Domestic server-based subsea tunnel information fusion identification and decision-making method

Through the information fusion judgment and decision-making method of undersea tunnel based on domestic servers, and the random forest algorithm and data fusion technology are used to solve the shortcomings of risk assessment and decision-making support systems in the construction of undersea tunnels, and efficient identification and risk of complex geological conditions are achieved. Quantification improves the scientificity and response speed of decision-making, reduces costs and promotes technological innovation and industrial upgrading.

CN119066535APending Publication Date: 2024-12-03INSPUR QILU SOFTWARE IND
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411157210.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The construction environment of undersea tunnels is complex, and the concealment and variability of geological conditions lead to insufficient risk assessment and decision-making support systems. It is difficult for existing technologies to effectively utilize big data and artificial intelligence, and cannot meet the needs of rapid response and precise decision-making of complex information.

Method used

The undersea tunnel information fusion judgment and decision-making method is adopted based on domestic servers. Through the integration and analysis of multiple information, the risk identification model is trained using a random forest algorithm, and combined with data feature extraction and weighted average fusion rules, risk quantification and response measures are generated.

Benefits of technology

It significantly enhances risk prevention and control capabilities, improves the scientific nature of decision-making and response speed, reduces costs, promotes technological innovation and industrial upgrading, ensures data security, and promotes the greening and sustainable development of marine engineering.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066535A_ABST
    Figure CN119066535A_ABST
Patent Text Reader

Abstract

The invention discloses an undersea tunnel information fusion identification and decision-making method based on a domestic server, and relates to the technical field of intelligent data analysis. Comprising the steps of 1, collecting multivariate information of a subsea tunnel, 2, integrating and analyzing the collected multivariate information, and 3, training a risk identification model by adopting a Random Forest supervised learning algorithm: forming a data set by using preprocessed data features, training the model by using training set data in the data set, and identifying the risk of the subsea tunnel. The hyper-parameters of the model are adjusted through cross validation to optimize the model performance; analyzing the importance of each data feature in the risk identification model, determining which features most influence risk assessment, setting a threshold value, setting the threshold value to distinguish normal and abnormal conditions, carrying out risk quantification, and quantifying a risk level through a probability value or a classification result output by the risk identification model; 4, obtaining a probability value or a classification result according to the collected multivariate information by using a risk identification model, and obtaining a corresponding risk level according to the probability value or the classification result, and 5, formulating corresponding countermeasures according to different risk levels.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention discloses an information fusion identification and decision-making method for an undersea tunnel based on a domestic server, and relates to the technical field of intelligent data analysis. Background Art

[0002] With the rapid development of global infrastructure construction, as an important channel connecting islands and land and crossing straits, the construction technology and safety management of submarine tunnels have increasingly become the research focus in the field of engineering technology. However, compared with land tunnels and other conventional construction projects, the construction environment and conditions of submarine tunnels are significantly more special and challenging, which have caused some problems:

[0003] Increased concealment and complexity: Subsea tunnels are buried deep underwater, and their geological structure is far more complex than that of the ground, often containing a variety of unfavorable geological bodies such as soft soil, faults, and karst caves. This concealment makes it difficult to intuitively and accurately grasp the overall geological picture before design and construction, increasing the uncertainty of the project.

[0004] Environmental factors are variable: The dynamic characteristics of the marine environment, such as seawater corrosion, currents, tidal changes, and seabed stability, pose a constant threat to tunnel structure design and construction safety. In addition, harsh offshore weather conditions also limit the construction window, increasing construction difficulty and cost.

[0005] Limitations of risk assessment and response mechanisms: Current risk assessments of undersea tunnels rely heavily on geological exploration technology and limited field data. The ability to identify and predict complex geological conditions is limited, which can easily lead to missed assessments of adverse geological conditions, thus affecting the safety and economy of the entire project.

