A tunnel portal slope risk identification and analysis system based on artificial intelligence

Abstract

The application discloses a tunnel portal slope risk identification and analysis system based on artificial intelligence, relates to the field of data processing, and comprises a slope risk identification system, wherein a risk factor investigation and collection module, a portal risk simulation processing module, a tunnel risk simulation processing module, a risk factor joint analysis module and a slope risk identification and analysis module are arranged; the risk factor investigation and collection module is used for collecting data information of a tunnel construction area; the portal risk simulation processing module is used for constructing a slope damage model according to obtained data information; the tunnel risk simulation processing module is used for constructing a support damage model according to obtained data information; the risk factor joint analysis module is used for generating a risk influence database; and the slope risk identification and analysis module is used for obtaining corresponding risk identification data; and the application improves the accuracy in the risk identification and analysis process to a certain extent.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an artificial intelligence-based system for identifying and analyzing the risks of tunnel entrance slopes. Background Technology

[0002] During tunnel construction, the tunnel entrance inevitably experiences the removal of soil and rock at the slope toe due to the excavation of the side slopes and tunnel body, creating an unfavorable open face. This disrupts the original equilibrium of the slope, causing instability of the entrance and exit slopes, which can lead to the tunnel entrance being covered and compressed. This poses a serious threat to the safety of tunnel construction and operation. Therefore, identifying and analyzing the risks of tunnel entrance slopes can improve the stability of tunnel construction to a certain extent.

[0003] A search revealed that existing tunnel portal risk identification methods all involve dividing the tunnel portal into regions, monitoring and analyzing the resulting regions, and then performing multiple divisions based on the monitoring results to determine the final risk area. However, this process of multiple region divisions and monitoring of the tunnel portal results in some resource waste, and the risk area may be segmented during these multiple divisions, thus affecting the accuracy of risk identification. Therefore, how to avoid the resource waste and inaccuracy caused by region division in tunnel portal risk identification is a problem we need to solve. To this end, this invention proposes an artificial intelligence-based tunnel portal slope risk identification and analysis system. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that rely too heavily on a single slope parameter, resulting in insufficient accuracy in risk identification. Therefore, this invention proposes an artificial intelligence-based tunnel entrance slope risk identification and analysis system.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] An artificial intelligence-based tunnel portal slope risk identification and analysis system includes a slope risk identification system, which comprises a risk factor investigation and collection module, a portal risk simulation and processing module, a tunnel risk simulation and processing module, a risk factor joint analysis module, and a slope risk identification and analysis module.

[0007] The risk factor investigation and collection module is used to collect geological feature data and construction monitoring data corresponding to the tunnel construction area, as well as risk factor data and risk event data of the corresponding tunnel portal slope;

[0008] The tunnel entrance risk simulation processing module is used to set up a physical model of the tunnel entrance slope based on the corresponding geological feature data, risk factor data and risk event data, obtain the corresponding slope deformation and failure feature data, and construct a slope failure model.

[0009] The tunnel risk simulation processing module is used to set up a physical model of the tunnel entrance based on the corresponding geological feature data, risk factor data and risk event data, obtain the corresponding support deformation and failure feature data, and construct a support failure model.

[0010] The risk factor joint analysis module is used to set up a three-dimensional construction model based on the geological feature data and construction monitoring data corresponding to the tunnel construction area, map the slope failure model and support failure model into the three-dimensional construction model, set risk feature sequences at various locations in the three-dimensional construction model, analyze and process the data information corresponding to the corresponding models in the risk feature sequences, obtain the corresponding joint influence coefficients, and construct a risk influence database based on the joint influence coefficients and the corresponding models.

[0011] The slope risk identification and analysis module is used to traverse the corresponding three-dimensional construction model according to the construction monitoring data, compare and analyze the risk impact database corresponding to the corresponding risk feature sequence, integrate the analysis results of each risk feature sequence, and obtain risk identification data.

[0012] The above technical solution further includes: the process of collecting geological feature data and construction monitoring data corresponding to the tunnel construction area includes:

[0013] A monitoring and acquisition unit is set up, which includes a geological monitoring terminal and multiple monitoring and acquisition terminals.

[0014] Geological feature data corresponding to the tunnel construction area are collected through geological monitoring terminals;

[0015] The monitoring data collected by the monitoring and acquisition terminal includes environmental monitoring data, working condition monitoring data, structural safety monitoring data, and equipment status monitoring data within the corresponding tunnel construction area.

