinear relationship between structural changes and various factors based on historical data. It can accurately predict the possible range of the structural safety factor of each section in the future period, thereby providing a quantitative future risk assessment, enabling managers to foresee potential problems in advance and leave valuable time for the formulation of preventive measures.

[0096] Step S105, calculating and obtaining a comprehensive risk score for each segment based on the segment feature prediction result;

[0097] Specifically, the section characteristic prediction results are integrated to conduct a quantitative assessment of the future safety risks of the tunnel section. A comprehensive risk score is given to each section to reflect the overall safety status of the section, thus achieving comprehensive consideration of risk assessment and ensuring the comprehensiveness and scientificity of decision-making.

[0098] Step S106, dynamically adjusting the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section;

[0099] In some embodiments, the warning threshold corresponding to the section can be dynamically adjusted according to the real-time changes in the traffic flow in the current section, ensuring that safety requirements are more stringent during high-flow periods, while the standards are appropriately relaxed during low-flow periods to avoid excessive alarms, thereby enhancing the flexibility and pertinence of the warning system, making warnings more accurate and reducing unnecessary interference.

[0100] Step S107, in response to the comprehensive risk score of the section being greater than the corresponding warning threshold, emergency warning prompt information is generated and sent to the warning terminal of the section.

[0101] Among them, once the comprehensive risk score exceeds the dynamically adjusted warning threshold, the system immediately triggers an alarm, generates a warning prompt containing information such as specific sections, risk levels, and recommended measures, and sends it to the corresponding warning terminal through the network to ensure the rapid transmission of information in an emergency, providing immediate action guidance for the on-site management team.

[0102] In the above implementation mode, advanced monitoring technology, data analysis methods and intelligent algorithms are integrated to conduct a comprehensive predictive analysis of the safety status of each section in the tunnel, fully considering the unique environment and structural characteristics of each section, and being able to proactively identify the potential risks of each section, forming a scientific decision-making framework that comprehensively considers multi-dimensional risk factors. This not only improves the accuracy and timeliness of emergency warnings, but also optimizes resource allocation by adjusting dynamic thresholds through real-time traffic flow, ensuring the safety and efficiency of tunnel operations, and providing strong technical support for tunnel emergency response management.

[0103] Reference Figure 2 As an implementation method of step S102, the steps of performing time series analysis and spatial data analysis on the monitoring data information of each section after preprocessing and establishing a spatiotemporal correlation model corresponding to each section include:

[0104] Step S201, performing time series analysis based on the preprocessed monitoring data information, identifying trend characteristics and periodic characteristics, and obtaining time series analysis results;

[0105] Among them, trend characteristics refer to the overall direction or pattern of data over time, and periodic characteristics refer to the repetitive patterns or regular fluctuations of data over time. By processing and analyzing the time series data in the monitoring data information, it is possible to identify the long-term trends and periodic changes of the monitoring data over time.

[0106] In one of the embodiments of the present application, ARIMA (autoregressive integrated moving average) can be used for time series analysis. By fitting and predicting the time series data of each segment, key time series features reflecting the long-term trend of the segment (such as trend coefficient), seasonal fluctuations (such as periodic intensity), and random fluctuations (such as residual variance) are extracted. These features directly reflect the changing pattern of structural stability over time.

[0107] Specifically, the ARIMA model first confirms the stationarity of the data through a unit root test (such as the ADF test), then performs difference processing on the non-stationary series until it becomes stationary, and then determines the autoregressive term (p), difference order (d), and sliding average term (q) of the model based on the autocorrelation and partial autocorrelation diagrams to construct the ARIMA (p, d, q) model.

[0108] Step S202, mapping the pre-processed environmental monitoring data and structural monitoring data in the spatial dimension and performing dimensionality reduction processing respectively to obtain spatial distribution characteristics of the environmental monitoring data and structural monitoring data;

[0109] In one embodiment of the present application, the location information of the monitoring points in each section is paired with the environmental monitoring data and the structural monitoring data, respectively, and mapped to the spatial coordinates for geocoding using GIS technology; then, PCA (principal component analysis) is applied for dimensionality reduction processing, which reduces the data dimension while retaining as much variation information as possible to facilitate subsequent cluster analysis;

[0110] Among them, the spatial distribution characteristics are the spatial distribution patterns of environmental monitoring data and structural monitoring data obtained through GIS technology mapping and PCA dimensionality reduction processing.