[0006] Insufficient information processing and decision support systems: Although many technologies have been applied to the construction and management of undersea tunnels, information islands are common, and the integration, analysis and application of multi-source data are inefficient. Existing decision support systems often lack effective use of big data and artificial intelligence technologies, and cannot meet the needs of rapid response to complex information and accurate decision-making. Summary of the invention

[0007] In view of the problems of the prior art, the present invention provides an information fusion identification and decision-making method for an undersea tunnel based on a domestic server, aiming to realize efficient identification of geological conditions of the undersea tunnel, risk assessment and generation of response strategies by integrating the powerful data processing capabilities of the domestic server.

[0008] The specific scheme proposed by the present invention is:

[0009] The present invention provides an information fusion identification and decision-making method for an undersea tunnel based on a domestic server, comprising:

[0010] Step 1: Collect multivariate information of the submarine tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, fractured zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information includes blasting excavation, tunnel spraying, and support construction information.

[0011] Step 2: Integrate and analyze the collected multi-dimensional information:

[0012] Step 21: Perform multivariate information preprocessing,

[0013] Step 22: Extract and select data features based on the preprocessed multivariate information.

[0014] Step 23: Perform data fusion on the selected data features: Use similarity measurement to match the patterns between different data sources; perform consistency check to ensure that the fused data is logically consistent; Use weighted average fusion rules to fuse data according to data type and application scenario.

[0015] Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data;

[0016] Step 3: Use the Random Forest supervised learning algorithm to train the risk identification model: Use the preprocessed data features to form a data set, use the training set data in the data set to train the model, and adjust the model's hyperparameters through cross-validation to optimize the model performance;

[0017] Analyze the importance of each data feature in the risk identification model and determine which features have the greatest impact on risk assessment.

[0018] Set thresholds: Set thresholds to distinguish between normal and abnormal situations.

[0019] Quantify risks: Quantify risk levels through probability values ​​or classification results output by risk identification models;

[0020] Step 4: Use the risk identification model to obtain the probability value or classification result based on the collected multivariate information, and obtain the corresponding risk level based on the probability value or classification result.

[0021] Step 5: Develop corresponding response measures based on different risk levels.

[0022] Furthermore, in step 21 of the method for fusion identification and decision-making of submarine tunnel information based on a domestic server, multivariate information preprocessing is performed, including:

[0023] Data cleaning: remove invalid values, fill or remove missing values, detect and process outliers,

[0024] Data conversion: Standardize or normalize raw data to ensure compatibility between different data sources.

[0025] Data integration: Integrate data from multiple data sources to resolve data redundancy and conflict issues.

[0026] Furthermore, in step 22 of the method for fusion identification and decision-making of submarine tunnel information based on a domestic server, data feature extraction and selection are performed, including: when performing feature extraction, meaningful data features are extracted from the original data, and the principal component analysis PCA method is used to reduce the data dimension,

[0027] When performing feature selection, a filtering method is used to select the most representative data features for subsequent analysis.

[0028] Furthermore, in step 3 of the method for fusion identification and decision-making of submarine tunnel information based on a domestic server, the pre-processed data features are used to form a data set, including:

[0029] The agreed preprocessed data features include the weak bottom layer, broken zone and fault, permeable stratum, seismic zone information in geological data, weather, temperature, humidity, tide data in environmental data, blasting excavation, tunnel spraying, support construction information in construction status information,

[0030] A data set is formed based on data characteristics.

[0031] Furthermore, in step 5 of the method for information fusion identification and decision-making of an undersea tunnel based on a domestic server, rules are pre-defined for risk scenarios, and when the risk level reaches the threshold of the preset rules, the emergency plan is activated, and a decision tree is used to automatically recommend an action plan based on the risk level and specific circumstances.

[0032] The present invention also provides an undersea tunnel information fusion identification and decision-making device based on a domestic server, comprising: a multi-source data acquisition module, a big data fusion processing module, an intelligent risk assessment module and a decision support module.