[0016] Furthermore, the process of collecting risk factor data and risk event data for the corresponding tunnel portal slope includes:

[0017] Set up survey and data collection units to collect survey data within the corresponding tunnel engineering area, and set up risk factor collection terminals and risk event collection terminals;

[0018] The risk factor collection terminal analyzes and processes the survey data based on artificial intelligence algorithms to obtain corresponding feature keywords. It then uses NER technology to extract and identify information from the obtained feature keywords to obtain corresponding risk factor data.

[0019] The risk event acquisition terminal identifies risk event types based on natural language processing algorithms, obtains feature datasets corresponding to the respective risk event types, analyzes and processes the obtained feature datasets sequentially using clustering analysis algorithms, obtains risk factor data evaluation indicators corresponding to the respective risk event types, and marks the obtained data as risk event data.

[0020] Furthermore, the process of constructing a slope failure model includes:

[0021] Set up a slope test simulation unit and a slope test treatment unit;

[0022] The corresponding geological feature data were obtained through the slope test simulation unit, and a three-dimensional image of the virtual tunnel entrance slope structure was constructed. The three-dimensional image of the virtual tunnel entrance slope structure was sliced ​​to obtain the corresponding slope structure slice image. The corresponding similar material mixing ratio was obtained for each pixel in the slope structure slice image based on the similarity ratio and dimensional analysis method. The physical model of the tunnel entrance slope was set according to the similar material mixing ratio of each pixel.

[0023] Obtain relevant risk factor data and historical construction monitoring data, and analyze and process the obtained data to obtain relevant slope condition information;

[0024] According to the type of slope working condition information, set the corresponding working condition variable data in sequence, combine the working condition variable data corresponding to different slope working conditions in sequence, obtain the corresponding slope simulation test data set, and perform numerical simulation on each slope simulation test data set based on FLAC3D numerical simulation software to obtain the corresponding test data information.

[0025] The slope test processing unit acquires and analyzes the test data information corresponding to the corresponding slope simulation test data group, sets up the corresponding test dataset according to the corresponding slope simulation test data group, compares and analyzes the corresponding test data information in the test dataset with the risk event data, and obtains the corresponding slope deformation and failure characteristic data.

[0026] The obtained slope deformation and failure characteristic data are analyzed and trained based on deep learning algorithms to construct corresponding slope failure models.

[0027] Furthermore, the process of constructing a support failure model includes:

[0028] Set up a tunnel test simulation unit and a tunnel test processing unit;

[0029] The corresponding geological feature data are obtained through the tunnel test simulation unit. The corresponding tunnel physical model is constructed according to the construction method of the tunnel entrance slope physical model. The corresponding tunnel working condition information is obtained according to the risk factor data and historical construction monitoring data.

[0030] Based on the tunnel working condition information, a tunnel simulation test data group is set up, and numerical simulation is performed on the corresponding tunnel simulation test data group using FLAC3D numerical simulation software to obtain the corresponding test data information.

[0031] The tunnel test processing unit compares and analyzes the test data information of the corresponding tunnel simulation test data group with the corresponding risk event data to obtain the corresponding support deformation failure characteristic data. The obtained support deformation failure characteristic data is then analyzed and trained based on a deep learning algorithm to construct the corresponding support failure model.

[0032] Furthermore, the process of setting up a three-dimensional construction model includes:

[0033] Set up a model analysis unit; obtain relevant geological feature data through the model analysis unit, construct an initial three-dimensional construction model, map the obtained construction monitoring data into the initial three-dimensional construction model for update processing, obtain the corresponding three-dimensional construction model, map the slope failure model and support failure model into the three-dimensional construction model according to their respective positions, and obtain the real-time three-dimensional construction model.

[0034] Furthermore, the process of building a risk impact database includes:

[0035] The real-time 3D construction model is intelligently sliced ​​to obtain corresponding tunnel entrance slice images, and each tunnel entrance slice image is marked.

[0036] Risk feature sequences are set according to the corresponding pixels in the slice images of each tunnel entrance. Within each risk feature sequence, corresponding risk feature sub-sequences are set according to the test data information corresponding to the slope simulation test data group and tunnel simulation test data group in the slope failure model and support failure model. The obtained risk feature sub-sequences are then combined horizontally, vertically, and comprehensively.

[0037] The horizontal combination refers to the combination of test data information corresponding to each risk feature subsequence within each risk feature sequence in the corresponding tunnel portal slice image under different working conditions.