[0111] Step S203, identifying spatial correlation characteristics of environmental monitoring data and structural monitoring data through cluster analysis according to spatial distribution characteristics, and obtaining spatial data analysis results;

[0112] As an implementation method of cluster analysis, the DBSCAN (density-based spatial clustering with applications and noise) algorithm can be used to identify the spatial correlation characteristics of environmental monitoring data and structural monitoring data. DBSCAN can determine the clustering according to the number of points in the neighborhood. It does not need to pre-specify the number of clusters, and can better handle data with different density areas.

[0113] Specifically, using the DBSCAN clustering results, each segment is marked with its category in the spatial clustering as a feature reflecting the consistency of the structural response and the spatial distribution of environmental factors. At the same time, the principal component scores of each segment after PCA dimensionality reduction are calculated. These scores can reflect the relative position of the segment in the multidimensional space, especially the spatial association with environmental sensitivity and structural vulnerability.

[0114] It is understandable that cluster analysis can be used to identify the spatial correlation patterns between environmental parameters and structural responses, and to explore the spatial correlation between structural responses and environmental factors, such as whether the structural response is more sensitive near areas with high humidity, or whether the structure is more prone to abnormalities near certain special geological structures, thereby revealing the spatial clustering patterns of environmental and structural responses in different sections, making it easier to understand the spatial dependence between them.

[0115] Step S204, constructing a space-time cube data structure according to the time series analysis results and the spatial data analysis results;

[0116] Specifically, the trend and periodic characteristics obtained from time series analysis and the spatial correlation characteristics obtained from spatial clustering analysis are combined with the original space-time coordinate information to construct a space-time cube; each cube unit represents a set of data features at a specific space-time position, including time series features and spatial features, forming a four-dimensional (three-dimensional space + time) data structure. According to the space-time coordinate information of each segment, all the characteristic values ​​of the segment are filled into the corresponding space-time cube unit. By integrating the results of time series and spatial analysis, a unified data framework is provided for the subsequent extraction of space-time correlation features.

[0117] Step S205, extracting spatiotemporal correlation features based on the spatiotemporal cube data structure;

[0118] Specifically, based on the space-time cube data structure, space-time correlation features such as space-time proximity and space-time consistency indices are calculated. Space-time proximity is measured by calculating the distance or adjacency relationship of data points in the space-time cube, while space-time consistency is evaluated by comparing the similarity of features of the same or adjacent space-time units.

[0119] It can be understood that through the extraction of spatiotemporal correlation features, the complex correlation of data in the spatiotemporal dimensions is captured, providing rich input information for the model.

[0120] Step S206, based on a multivariate analysis algorithm and according to the spatiotemporal correlation characteristics, a spatiotemporal correlation model is established.

[0121] In one of the embodiments of the present application, partial least squares regression (PLSR) can be used as a multivariate analysis algorithm. PLSR is a dimensionality reduction regression method that can simultaneously consider the multicollinearity problem between predictor variables when processing high-dimensional data. It is particularly suitable for analyzing the relationship between spatiotemporal correlation characteristics and response variables of interest. Through the PLSR model, a quantitative relationship between spatiotemporal characteristics and section safety status or performance indicators is established, providing a scientific basis for prediction and risk assessment.

[0122] Specifically, in the process of building the PLSR model, a low-dimensional structure (latent variable) can be found through an iterative process to optimally explain the variation between the independent variables (spatial-temporal characteristics) and the dependent variables (section safety status or performance indicators). The number of latent variables needs to be determined through cross-validation to balance the complexity and explanatory power of the model; and for each latent variable, the relevant weight vectors of the independent variables (spatial-temporal characteristics) and the dependent variables are calculated. These weights reflect the contribution of each variable to the latent variable, while also considering the multicollinearity problem between the independent variables. According to the above weight vector, the model parameters are fitted by minimizing the covariance of the residual sum of squares and the predicted variables, thereby enhancing the prediction accuracy and model explanatory power.