[0033] The multi-source data acquisition module collects multi-dimensional information of the submarine tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, broken zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information includes blasting excavation, tunnel spraying, and support construction information.

[0034] The big data fusion processing module integrates and analyzes the collected multi-dimensional information:

[0035] Step 21: Perform multivariate information preprocessing,

[0036] Step 22: Extract and select data features based on the preprocessed multivariate information.

[0037] Step 23: Perform data fusion on the selected data features: Use similarity measurement to match the patterns between different data sources; perform consistency check to ensure that the fused data is logically consistent; Use weighted average fusion rules to fuse data according to data type and application scenario.

[0038] Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data;

[0039] The intelligent risk assessment module uses the Random Forest supervised learning algorithm to train the risk identification model: the preprocessed data features are used to form a data set, the training set data in the data set is used to train the model, and the model's hyperparameters are adjusted through cross-validation to optimize the model performance;

[0040] Analyze the importance of each data feature in the risk identification model and determine which features have the greatest impact on risk assessment.

[0041] Set thresholds: Set thresholds to distinguish between normal and abnormal situations.

[0042] Quantify risks: Quantify risk levels through probability values ​​or classification results output by risk identification models;

[0043] The risk identification model is used to obtain probability values ​​or classification results based on the collected multivariate information, and the corresponding risk level is obtained based on the probability values ​​or classification results.

[0044] The decision support module formulates corresponding response measures according to different risk levels.

[0045] Furthermore, the big data fusion processing module of the domestic server-based undersea tunnel information fusion identification and decision-making device executes step 21 to perform multivariate information preprocessing, including:

[0046] Data cleaning: remove invalid values, fill or remove missing values, detect and process outliers,

[0047] Data conversion: Standardize or normalize raw data to ensure compatibility between different data sources.

[0048] Data integration: Integrate data from multiple data sources to resolve data redundancy and conflict issues.

[0049] Furthermore, the big data fusion processing module of the domestic server-based undersea tunnel information fusion identification and decision-making device executes step 22 to extract and select data features, including: when extracting features, extracting meaningful data features from the original data, using the principal component analysis PCA method to reduce the data dimension,

[0050] When performing feature selection, a filtering method is used to select the most representative data features for subsequent analysis.

[0051] Furthermore, the analytical risk identification model of the undersea tunnel information fusion identification and decision-making device based on a domestic server uses the preprocessed data features to form a data set, including:

[0052] The agreed preprocessed data features include the weak bottom layer, broken zone and fault, permeable stratum, seismic zone information in geological data, weather, temperature, humidity, tide data in environmental data, blasting excavation, tunnel spraying, support construction information in construction status information,

[0053] A data set is formed based on data characteristics.

[0054] Furthermore, the decision support module of the undersea tunnel information fusion identification and decision-making device based on a domestic server predefines rules for risk scenarios, activates the emergency plan when the risk level reaches the threshold of the preset rules, and uses a decision tree to automatically recommend action plans based on the risk level and specific circumstances.

[0055] The benefits of the present invention are:

[0056] (1) Significantly enhance risk prevention and control capabilities: By effectively integrating multiple monitoring data, the speed and accuracy of identifying potential risks in undersea tunnel construction are improved, which helps to take measures to avoid disasters in advance and ensure the safety of personnel and the smooth progress of the project.

[0057] (2) Improving the scientific nature and response speed of decision-making: Intelligent decision support systems can analyze complex data in real time and quickly generate response strategy suggestions, helping managers make reasonable decisions quickly at critical moments, reducing the decision-making cycle and enhancing the effectiveness and flexibility of decisions.

[0058] (3) Reduce costs and improve efficiency: Accurate risk prediction reduces unnecessary preventive construction and repair work, reducing the total cost of the project. At the same time, the system automatically processes large amounts of data, reducing manpower requirements and improving overall work efficiency.