[0038] The longitudinal combination refers to the combination of test data information corresponding to each risk feature subsequence in the risk feature sequence at the corresponding position in the slice image of different tunnel entrances under different working conditions.

[0039] The comprehensive combination is the result of combining the test data information corresponding to each risk feature subsequence in the risk feature sequence at different locations within different tunnel portal slice images under different working conditions.

[0040] For the test data information corresponding to the slope simulation test data group and tunnel simulation test data group within the horizontal combination, vertical combination and comprehensive combination, set up corresponding combination datasets, perform joint analysis on the obtained combination datasets, and obtain the joint influence coefficient between the risk event data corresponding to different working conditions at the corresponding locations;

[0041] Based on the obtained joint impact coefficients, risk impact data tables are set up according to the corresponding risk characteristic sequences in the real-time three-dimensional construction model.

[0042] The risk impact data tables within each risk characteristic sequence are integrated to construct a risk impact database.

[0043] Furthermore, the process of acquiring risk identification data includes:

[0044] Acquire construction monitoring data and map the latest acquired construction monitoring data into the real-time three-dimensional construction model. The corresponding risk feature sequence in the real-time three-dimensional construction model obtains the corresponding working condition information based on the corresponding construction monitoring data. The obtained working condition information is then input into the corresponding slope failure model and support failure model to obtain the corresponding slope deformation failure feature data and tunnel deformation failure feature data. The obtained feature data is then marked as sequence identification data.

[0045] Based on the joint impact coefficients in the risk impact database, the risk characteristic data in each risk characteristic sequence are verified and analyzed to obtain the corresponding verification and identification data.

[0046] The sequence identification data corresponding to each risk feature sequence is compared and analyzed with the verification identification data to obtain the corresponding verification accuracy. Based on the corresponding verification accuracy, the risk identification data corresponding to the corresponding risk feature sequence is obtained.

[0047] The present invention has the following beneficial effects:

[0048] 0. In this invention, a real-time three-dimensional construction model is set up to dynamically manage the risk identification process according to the tunnel construction progress. In addition, by setting corresponding risk feature sequences at different locations within the real-time three-dimensional construction model, the working condition information corresponding to different types and variables within different risk feature sequences is combined horizontally, vertically, and comprehensively to obtain the joint influence coefficient between different risk feature sequences. Based on the real-time three-dimensional construction model and the joint influence coefficient, a risk impact database corresponding to the risk event data is constructed. Based on the risk impact database, multiple verifications are performed from multiple dimensions, including the sequence identification data of the risk feature sequence at the corresponding location and the verification identification data of other risk feature sequences, thereby improving the accuracy of the risk identification process to a certain extent.

[0049] 1. In this invention, by setting up corresponding physical models of the tunnel entrance slope and tunnel entrance within the corresponding tunnel construction area, and conducting simulation tests based on the corresponding physical models, slope failure models and support failure models corresponding to different risk events are set up according to the test data information corresponding to the simulation test results. Moreover, the corresponding slope failure models and support failure models are set according to the corresponding tunnel construction area, thereby improving the flexibility of the tunnel entrance slope risk identification and analysis system. In addition, by constructing the corresponding slope failure models and support failure models through simulation test analysis methods, the cost of constructing the risk identification system is reduced to a certain extent. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the structure of an artificial intelligence-based tunnel entrance slope risk identification and analysis system proposed in this invention;

[0051] Figure 2 This is a flowchart illustrating an artificial intelligence-based tunnel entrance slope risk identification and analysis system proposed in this invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Example 1

[0054] like Figure 1As shown, the present invention proposes an artificial intelligence-based tunnel portal slope risk identification and analysis system, which includes a slope risk identification system. The slope risk identification system includes a risk factor investigation and collection module, a portal risk simulation and processing module, a tunnel risk simulation and processing module, a risk factor joint analysis module, and a slope risk identification and analysis module.