[0123] In the above technical solution, the spatiotemporal correlation of the monitoring data of each section of the tunnel is identified and quantified, and an independent spatiotemporal correlation model is constructed for a single section. This model can describe and predict how specific environmental factors affect the response of the structure at a specific spatial location, providing a more accurate basis for risk assessment and early warning, effectively improving the intelligence level of tunnel monitoring, and enhancing the ability to predict and respond to potential risks.

[0124] Reference Figure 3 As an implementation of step S105, the step of extracting a comprehensive feature set of each segment according to the spatiotemporal correlation model includes:

[0125] Step S301, obtaining a feature weight list based on a spatiotemporal correlation model;

[0126] Among them, the weight values ​​of all features are derived from the model. These weights reflect the contribution of each feature to risk prediction and environmental sensitivity identification. High-weight features have a higher influence on model prediction.

[0127] In the present application, specifically in the monitoring data scenarios of each section of the tunnel, the features can be various quantitative indicators extracted directly from the original monitoring data, which can reflect information on tunnel structure safety, environmental conditions, traffic flow, etc.

[0128] For example, tunnel monitoring data may include structural parameter characteristics, environmental parameter characteristics and traffic flow characteristics, among which structural parameter characteristics include stress / strain, concrete strength, crack data, environmental parameter characteristics include temperature and humidity, carbon dioxide concentration, etc., and traffic flow characteristics include vehicle volume, vehicle type ratio, average vehicle speed, etc.

[0129] Step S302, screening the feature weight list based on a preset weight threshold;

[0130] In some embodiments, thresholds can be pre-set based on domain knowledge and historical experience (e.g., features with weight values ​​in the top percentage or features with absolute values ​​greater than a certain threshold) to preliminarily screen out features that contribute significantly to model prediction. This step is intended to reduce unnecessary computational complexity and focus on analyzing the most critical features.

[0131] Step S303, performing correlation analysis and recursive feature elimination on the filtered feature set to obtain feature subsets and corresponding weights;

[0132] In one embodiment of the present application, for the feature set that has been preliminarily screened, a correlation analysis between the features (such as the Pearson correlation coefficient and the Spearman rank correlation coefficient) is run to identify and remove highly correlated features to reduce multicollinearity problems and ensure the independence and diversity of the feature set; then, the feature set that has been streamlined after the correlation analysis is subjected to recursive feature elimination (RFE), and in each iteration, the least important feature (based on the model coefficient or other scoring criteria) is removed, and the iteration is repeated until a predetermined number of features is reached or the feature importance is no longer significantly increased.

[0133] Step S304, calculating a weighted average score of the features according to the feature subsets and the corresponding weights;

[0134] Among them, according to the feature weights obtained in the RFE process, each feature in the selected feature subset is weighted. The weighting can be obtained by multiplying the feature value by its corresponding weight value, and the relative importance of each feature is determined by calculating the weighted average score of the feature.

[0135] Specifically, the weighted average score is calculated as follows:

[0136] ;

[0137] In the above formula, w i is the weight, x i is the eigenvalue, n is the number of characteristics;

[0138] Step S305: perform feature selection based on the weighted average score to obtain a comprehensive feature set.

[0139] Specifically, the weighted average score of the features is calculated, and the features to be retained are selected based on the score; for example, in some specific embodiments, the first N features with the highest weighted scores may be selected for retention.

[0140] In the above implementation, the feature weights output by the model are used for screening to ensure that subsequent predictive analysis can focus on the most explanatory features for section safety and environmental sensitivity. Through further correlation analysis and recursive feature elimination, redundant information is effectively removed, the dimension is reduced, and the diversity and independence of the data are maintained. The weighted average score is calculated and feature selection is performed based on it, which not only scientifically quantifies the relative importance of each feature, but also constructs a highly concentrated and most predictive comprehensive feature set.

[0141] The comprehensive feature set constructed in this application not only comprehensively reflects the time series dynamics and spatial distribution characteristics of the section, but also deeply reflects the complex interaction between structural risks and environmental impacts, providing a powerful and accurate data foundation for tunnel safety monitoring, risk assessment and intelligent management, greatly improving the efficiency and accuracy of the decision support system.