[0059] Promote technological innovation and industrial upgrading: The development and application of the system have promoted the cross-integration of information technology, artificial intelligence and tunnel engineering technology, provided a new technical paradigm for related industries, and promoted technological progress and industrial upgrading of the entire industrial chain.

[0060] (4) Ensure data security: Using domestic servers as infrastructure ensures information security during data processing, complies with the country's data protection requirements for critical infrastructure, and enhances national strategic security.

[0061] (5) Environmental friendliness and sustainable development: Through precise management of construction activities, interference with marine ecology is reduced, promoting the greening and sustainable development of marine engineering projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic flow chart of the method of the present invention.

[0063] Figure 2 It is a system platform architecture diagram involved in the present invention. DETAILED DESCRIPTION

[0064] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0065] Example 1

[0066] The present invention provides an information fusion identification and decision-making method for an undersea tunnel based on a domestic server, comprising:

[0067] Step 1: Collect multivariate information of the undersea tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, fractured zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information is divided into blasting excavation, tunnel spraying, and support construction information.

[0068] Step 2: Integrate and analyze the collected multi-dimensional information:

[0069] Step 21: Perform multivariate information preprocessing, including, for example:

[0070] Data cleaning: remove invalid values, fill or remove missing values, detect and process outliers,

[0071] Data conversion: Standardize or normalize raw data to ensure compatibility between different data sources.

[0072] Data integration: Integrate data from multiple data sources to resolve data redundancy and conflict issues.

[0073] Step 22: Extract and select data features based on the preprocessed multivariate information. Extracting and selecting data features may further include: extracting meaningful data features from the original data, using the principal component analysis (PCA) method to reduce the data dimension,

[0074] When performing feature selection, a filtering method is used to select the most representative data features for subsequent analysis.

[0075] Step 23: Perform data fusion on the selected data features: Use similarity measurement to match the patterns between different data sources; perform consistency check to ensure that the fused data is logically consistent; Use weighted average fusion rules to fuse data according to data type and application scenario.

[0076] Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data;

[0077] Step 3: Use the Random Forest supervised learning algorithm to train the risk identification model:

[0078] Using the preprocessed data features to form a data set, wherein using the preprocessed data features to form a data set may include:

[0079] The data characteristics after the agreed preprocessing involve the weak bottom layer X in the geological data weak , fracture zones and faultsX fracture , Permeable formation X permeable 、Earthquake Zone InformationX seismic , Weather X in environmental data weather , Temperature X temperature 、HumidityX humidity , TideX tide Data, construction status information in blasting excavation X blasting 、Tunnel Shotcrete spraying , support construction informationX support ,

[0080] A data set is formed based on data characteristics.

[0081] Use the training set data in the dataset to train the model and adjust the model's hyperparameters through cross-validation to optimize the model performance, which may include:

[0082] Input data set, where the feature vector (X):

[0083] X=[X weak ,X fracture ,X permeable ,X seismic ,X weather,X temperature ,X humidity ,X tide ,X blasting ,X spraying

[0084] ,X support ]

[0085] Target variable(Y):

[0086] Risk level (Y) is a continuous score used to indicate the risk level, such as low, medium, and high.

[0087] Construction of a single decision tree:

[0088]

[0089] in:

[0090] X represents the input feature vector.

[0091] R represents the region of the leaf node.

[0092] c represents the category or predicted value of the leaf node.

[0093] R left and R right Represent the areas of the left branch and the right branch respectively.

[0094] The random forest algorithm is as follows:

[0095]

[0096] in:

[0097] N represents the number of decision trees in the random forest.

[0098] T i Represents the i-th decision tree.

[0099] X represents the input feature vector.

[0100] In addition, the importance of each data feature in the risk identification model is analyzed to determine which features have the greatest impact on risk assessment.

[0101] Set thresholds: Set thresholds to distinguish between normal and abnormal situations.