[0055] In this embodiment, the slope risk identification system is used to identify risks associated with the stability control of the tunnel entrance slope in water-rich weathered granite gneiss tunnels and the corresponding tunnel entrance slope during tunnel construction, thereby improving safety during tunnel construction. Its specific implementation process includes, for example: Figure 2 As shown, it includes the following steps:

[0056] Step S1: Collect geological feature data and construction monitoring data corresponding to the tunnel construction area, as well as risk factor data and risk event data of the corresponding tunnel portal slope;

[0057] Step S2: Based on the corresponding geological feature data, risk factor data, and risk event data, set up a physical model of the tunnel entrance slope, obtain the corresponding slope deformation and failure characteristic data, and construct a slope failure model;

[0058] Step S3: Based on the corresponding geological feature data, risk factor data, and risk event data, set up the physical model of the tunnel entrance, obtain the corresponding support deformation and failure characteristic data, and construct the support failure model;

[0059] Step S4: Set up a three-dimensional construction model based on the geological feature data and construction monitoring data corresponding to the tunnel construction area, and map the slope failure model and support failure model into the three-dimensional construction model;

[0060] Step S5: Set risk feature sequences at various locations within the 3D construction model, analyze and process the data information corresponding to the models within the risk feature sequences, obtain the corresponding joint impact coefficients, and construct a risk impact database based on the joint impact coefficients and the corresponding models.

[0061] Step S6: Based on the construction monitoring data, traverse the corresponding three-dimensional construction models, compare and analyze the risk impact database corresponding to the corresponding risk feature sequence, verify the analysis results of each risk feature sequence, and obtain risk identification data.

[0062] The risk factor investigation and data collection module is used to collect geological feature data and construction monitoring data corresponding to the tunnel construction area, as well as risk factor data and risk event data of the corresponding tunnel portal slope. Its specific implementation process includes:

[0063] Set up monitoring and data collection units and survey and data collection units;

[0064] The monitoring and acquisition unit is equipped with a geological monitoring terminal and multiple monitoring and acquisition terminals, which are used to collect geological feature data and construction monitoring data corresponding to the corresponding tunnel engineering area.

[0065] The geological monitoring terminal is used to acquire geological feature data corresponding to the tunnel construction area. The geological feature data includes data related to tunnel construction, such as topographic and geomorphological feature data and geological structural feature data.

[0066] The monitoring and data acquisition terminals include environmental monitoring equipment, operating condition monitoring equipment, structural safety monitoring equipment, and equipment status monitoring equipment. These terminals are used to remotely and intelligently manage the data acquisition process based on artificial intelligence technology.

[0067] Environmental monitoring equipment is used to acquire meteorological and air quality monitoring data corresponding to the tunnel construction area;

[0068] The working condition monitoring equipment is used to acquire the corresponding construction working condition data within the tunnel construction area;

[0069] Structural safety monitoring equipment is used to acquire stress-strain data and displacement monitoring data at corresponding locations during tunnel construction.

[0070] Equipment condition monitoring equipment is used to acquire relevant vibration signal data and equipment condition data;

[0071] The monitoring data obtained from the corresponding monitoring and acquisition terminals are analyzed and processed to obtain the corresponding construction monitoring data. The monitoring risk factor data includes the corresponding meteorological monitoring data, air quality monitoring data, construction condition data, stress and strain data, displacement monitoring data, vibration signal data, and equipment status data.

[0072] The survey and data collection unit is used to survey and collect survey data corresponding to the tunnel engineering area, and to obtain relevant survey data provided by the geological exploration party, design party, construction party, supervision party and construction party of the tunnel in the engineering area;

[0073] Risk factor collection terminals and risk event collection terminals are set up separately in the survey and collection unit;

[0074] The risk factor collection terminal is used to perform risk factor analysis on the relevant survey data obtained, and the process includes:

[0075] Obtain relevant survey data, extract keywords and feature information from the obtained survey data based on the natural language processing algorithm corresponding to artificial intelligence technology, and pre-set a feature keyword database corresponding to the corresponding type of feature data;

[0076] Based on the feature keyword database corresponding to the corresponding feature type, keywords are extracted using the TF-IDF algorithm, and the corresponding feature keywords are obtained based on the keyword extraction results.

[0077] The obtained feature keywords are extracted and identified using NER technology, and the corresponding feature data is obtained based on the information extraction and identification results.

[0078] Based on the feature data obtained from the corresponding feature keyword database type, the obtained feature data is marked as the corresponding risk factor data. The risk factor data includes the corresponding climate feature risk factor data, geological feature risk factor data, and construction feature risk factor data.

[0079] The risk event collection terminal is used to analyze relevant survey data to identify risk events. The process includes:

[0080] The relevant survey data obtained are used to identify risk event types using natural language processing algorithms within artificial intelligence technology, and feature datasets corresponding to the respective risk event types are obtained.