[0142] Reference Figure 4 As an implementation of the risk prediction model, the method further includes the step of training the risk prediction model for the segment, wherein the training step includes:

[0143] Step S401, obtaining a sample data set;

[0144] Among them, the sample data set includes the historical monitoring data of the section and the corresponding historical structural safety factor;

[0145] Specifically, the historical monitoring data include the historical structural monitoring data, historical environmental monitoring data and historical traffic flow monitoring data of the section. These data directly reflect the changing patterns of the structure, environment and traffic flow status over time and the influence of external factors, and are the basis of model learning.

[0146] Step S402, performing data preprocessing on historical monitoring data;

[0147] Specifically, the data preprocessing steps may include data cleaning, removing outliers, filling missing values, data smoothing and other operations to ensure data quality. By normalizing the data, the dimensional effect can be eliminated to facilitate subsequent analysis.

[0148] Step S403, performing time series analysis and spatial data analysis on the pre-processed historical monitoring data to extract a comprehensive sample feature set;

[0149] The comprehensive sample feature set not only contains the information of a single segment, but also incorporates the complex associations of time and space dimensions, thus enhancing the prediction capability of the model. Specifically, the steps for obtaining the comprehensive sample feature set can refer to the contents of steps S201 to S206 and steps 301 to S305 above.

[0150] Step S404, dividing the comprehensive sample feature set into training samples and test samples;

[0151] Among them, the comprehensive sample feature set is divided into training samples and test samples, usually allocated according to a certain ratio (such as 70% training and 30% testing) to verify the model performance with an independent data set.

[0152] Step S405, inputting the training samples into a pre-built machine learning model for training, optimizing the model parameters, and obtaining a pre-trained risk prediction model;

[0153] In one of the embodiments of the present application, a long short-term memory network (LSTM) can be used as the core algorithm model to pre-build and train the risk prediction model for each segment. Because it is good at processing long-term dependencies in time series data, it is more suitable for predicting the dynamic changes of structural safety factors in the present application.

[0154] As the architecture of the long short-term memory network (LSTM) model, it mainly includes an input layer, an LSTM layer, a dropout layer, an output layer and a loss function; among them, the input layer is defined, and the input size of the LSTM model can be set according to the dimension of the comprehensive feature set; a multi-layer LSTM structure is constructed, and each layer contains a certain number of hidden units to capture the long-term dependencies in the time series; a dropout layer is introduced to reduce overfitting; since the prediction is the prediction interval of the structural safety factor rather than a single value, the mean square error (MSE) can be used as the main loss function, and combined with a certain form of probability distribution (such as parameter prediction of normal distribution) to express the uncertainty of the structural safety factor; the output layer is designed to predict the parameters of the corresponding structural safety factor interval.

[0155] During the model training process, training sample data can be used to adjust model parameters through optimization algorithms such as back propagation and gradient descent, such as selecting appropriate hyperparameters such as batch size and learning rate.

[0156] Step S406, testing the risk prediction model and calibrating model parameters based on the test sample until the loss function of the model meets the preset conditions, thereby obtaining a trained risk prediction model.

[0157] Among them, the model performance is evaluated on the test set that has not participated in the training, focusing on the prediction accuracy and the coverage of interval prediction. If the model performs poorly (the loss function is higher than the preset threshold), it is necessary to adjust the model structure or parameters and train again until the preset conditions are met.

[0158] In the above implementation mode, a risk prediction model is established that can predict the range of structural safety factors in future periods. This model not only provides a scientific basis for the safety management of infrastructure, but also helps decision makers take preventive measures in advance through quantitative risk assessment, effectively avoids potential risks, and greatly improves the foresight and effectiveness of safety management.

[0159] Reference Figure 5 As an implementation of step S105, the step of calculating the comprehensive risk score of each segment according to the segment feature prediction result includes:

[0160] Step S501, extracting the predicted interval of the structural safety factor of each section in the future period according to the section characteristic prediction result;

[0161] Step S502, according to the structural safety factor prediction interval, obtain the safety factor prediction interval mean and interval width of each section;

[0162] Specifically, assuming that for a certain section, the predicted structural safety factor interval for the future period is [L, U], where L is the lower limit and U is the upper limit, the mean of the safety factor prediction interval is μ=(L+U) / 2, and the interval width is W=UL;

[0163] Step S503, based on the preset risk level standard, according to the mean of the safety factor prediction interval, obtain a risk score based on the mean;

[0164] In some embodiments, the preset risk level standard is, for example, low risk, medium risk, high risk, and the safety factor threshold of each risk level is μ respectively. low , μ mid and μ high , and μ low <μ mid <μ high ; According to the relationship between the mean μ and the preset threshold, the risk level division can be achieved by using conditional statements, and f(μ) is the risk scoring function based on the mean μ.