[0102] Quantify risk: Quantify the risk level through the probability value or classification result output by the risk identification model.

[0103] Step 4: Use the risk identification model to obtain probability values ​​or classification results based on the collected multivariate information, and obtain the corresponding risk level, such as low, medium, or high, based on the probability value or classification result.

[0104] Step 5: Develop corresponding response measures based on different risk levels.

[0105] Rules are pre-defined for risk scenarios, and emergency plans are initiated when the risk level reaches the threshold of the preset rules. A decision tree is used to automatically recommend action plans based on the risk level and specific circumstances.

[0106] The present invention is aimed at the construction of submarine tunnels. It can effectively integrate multi-source information, deeply mine the value of data, and adapt to the changes in complex marine environment to make intelligent decisions, improve the accuracy of identifying adverse geological conditions, optimize construction strategies, reduce construction risks, and ensure the smooth progress of the project.

[0107] Example 2

[0108] The present invention also provides an undersea tunnel information fusion identification and decision-making device based on a domestic server, comprising: a multi-source data acquisition module, a big data fusion processing module, an intelligent risk assessment module and a decision support module.

[0109] The multi-source data acquisition module collects multi-dimensional information of the submarine tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, broken zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information includes blasting excavation, tunnel spraying, and support construction information.

[0110] The big data fusion processing module integrates and analyzes the collected multi-dimensional information:

[0111] Step 21: Perform multivariate information preprocessing,

[0112] Step 22: Extract and select data features based on the preprocessed multivariate information.

[0113] Step 23: Perform data fusion on the selected data features: Use similarity measurement to match the patterns between different data sources; perform consistency check to ensure that the fused data is logically consistent; Use weighted average fusion rules to fuse data according to data type and application scenario.

[0114] Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data;

[0115] The intelligent risk assessment module uses the Random Forest supervised learning algorithm to train the risk identification model: the preprocessed data features are used to form a data set, the training set data in the data set is used to train the model, and the model's hyperparameters are adjusted through cross-validation to optimize the model performance;

[0116] Analyze the importance of each data feature in the risk identification model and determine which features have the greatest impact on risk assessment.

[0117] Set thresholds: Set thresholds to distinguish between normal and abnormal situations.

[0118] Quantify risks: Quantify risk levels through probability values ​​or classification results output by risk identification models;

[0119] The risk identification model is used to obtain probability values ​​or classification results based on the collected multivariate information, and the corresponding risk level is obtained based on the probability values ​​or classification results.

[0120] The decision support module formulates corresponding response measures according to different risk levels.

[0121] As the information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention, the specific contents can be found in the description of the embodiment of the method of the present invention and will not be repeated here.

[0122] Likewise, the benefits of the device of the present invention are:

[0123] (1) Significantly enhance risk prevention and control capabilities: By effectively integrating multiple monitoring data, the speed and accuracy of identifying potential risks in undersea tunnel construction are improved, which helps to take measures to avoid disasters in advance and ensure the safety of personnel and the smooth progress of the project.

[0124] (2) Improving the scientific nature and response speed of decision-making: Intelligent decision support systems can analyze complex data in real time and quickly generate response strategy suggestions, helping managers make reasonable decisions quickly at critical moments, reducing the decision-making cycle and enhancing the effectiveness and flexibility of decisions.

[0125] (3) Reduce costs and improve efficiency: Accurate risk prediction reduces unnecessary preventive construction and repair work, reducing the total cost of the project. At the same time, the system automatically processes large amounts of data, reducing manpower requirements and improving overall work efficiency.

[0126] Promote technological innovation and industrial upgrading: The development and application of the system have promoted the cross-integration of information technology, artificial intelligence and tunnel engineering technology, provided a new technical paradigm for related industries, and promoted technological progress and industrial upgrading of the entire industrial chain.

[0127] (4) Ensure data security: Using domestic servers as infrastructure ensures information security during data processing, complies with the country's data protection requirements for critical infrastructure, and enhances national strategic security.