[0081] The obtained feature dataset is evaluated using intensity criteria. The corresponding risk event types in the feature dataset are labeled as sample data. The corresponding risk factor data types in the obtained sample data are comprehensively sorted to obtain the feature data subset corresponding to the sample data.

[0082] The obtained feature data subset is analyzed and processed using a clustering analysis algorithm to obtain the risk factor data evaluation indicators corresponding to the risk event types. The process includes:

[0083] The evaluation indicators corresponding to each risk factor data within the feature data subset are integrated and processed to obtain the intensity level corresponding to the corresponding risk event type. The intensity level includes the evaluation indicator range corresponding to the corresponding risk factor data.

[0084] The obtained risk event types and corresponding evaluation indicators are labeled to obtain the corresponding risk event data.

[0085] The portal risk simulation processing module is used to set up a physical model of the portal slope based on relevant geological feature data, risk factor data, and risk event data, obtain relevant slope deformation and failure characteristic data, and construct a slope failure model. Its specific implementation process includes:

[0086] Set up a slope test simulation unit and a slope test treatment unit;

[0087] The slope test simulation unit is used to set up a physical model test of the tunnel entrance slope and obtain corresponding test data. The process includes:

[0088] Geological feature data of the corresponding tunnel portal section is obtained, and a three-dimensional image of the virtual tunnel portal slope structure is constructed based on the geological feature data. The three-dimensional image of the virtual tunnel portal slope structure is sliced ​​to obtain the corresponding slope structure slice image.

[0089] Each pixel in the obtained slope structure slice image is marked, and each pixel is analyzed and processed according to the marking and processing results. The soil and rock mass at the corresponding pixel in each slope structure slice image is analyzed and processed in turn.

[0090] The analysis and processing are performed based on the properties of the soil and rock materials corresponding to the corresponding pixels, and the corresponding soil and rock material database is set up according to the corresponding properties of the soil and rock materials.

[0091] Set the corresponding pixel dimension data for the soil and rock mass at the corresponding pixel point according to the soil and rock material database.

[0092] Based on the pixel dimension data corresponding to the corresponding pixel point, similarity material matching analysis is performed to obtain the material mixing dimension data of the corresponding test material. The material mixing dimension data is the dimension data corresponding to the similar material mix ratio determined by mechanical test.

[0093] The material mixing dimension data corresponding to the pixel dimension data at each pixel point are integrated to construct the corresponding physical model of the tunnel entrance slope in sequence.

[0094] The obtained physical model of the tunnel entrance slope was analyzed and processed. The test process was set up sequentially for the obtained physical model of the tunnel entrance slope to obtain the corresponding test data information.

[0095] Obtain the physical model of the tunnel entrance slope, and conduct a damage simulation experiment based on the material information corresponding to the corresponding pixel in the physical model of the tunnel entrance slope.

[0096] Set slope working condition information, which includes working conditions such as slope cutting, tunnel construction, and climate change;

[0097] Based on the corresponding slope working condition information, the corresponding pixels in the physical model of the corresponding opening slope are traversed sequentially. With the corresponding pixel as the center, the corresponding working condition variable data are set according to the slope working condition information. The working condition variable data is the quantitative data corresponding to the corresponding type of slope working condition information, and each quantitative data increases sequentially.

[0098] Based on the corresponding pixel points and the corresponding slope working condition information, simulation experimental data are obtained by performing simulation experiments on the corresponding slope simulation experimental data. The obtained slope simulation experimental data are then used to perform numerical simulations on numerical simulation software such as FLAC3D and Geo-slope2D to obtain the corresponding experimental data information.

[0099] The slope test processing unit is used to analyze and process the obtained test data to construct a slope failure model. The process includes:

[0100] Obtain the relevant experimental data information, and analyze and process the experimental data information corresponding to the slope working conditions of the corresponding pixels in the physical model of the tunnel entrance slope and the corresponding working condition variables in the slope working conditions information. Set the corresponding experimental dataset, experimental data subset, and experimental data elements respectively, where:

[0101] The test datasets correspond to the test data subsets corresponding to the corresponding pixels in the physical model of the tunnel entrance slope;

[0102] The subsets of test data correspond to test data elements for different slope conditions within a pixel.

[0103] The test data elements correspond to the test data information corresponding to different working condition variable data of the corresponding type of slope working condition information;

[0104] Outlier analysis was performed on the obtained experimental dataset, experimental data subset, and corresponding experimental data elements based on machine learning models, and the corresponding outliers were filtered out.