[0165] For example, if μ<μ low , then set f(μ)=10 (high risk score). Otherwise, if μ low ≤μ<μ mid , then set f(μ)=5 (medium risk score), otherwise μ≥μ mid , set f(μ)=1 (low risk score).

[0166] Step S504, obtaining a risk gain based on the interval width according to the interval width;

[0167] Among them, the adjustment factor K of the interval width can be set in advance, and the risk gain g(W) can be directly calculated by applying the formula g(W)=K·W; it can be understood that the wider the interval, the higher the risk score, and k can be pre-set as an adjustment factor according to the sensitivity to uncertainty; in the embodiment of the present application, k can be set to 0.01.

[0168] Step S505: Based on the preset calculation weights, a comprehensive risk score is calculated according to the risk score and the risk gain.

[0169] Specifically, the preset calculation weights include the central tendency weight w1 and the uncertainty weight w2, and the comprehensive risk score is Score=w1·f(μ)+w2·g(W). Among them, the central tendency weight w1 is used to reflect the importance of the mean μ of the safety factor prediction interval, and the uncertainty weight w2 reflects the importance of the interval width W, that is, the uncertainty of the future safety factor. In the embodiment of the present application, w1 can be set to 0.7 and w2 can be set to 0.3.

[0170] It should be noted that the weights w1, w2 and the threshold μ can be adjusted. low , μ mid , μ high , and the adjustment factor k of the interval width to optimize the performance of the scoring model.

[0171] In the above implementation, the future safety risk of the tunnel section is quantitatively assessed by combining the mean and interval width of the structural safety factor prediction interval, providing an accurate, comprehensive and adjustable risk assessment method for tunnel safety management, enhancing the timeliness and accuracy of risk warnings, and providing solid decision-making support for the formulation of preventive measures and emergency response.

[0172] Reference Figure 6 As an implementation method of step S106, the step of dynamically adjusting the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section includes:

[0173] Step S601, obtaining historical traffic flow monitoring data of each section;

[0174] Among them, through the sensor network installed in each section, key parameters such as the number and speed of passing vehicles are continuously collected to form a historical data set. These data are crucial for analyzing the traffic flow characteristics of each time period.

[0175] Step S602, identifying the periodic pattern of each section based on historical traffic flow monitoring data;

[0176] Among them, the periodic pattern includes peak period, off-peak period and valley period;

[0177] Specifically, through time series analysis of historical data, we can identify the patterns of traffic flow changes in different sections within a day, a week, or even longer periods of time, and clearly distinguish between peak hours, off-peak hours, and trough hours; by revealing the periodicity and predictability of traffic flow, we can provide theoretical support for subsequent differentiated management by time periods, which will help to rationally allocate resources and improve traffic management efficiency.

[0178] Step S603, determining a dynamic adjustment factor for each section based on the real-time monitoring data of traffic flow in each section according to the periodic pattern corresponding to the current time period;

[0179] Among them, the dynamic adjustment factors include time factor and real-time flow ratio;

[0180] Specifically, according to the different periodic patterns (peak, flat, and valley), different time factors (time factor is a coefficient used to reflect the importance and risk level of traffic flow in different time periods, and is set according to the periodic pattern to which the current period belongs) and real-time traffic ratios (real-time traffic ratio is a coefficient used to reflect the ratio of actual traffic flow to expected flow in the current period, and can be estimated based on the average flow in the same period in history). Among them, a higher time factor can be set during peak hours to increase vigilance; a lower time factor can be set during valley hours to reduce unnecessary alarms.