[0128] (5) Environmental friendliness and sustainable development: Through precise management of construction activities, interference with marine ecology is reduced, promoting the greening and sustainable development of marine engineering projects.

[0129] A display front desk and a management back desk of the device of the present invention can also be set up to carry out information fusion, identification and decision-making. The data center module can be used as a display front desk, playing a central role in data aggregation and intuitive display, providing business personnel with a window to comprehensively review the operation status of the system. Through a highly interactive interface, this module integrates data from different sources, such as real-time construction monitoring, geological analysis reports, environmental monitoring data, etc., to ensure the timeliness and accuracy of the data. In terms of design, the data center uses advanced data visualization technology to display complex statistical analysis results in the form of charts, including but not limited to dynamic line charts showing construction progress and cost changes, heat maps reflecting geological stability distribution, waterfall charts analyzing cost structure, etc.

[0130] Core functions include: Real-time monitoring panel: Centrally display key performance indicators (KPIs), such as tunneling rate, geological stability index, construction equipment status, etc., so that business personnel can quickly grasp the overall situation of the project.

[0131] Data filtering and sorting: Users can customize the data filtering and sorting according to their needs to quickly focus on key information, such as viewing the changing trend of construction efficiency by time series.

[0132] Quick maintenance entry: While displaying data, the data center has a built-in rapid response mechanism. Once data anomalies are found, business personnel can directly enter the corresponding data correction or exception handling process through the embedded link, greatly shortening the problem response time.

[0133] The data center module's data screen can transform massive amounts of data into visual information that is easy to understand and make decisions, serving senior decision makers and project managers. Through the high-definition large screen, core business indicators and key risk warnings are displayed in a high-resolution and dynamic manner. The data screen not only focuses on the overall operational health, but also goes deep into specific business links, such as geological risk level distribution, construction safety indicators, resource utilization, etc. Through colorful charts, dynamic map tracking, real-time updated warning lights and other elements, it ensures the intuitiveness and urgency of information transmission.

[0134] Features of the data screen:

[0135] Customized view configuration: supports customized data display interface according to different concerns of management, such as configuring views of security violation counts and processing progress for security supervisors, and providing cost budget and actual expenditure comparison charts for financial directors.

[0136] Risk warning prompts: Integrated intelligent algorithms can warn of potential abnormal construction risks, and display them in real time on the large screen with eye-catching colors and icons to encourage quick response.

[0137] Cross-screen linkage: Data linkage is realized between the data screen and the data center and other modules. Clicking a specific area on the screen can jump to the detailed data page for in-depth analysis.

[0138] The management backend can be used for data maintenance to ensure that the system data is complete, accurate and timely. It supports multi-dimensional data management functions, including data import and export, data verification, data cleaning, data backup and recovery, etc. This module uses automated tools and scripts to reduce manual intervention and improve the efficiency and quality of data processing.

[0139] Core Competencies:

[0140] Intelligent data verification: Introduce a rule engine to automatically detect data integrity, logical consistency, and format correctness, discover and mark abnormal data, and provide repair suggestions or automatic repairs.

[0141] Data lifecycle management: full process management from data entry to abandonment, supporting data version control to ensure that every data change can be traced, facilitating the query and reuse of historical data.

[0142] Data permission management: Refine data access permissions to ensure that users in different roles can only access data within their scope of responsibility and maintain data security.

[0143] The algorithm center of the management background can be used as an intelligent engine, supporting multiple algorithm frameworks, making it easy for algorithm engineers to quickly iterate and optimize models.

[0144] Main functions of the algorithm center:

[0145] Algorithm development and debugging: Provides an integrated development environment (IDE) to support algorithm code writing, debugging, and performance testing, accelerating the transformation of algorithms from the laboratory to production.

[0146] Model management and version control: Implement version management of algorithm models, support model comparison, selection, and rollback, and ensure the optimal model online service.