[0105] Based on the experimental dataset, experimental data subsets and corresponding experimental data elements obtained by filtering, data processing is performed to obtain slope deformation and failure characteristic data of corresponding experimental data elements in each experimental data subset within the corresponding experimental dataset, as well as slope deformation and failure characteristic data of corresponding experimental data elements between each experimental data subset.

[0106] Based on the obtained slope deformation and failure characteristic data, deep learning algorithms were used to analyze and train the data to construct corresponding slope failure models.

[0107] It should be further explained that, in the specific implementation process, the slope failure model is verified and analyzed based on real tunnel portal slope events to determine whether the corresponding slope failure model conforms to the occurrence process of tunnel portal slope events. If it does, the corresponding slope failure model is saved; otherwise, the slope failure model is retrained.

[0108] The tunnel risk simulation processing module is used to set up a physical model of the tunnel entrance based on relevant geological feature data, risk factor data, and risk event data, obtain relevant support deformation and failure characteristic data, and construct a support failure model. Its specific implementation process includes:

[0109] Set up a tunnel test simulation unit and a tunnel test processing unit;

[0110] The tunnel test simulation unit acquires risk factor data for the excavation of the tunnel portal section, sets up a physical model test of the tunnel portal section, and obtains corresponding test numerical data.

[0111] The process involves acquiring risk factor data for the excavation of the tunnel portal section, obtaining the physical model of the portal slope based on the slope risk simulation processing module, and deriving the similarity ratio of the tunnel physical model based on similarity theory and dimensional analysis based on the corresponding risk factor data.

[0112] Based on the similarity ratio derivation results of the corresponding tunnel portal section model, the mix ratio of the similar materials corresponding to each pixel point of the tunnel portal section physical model is obtained. The mix ratios of the similar materials corresponding to each pixel point are then integrated to construct the corresponding tunnel portal section physical model.

[0113] The obtained physical model of the tunnel entrance section is subjected to experimental simulation analysis to obtain tunnel working condition information involved in the tunnel excavation process. The tunnel working condition information includes various tunnel working condition types such as the corresponding entrance slope collapse, tunnel construction, climate change, and support type.

[0114] Based on the corresponding tunnel working condition information, a corresponding tunnel simulation test data set is set up for the physical model of the tunnel entrance. The obtained corresponding tunnel simulation test data set is used for numerical simulation based on numerical simulation software such as FLAC3D and Geo-slope2D to obtain the corresponding test data information.

[0115] The tunnel test processing unit is used to compare and analyze the test data information of the corresponding tunnel simulation test data group with the corresponding risk event data to obtain the corresponding support deformation and failure characteristic data.

[0116] The obtained support deformation failure feature data are analyzed and trained based on deep learning algorithms to construct a corresponding support failure model;

[0117] It should be further explained that, in the specific implementation process, the support failure model is verified and analyzed based on real tunnel portal slope events to determine whether the corresponding support failure model conforms to the occurrence process of tunnel portal slope events. If it does, the corresponding support failure model is saved; otherwise, the support failure model is retrained.

[0118] The risk factor joint analysis module is used to set up a three-dimensional construction model based on the geological feature data and construction monitoring data corresponding to the tunnel construction area, map the slope failure model and support failure model into the three-dimensional construction model, set risk feature sequences at various locations within the three-dimensional construction model, analyze and process the data information corresponding to the models within the risk feature sequences, obtain the corresponding joint influence coefficients, and construct a risk influence database based on the joint influence coefficients and the corresponding models. Its specific implementation process includes:

[0119] Set up model analysis units and joint analysis units;

[0120] The model analysis unit is used to acquire relevant geological feature data, construct an initial three-dimensional construction model within the tunnel construction area, map the acquired construction monitoring data into the initial three-dimensional construction model for update processing, acquire the corresponding three-dimensional construction model, and map the slope failure model and support failure model into the three-dimensional construction model according to their respective locations to acquire a real-time three-dimensional construction model.

[0121] It should be further explained that, in the specific implementation process, the real-time three-dimensional construction model is obtained according to the construction progress within the tunnel construction area. The initial three-dimensional construction model is obtained based on the analyzed tunnel construction monitoring data, while the real-time three-dimensional construction model is obtained based on the unanalyzed tunnel construction monitoring data. The corresponding slope failure model and support failure model are updated and verified in real time according to the corresponding construction monitoring process. Based on artificial intelligence algorithms, the corresponding deep learning algorithms are used to train, verify, analyze and process the corresponding slope failure model and support failure model, thereby improving the accuracy of the slope risk identification process within the corresponding tunnel construction area.