[0181] Exemplarily, first determine which traffic periodicity pattern (peak, flat, or valley) the current time period (time point) falls into. When it is determined to be in a peak period, the configurable time factor is greater than 1 (for example, set to 1.3); when it is in a flat period, the configurable time factor is 1; when it is in a valley period, the configurable time factor is less than 1 (for example, set to 0.8); at the same time, calculate the ratio of the current real-time monitoring data of traffic flow to the expected flow in the period, and the expected flow can be the average value of the historical traffic flow in the same period; it can be understood that if the current real-time monitoring data of traffic flow is higher than the average value, the real-time flow ratio is greater than 1, otherwise it is less than 1.

[0182] It is understandable that by introducing dynamic adjustment factors, the early warning mechanism can flexibly adapt to traffic conditions in different time periods, which can not only strengthen monitoring during high-risk periods, but also reduce false alarms during low-traffic periods, thereby improving the pertinence and practicality of the early warning.

[0183] Step S604: Based on the basic warning threshold of each section, dynamically adjust the basic warning threshold of each section according to the dynamic adjustment factor.

[0184] Specifically, the basic warning threshold set in each section is multiplied by the dynamic adjustment factor (ratio of time factor to real-time traffic) to obtain the actual warning threshold for the current period, to ensure that the warning threshold responds quickly to changes in actual conditions.

[0185] Among them, the basic warning threshold of each segment can be adjusted and configured based on historical data analysis and professional standards, aiming to provide a baseline for the risk score. When the risk score exceeds this threshold, there may be potential security risks.

[0186] In the above implementation mode, dynamic adjustment of the warning threshold is realized, which not only takes into account the real-time changes in traffic flow, but also incorporates intelligent analysis of time periodicity and traffic fluctuations, ensuring the flexibility, pertinence and accuracy of the warning system, effectively reducing false alarms and missed alarms, and improving the intelligence level of road safety management.

[0187] This application significantly improves the intelligence level and response speed of tunnel safety management and reduces operating costs. It is of great significance for improving tunnel operation safety and ensuring the safety of personnel and property, and demonstrates the huge potential of technological innovation in the field of infrastructure maintenance.

[0188] The embodiment of the present application also discloses a tunnel monitoring data emergency response warning system.

[0189] A tunnel monitoring data emergency response warning system, the system comprising:

[0190] An acquisition module is used to acquire monitoring data information of each section in the tunnel; the monitoring data information includes structure monitoring data, environmental monitoring data and real-time traffic flow monitoring data;

[0191] A preprocessing module is used to preprocess the monitoring data information;

[0192] The spatiotemporal correlation model building module is used to perform time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establish the spatiotemporal correlation model corresponding to each section;

[0193] A comprehensive feature set extraction module is used to extract the comprehensive feature set of each segment according to the spatiotemporal correlation model;

[0194] The section feature prediction module is used to input the comprehensive feature set into the pre-built risk prediction model of each section to predict the section feature prediction results of each section; the section feature prediction results include the predicted interval of the structural safety factor of the section in the future period;

[0195] The comprehensive risk score calculation module is used to calculate the comprehensive risk score of each segment based on the segment feature prediction results;

[0196] The threshold dynamic adjustment module is used to dynamically adjust the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section;

[0197] The emergency warning prompt module is used to generate emergency warning prompt information and send it to the warning terminal of the section in response to the comprehensive risk score of the section being greater than the corresponding warning threshold.

[0198] As an implementation of the spatiotemporal correlation model construction module, the spatiotemporal correlation model construction module includes:

[0199] A time series analysis unit is used to perform time series analysis based on the preprocessed monitoring data information, identify trend characteristics and periodic characteristics, and obtain time series analysis results;

[0200] A spatial distribution feature generating unit is used to map the pre-processed environmental monitoring data and structural monitoring data in the spatial dimension and perform dimensionality reduction processing to obtain spatial distribution features of the environmental monitoring data and structural monitoring data;

[0201] A spatial data analysis unit is used to identify the spatial correlation characteristics of environmental monitoring data and structural monitoring data through cluster analysis according to the spatial distribution characteristics, and obtain spatial data analysis results;

[0202] A space-time cube construction unit is used to construct a space-time cube data structure according to the time series analysis results and the spatial data analysis results;

[0203] A spatiotemporal correlation feature extraction unit, used for extracting spatiotemporal correlation features based on a spatiotemporal cube data structure;

[0204] The modeling unit is used to establish a spatiotemporal correlation model based on a multivariate analysis algorithm and according to the spatiotemporal correlation characteristics.