[0147] Algorithm performance monitoring: Real-time monitoring of algorithm operation status, including processing speed, accuracy, resource usage, etc., automatically alarms when performance degrades or becomes abnormal, ensuring the stability and reliability of the decision-making system.

[0148] The management background is responsible for the daily operation and maintenance and security management of the device, covering user permissions, system configuration, log auditing and other aspects, and is the key to ensuring the stable and efficient operation of the device.

[0149] The management backend can achieve:

[0150] User and permission management: Detailed user role division, supporting multi-level permission settings, ensuring that each user can only access and operate system functions that match their responsibilities.

[0151] Menu and interface configuration: allows administrators to customize system menus and interface layouts according to organizational structure and business needs to improve user experience.

[0152] System log and audit: Automatically record all system operation logs, including login records, operation records, exception records, etc., to facilitate audit tracking and troubleshooting, while providing evidence support for system security.

[0153] Security policy management: Integrate advanced security protection measures, such as data encryption transmission, regular security scanning, access control lists, etc., to ensure the security of the system and data.

[0154] It should be noted that not all steps and modules in the above-mentioned processes and device structures are necessary, and some steps or modules can be ignored according to actual needs. The execution order of each step is not fixed and can be adjusted as needed. The system structure described in the above-mentioned embodiments can be a physical structure or a logical structure, that is, some modules may be implemented by the same physical entity, or some modules may be implemented by multiple physical entities, or some components in multiple independent devices may be implemented together.

[0155] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A method for identifying and making decisions on submarine tunnel information fusion based on a domestic server, characterized by: include: Step 1: Collect multivariate information of the submarine tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, fractured zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information includes blasting excavation, tunnel spraying, and support construction information. Step 2: Integrate and analyze the collected multi-dimensional information: Step 21: Perform multivariate information preprocessing, Step 22: Extract and select data features based on the preprocessed multivariate information. Step 23: Perform data fusion on the selected data features: Use similarity metrics to match patterns between different data sources; Perform consistency checks to ensure that the fused data is logically consistent; use weighted average fusion rules to fuse data based on data type and application scenario. Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data; Step 3: Use the Random Forest supervised learning algorithm to train the risk identification model: Use the preprocessed data features to form a data set, use the training set data in the data set to train the model, and adjust the model's hyperparameters through cross-validation to optimize the model performance; Analyze the importance of each data feature in the risk identification model and determine which features have the greatest impact on risk assessment. Set thresholds: Set thresholds to distinguish between normal and abnormal situations. Quantify risks: Quantify risk levels through probability values ​​or classification results output by risk identification models; Step 4: Use the risk identification model to obtain probability values ​​or classification results based on the collected multivariate information, and obtain the corresponding risk level based on the probability value or classification result. Step 5: Develop corresponding response measures based on different risk levels.

2. According to claim 1, a method for identifying and making decisions on information fusion of submarine tunnels based on a domestic server is characterized by: In step 21, multivariate information preprocessing is performed, including: Data cleaning: remove invalid values, fill or remove missing values, detect and process outliers, Data conversion: Standardize or normalize raw data to ensure compatibility between different data sources. Data integration: Integrate data from multiple data sources to resolve data redundancy and conflict issues.

3. According to claim 1, a method for identifying and making decisions on submarine tunnel information fusion based on a domestic server is characterized by: In step 22, data feature extraction and selection are performed, including: when performing feature extraction, meaningful data features are extracted from the original data, and the principal component analysis PCA method is used to reduce the data dimension. When performing feature selection, a filtering method is used to select the most representative data features for subsequent analysis.

4. According to claim 1, a method for identifying and making decisions on information fusion of submarine tunnels based on a domestic server is characterized by: In step 3, the preprocessed data features are used to form a data set, including: The agreed preprocessed data features include the weak bottom layer, broken zone and fault, permeable stratum, seismic zone information in geological data, weather, temperature, humidity, tide data in environmental data, blasting excavation, tunnel spraying, support construction information in construction status information, A data set is formed based on data characteristics.