[0122] The joint analysis unit is used to obtain the corresponding joint impact coefficients and construct a risk impact database. The process includes:

[0123] The real-time 3D construction model is intelligently sliced ​​to obtain corresponding tunnel entrance slice images, and each tunnel entrance slice image is marked.

[0124] Risk feature sequences were set based on the corresponding pixels in the slice images of each tunnel entrance, and these sequences were marked as follows: ,in, For the corresponding risk feature sequence, P is the corresponding marked tunnel portal slice image;

[0125] Within the risk characteristic sequence, corresponding risk characteristic sub-sequences are set based on the test data information corresponding to the slope failure model and the tunnel simulation test data group within the slope failure model and the support failure model, respectively. The slope simulation test data group is marked as... The corresponding experimental data information is marked as , , The tunnel simulation test data set was marked as... The corresponding experimental data information is marked as , , The corresponding risk feature subsequences are labeled as follows: and ;

[0126] The obtained risk feature subsequences are combined horizontally, vertically, and comprehensively, respectively.

[0127] The horizontal combination refers to the combination of test data information corresponding to each risk feature subsequence within each risk feature sequence in the corresponding tunnel portal slice image under different working conditions. The corresponding combination results are marked as follows: and And label the corresponding combined datasets as ;

[0128] The longitudinal combination refers to the combination of experimental data information corresponding to each risk feature subsequence within the risk feature sequence at corresponding positions in different tunnel portal slice images under different working conditions. The corresponding combination results are marked as follows: And label the corresponding combined datasets as ;

[0129] The comprehensive combination refers to the combination of experimental data information corresponding to each risk feature subsequence within the risk feature sequence at different locations within different tunnel portal slice images under different working conditions. The corresponding combination results are marked as follows: And label the corresponding combined datasets as ;

[0130] The combined datasets corresponding to the slope simulation test data sets and tunnel simulation test data sets within the horizontal combination, vertical combination, and comprehensive combination are jointly analyzed to obtain the joint influence coefficients between the risk event data corresponding to different working conditions at the corresponding locations.

[0131] The corresponding joint impact coefficient is denoted as F, where:

[0132] ,in, and These are the weighting coefficients for the corresponding experimental data information within the respective combined datasets. These are the coefficients of the corresponding error terms;

[0133] ,in These are the weighting coefficients for the vertical combination results of the corresponding risk feature sequences. These are the coefficients of the corresponding error terms;

[0134] ,in , and These are the weight coefficients within the corresponding combined dataset. These are the coefficients of the corresponding error terms;

[0135] Based on the obtained joint impact coefficients, risk impact data tables are set up according to the corresponding risk characteristic sequences in the real-time three-dimensional construction model.

[0136] The risk impact data tables within each risk characteristic sequence are integrated to construct a risk impact database.

[0137] The slope risk identification and analysis module is used to traverse the corresponding three-dimensional construction models based on construction monitoring data, compare and analyze the risk impact database corresponding to the corresponding risk feature sequence, verify the analysis results at each risk feature sequence, and obtain risk identification data. Its specific implementation process includes:

[0138] Set up a risk identification unit and a risk verification unit;

[0139] The risk identification unit is used to obtain the identification and analysis results in the risk impact database corresponding to each risk feature sequence. The process includes:

[0140] Acquire construction monitoring data and map the latest acquired construction monitoring data into the real-time three-dimensional construction model. The corresponding risk feature sequence in the real-time three-dimensional construction model obtains the corresponding working condition information based on the corresponding construction monitoring data. The obtained working condition information is then input into the corresponding slope failure model and support failure model to obtain the corresponding slope deformation failure feature data and tunnel deformation failure feature data. The obtained feature data is then marked as sequence identification data.

[0141] Based on the joint impact coefficients in the risk impact database, the risk characteristic data in each risk characteristic sequence are verified and analyzed to obtain the corresponding verification and identification data.

[0142] The risk verification unit is used to verify the analysis results of each risk feature sequence and obtain risk identification data. The process includes:

[0143] Compare and analyze the sequence identification data corresponding to each risk feature sequence with the verification identification data to obtain the corresponding verification accuracy, preset the verification threshold, and compare and analyze the obtained verification accuracy against the corresponding verification threshold.