[0205] The tunnel monitoring data emergency response warning system of the embodiment of the present application can implement any of the above-mentioned tunnel monitoring data emergency response warning methods, and the specific working process of each module in the tunnel monitoring data emergency response warning system can refer to the corresponding process in the above-mentioned method embodiment.

[0206] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are only illustrative; for example, the division of a certain module is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0207] The embodiment of the present application also discloses a computer device.

[0208] The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the tunnel monitoring data emergency response warning method as described above is implemented.

[0209] The embodiment of the present application also discloses a computer-readable storage medium.

[0210] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any of the above-mentioned tunnel monitoring data emergency response and early warning methods.

[0211] Among them, computer-readable storage media can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus or device; the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0212] It should be noted that in the above embodiments, the description of each embodiment has different emphases, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0213] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A tunnel monitoring data emergency response warning method, characterized in that: The method comprises: Acquire monitoring data information of each section in the tunnel and perform data preprocessing; the monitoring data information includes structure monitoring data, environmental monitoring data and real-time traffic flow monitoring data; Perform time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establish a spatiotemporal correlation model corresponding to each section; Extracting a comprehensive feature set of each segment according to the spatiotemporal correlation model; The comprehensive feature set is input into the pre-built risk prediction model of each section respectively to predict the section feature prediction results of each section; the section feature prediction results include the structural safety factor prediction interval of the section in the future period; Calculate the comprehensive risk score of each segment according to the segment feature prediction results; According to the real-time monitoring data of the traffic flow in each section, dynamically adjust the warning threshold corresponding to each section; In response to the comprehensive risk score of the section being greater than the corresponding warning threshold, emergency warning prompt information is generated and sent to the warning terminal of the section.

2. The tunnel monitoring data emergency response warning method according to claim 1 is characterized in that: The steps of performing time series analysis and spatial data analysis on the monitoring data information of each section after preprocessing and establishing the spatiotemporal correlation model corresponding to each section include: Perform time series analysis based on the preprocessed monitoring data information, identify trend characteristics and periodic characteristics, and obtain time series analysis results; Respectively mapping the preprocessed environmental monitoring data and structural monitoring data in the spatial dimension and performing dimensionality reduction processing to obtain the spatial distribution characteristics of the environmental monitoring data and structural monitoring data; According to the spatial distribution characteristics, the spatial correlation characteristics of the environmental monitoring data and the structural monitoring data are identified through cluster analysis to obtain spatial data analysis results; Constructing a space-time cube data structure according to the time series analysis results and the spatial data analysis results; Extracting spatiotemporal correlation features based on the spatiotemporal cube data structure; Based on the multivariate analysis algorithm, a spatiotemporal correlation model is established according to the spatiotemporal correlation characteristics.

3. A tunnel monitoring data emergency response warning method according to claim 2, characterized in that: According to the spatiotemporal correlation model, the step of extracting a comprehensive feature set for each segment includes: Based on the spatiotemporal association model, a feature weight list is obtained; Based on a preset weight threshold, the feature weight list is screened; Perform correlation analysis and recursive feature elimination on the filtered feature set to obtain feature subsets and corresponding weights; Calculate the weighted average score of the features according to the feature subsets and the corresponding weights; The calculation formula for the weighted average score is: ; In the above formula, w i is the weight, x i is the eigenvalue, n is the number of characteristics; Feature selection is performed based on the weighted average score to obtain a comprehensive feature set.

4. The tunnel monitoring data emergency response warning method according to claim 1 is characterized in that: The method further comprises the step of training the risk prediction model for the segment, wherein the training step comprises: Acquire a sample data set; the sample data set includes historical monitoring data of the section and corresponding historical structural safety factors; Performing data preprocessing on the historical monitoring data; Perform time series analysis and spatial data analysis on the preprocessed historical monitoring data to extract a comprehensive sample feature set; Dividing the comprehensive sample feature set into training samples and test samples; Inputting the training samples into a pre-built machine learning model for training, optimizing model parameters, and obtaining a pre-trained risk prediction model; The risk prediction model is tested and model parameters are corrected based on the test sample until the loss function of the model meets the preset conditions, thereby obtaining the trained risk prediction model.