5. According to claim 1, a method for identifying and making decisions on submarine tunnel information fusion based on a domestic server is characterized by: In step 5, rules are predefined for risk scenarios. When the risk level reaches the threshold of the preset rules, the emergency plan is activated, and a decision tree is used to automatically recommend an action plan based on the risk level and specific circumstances.

6. An information fusion identification and decision-making device for submarine tunnels based on a domestic server, characterized by: include: Multi-source data acquisition module, big data fusion processing module, intelligent risk assessment module and decision support module, The multi-source data acquisition module collects multi-dimensional information of the submarine tunnel, including geological data, environmental data, and construction status information. The geological data covers information on weak substrata, broken zones and faults, permeable strata, and seismic zones. The environmental data involves weather, temperature, humidity, and tidal data. The construction status information includes blasting excavation, tunnel spraying, and support construction information. The big data fusion processing module integrates and analyzes the collected multi-dimensional information: Step 21: Perform multivariate information preprocessing, Step 22: Extract and select data features based on the preprocessed multivariate information. Step 23: Perform data fusion on the selected data features: Use similarity metrics to match patterns between different data sources; Perform consistency checks to ensure that the fused data is logically consistent; use weighted average fusion rules to fuse data based on data type and application scenario. Step 24: Analyze the fused data: Perform descriptive statistical analysis on the fused data to understand the basic attributes of the fused data, and apply predictive analysis to reveal the potential connections between the fused data; The intelligent risk assessment module uses the Random Forest supervised learning algorithm to train the risk identification model: the preprocessed data features are used to form a data set, the training set data in the data set is used to train the model, and the model's hyperparameters are adjusted through cross-validation to optimize the model performance; Analyze the importance of each data feature in the risk identification model and determine which features have the greatest impact on risk assessment. Set thresholds: Set thresholds to distinguish between normal and abnormal situations. Quantify risks: Quantify risk levels through probability values ​​or classification results output by risk identification models; The risk identification model is used to obtain probability values ​​or classification results based on the collected multivariate information, and the corresponding risk level is obtained based on the probability values ​​or classification results. The decision support module formulates corresponding response measures according to different risk levels.

7. According to claim 6, a submarine tunnel information fusion identification and decision-making device based on a domestic server is characterized by big data The fusion processing module executes step 21 to perform multivariate information preprocessing, including: Data cleaning: remove invalid values, fill or remove missing values, detect and process outliers, Data conversion: Standardize or normalize raw data to ensure compatibility between different data sources. Data integration: Integrate data from multiple data sources to resolve data redundancy and conflict issues.

8. According to claim 6, a submarine tunnel information fusion identification and decision-making device based on a domestic server is characterized by big data The fusion processing module executes step 22 to extract and select data features, including: when extracting features, extracting meaningful data features from the original data, using the principal component analysis PCA method to reduce the data dimension, When performing feature selection, a filtering method is used to select the most representative data features for subsequent analysis.

9. The device for identifying and making decisions on submarine tunnel information fusion based on a domestic server according to claim 6 is characterized by: The analytical risk identification model uses the preprocessed data features to form a data set, including: The agreed preprocessed data features include the weak bottom layer, broken zone and fault, permeable stratum, seismic zone information in geological data, weather, temperature, humidity, tide data in environmental data, blasting excavation, tunnel spraying, support construction information in construction status information, A data set is formed based on data characteristics.

10. The device for identifying and making decisions on submarine tunnel information fusion based on a domestic server according to claim 6 is characterized by: The decision support module predefines rules for risk scenarios, activates emergency plans when the risk level reaches the threshold of the preset rules, and uses a decision tree to automatically recommend action plans based on the risk level and specific circumstances.