5. The tunnel monitoring data emergency response warning method according to claim 1 is characterized in that: The step of calculating the comprehensive risk score of each segment according to the segment feature prediction result includes: According to the prediction results of the section characteristics, a prediction interval of the structural safety factor of each section in the future period is extracted; According to the structural safety factor prediction interval, the safety factor prediction interval mean and interval width of each section are obtained; Based on the preset risk level standard, according to the mean of the safety factor prediction interval, a risk score based on the mean is obtained; According to the interval width, a risk gain based on the interval width is obtained; The calculation formula of risk gain is: g(W)=K·W, where g(W) is risk gain, K is adjustment factor, and W is interval width; Based on the preset calculation weights, the comprehensive risk score is calculated according to the risk score and the risk gain; Among them, the calculation formula for the comprehensive risk score is: Score=w1·f(μ)+w2·g(W); in the above formula, w1 is the central trend weight, w2 is the uncertainty weight, μ is the mean of the safety factor prediction interval, f(μ) is the risk scoring function based on the mean μ, and g(W) is the risk gain.

6. A tunnel monitoring data emergency response warning method according to any one of claims 1 to 5, characterized in that: The step of dynamically adjusting the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section comprises: Obtain historical traffic flow monitoring data for each section; According to the historical traffic flow monitoring data, identifying the periodic pattern of each section; the periodic pattern includes peak period, off-peak period and valley period; According to the periodic pattern corresponding to the current time period, based on the real-time monitoring data of the traffic flow in each section, the dynamic adjustment factor of each section is determined; the dynamic adjustment factor includes a time factor and a real-time flow ratio, the time factor is used to reflect the importance and risk level of traffic flow in different time periods, and the real-time flow ratio is used to reflect the ratio of the actual traffic flow in the current time period to the expected flow; Based on the basic warning threshold of each section, the basic warning threshold of each section is dynamically adjusted according to the dynamic adjustment factor.

7. A tunnel monitoring data emergency response warning system, characterized in that: The system comprises: An acquisition module is used to acquire monitoring data information of each section in the tunnel; the monitoring data information includes structure monitoring data, environmental monitoring data and real-time traffic flow monitoring data; A preprocessing module, used for performing data preprocessing on the monitoring data information; The spatiotemporal correlation model building module is used to perform time series analysis and spatial data analysis on the pre-processed monitoring data information of each section, and establish the spatiotemporal correlation model corresponding to each section; A comprehensive feature set extraction module, used to extract a comprehensive feature set of each section according to the spatiotemporal correlation model; A section feature prediction module is used to input the comprehensive feature set into a pre-built risk prediction model for each section to predict the section feature prediction results of each section; the section feature prediction results include a predicted interval of the structural safety factor of the section in the future period; A comprehensive risk score calculation module, used to calculate the comprehensive risk score of each segment according to the segment feature prediction result; A threshold dynamic adjustment module, used to dynamically adjust the warning threshold corresponding to each section according to the real-time monitoring data of the traffic flow in each section; The emergency warning prompt module is used to generate emergency warning prompt information and send it to the warning terminal of the section in response to the comprehensive risk score of the section being greater than the corresponding warning threshold.

8. The tunnel monitoring data emergency response warning system according to claim 7, characterized in that: The spatiotemporal correlation model building module includes: A time series analysis unit is used to perform time series analysis based on the preprocessed monitoring data information, identify trend characteristics and periodic characteristics, and obtain time series analysis results; A spatial distribution feature generating unit is used to map the pre-processed environmental monitoring data and structural monitoring data in the spatial dimension and perform dimensionality reduction processing to obtain the spatial distribution features of the environmental monitoring data and structural monitoring data; A spatial data analysis unit, configured to identify spatial correlation characteristics of the environmental monitoring data and the structural monitoring data through cluster analysis according to the spatial distribution characteristics, and obtain spatial data analysis results; A space-time cube construction unit, used to construct a space-time cube data structure according to the time series analysis results and the spatial data analysis results; A spatiotemporal correlation feature extraction unit, used for extracting spatiotemporal correlation features based on the spatiotemporal cube data structure; The modeling unit is used to establish a spatiotemporal correlation model based on a multivariate analysis algorithm and the spatiotemporal correlation characteristics.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 6 when executing the program.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 6.