[0144] If the corresponding verification accuracy is greater than or equal to the verification threshold, the corresponding risk identification data will be output.

[0145] If the corresponding verification accuracy is less than the verification threshold, the corresponding risk identification data will not be output.

[0146] Based on the comparative analysis structure, obtain the risk identification data corresponding to the risk feature sequence.

[0147] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An artificial intelligence-based tunnel portal slope risk identification and analysis system, comprising a slope risk identification system, characterized in that, The slope risk identification system includes a risk factor investigation and collection module, a portal risk simulation and processing module, a tunnel risk simulation and processing module, a risk factor joint analysis module, and a slope risk identification and analysis module. The risk factor investigation and collection module is used to collect geological feature data and construction monitoring data corresponding to the tunnel construction area, as well as risk factor data and risk event data of the corresponding tunnel portal slope; The tunnel entrance risk simulation processing module is used to set up a physical model of the tunnel entrance slope based on relevant geological feature data, risk factor data, and risk event data, obtain relevant slope deformation and failure characteristic data, and construct a slope failure model. The process of constructing the slope failure model includes: Set up a slope test simulation unit and a slope test treatment unit; The corresponding geological feature data were obtained through the slope test simulation unit, and a three-dimensional image of the virtual tunnel entrance slope structure was constructed. The three-dimensional image of the virtual tunnel entrance slope structure was sliced ​​to obtain the corresponding slope structure slice image. The corresponding similar material mixing ratio was obtained for each pixel in the slope structure slice image based on the similarity ratio and dimensional analysis method. The physical model of the tunnel entrance slope was set according to the similar material mixing ratio of each pixel. Obtain relevant risk factor data and historical construction monitoring data, and analyze and process the obtained data to obtain relevant slope condition information; According to the type of slope working condition information, set the corresponding working condition variable data in sequence, combine the working condition variable data corresponding to different slope working conditions in sequence, obtain the corresponding slope simulation test data set, and perform numerical simulation on each slope simulation test data set based on FLAC3D numerical simulation software to obtain the corresponding test data information. The slope test processing unit acquires and analyzes the test data information corresponding to the corresponding slope simulation test data group, sets up the corresponding test dataset according to the corresponding slope simulation test data group, compares and analyzes the corresponding test data information in the test dataset with the risk event data, and obtains the corresponding slope deformation and failure characteristic data. The obtained slope deformation and failure characteristic data are analyzed and trained based on deep learning algorithms to construct a corresponding slope failure model; The tunnel risk simulation processing module is used to set up a physical model of the tunnel entrance based on relevant geological feature data, risk factor data, and risk event data, obtain relevant support deformation and failure characteristic data, and construct a support failure model. The process of constructing the support failure model includes: Set up a tunnel test simulation unit and a tunnel test processing unit; The corresponding geological feature data are obtained through the tunnel test simulation unit. The corresponding tunnel physical model is constructed according to the construction method of the tunnel entrance slope physical model. The corresponding tunnel working condition information is obtained according to the risk factor data and historical construction monitoring data. Based on the tunnel working condition information, a tunnel simulation test data group is set up, and numerical simulation is performed on the corresponding tunnel simulation test data group using FLAC3D numerical simulation software to obtain the corresponding test data information. The tunnel test processing unit compares and analyzes the test data information of the corresponding tunnel simulation test data group with the corresponding risk event data to obtain the corresponding support deformation failure characteristic data. The obtained support deformation failure characteristic data is then analyzed and trained based on a deep learning algorithm to construct the corresponding support failure model. The risk factor joint analysis module is used to set up a three-dimensional construction model based on the geological feature data and construction monitoring data corresponding to the tunnel construction area, map the slope failure model and support failure model into the three-dimensional construction model, set risk feature sequences at various locations in the three-dimensional construction model, analyze and process the data information corresponding to the corresponding models in the risk feature sequences, obtain the corresponding joint influence coefficients, and construct a risk influence database based on the joint influence coefficients and the corresponding models. The real-time 3D construction model is intelligently sliced ​​to obtain corresponding tunnel entrance slice images, and each tunnel entrance slice image is marked. Risk feature sequences were set based on the corresponding pixels in the slice images of each tunnel entrance, and these sequences were marked as follows: ,in, For the corresponding risk feature sequence, P is the corresponding marked tunnel portal slice image; Within the risk characteristic sequence, corresponding risk characteristic sub-sequences are set based on the test data information corresponding to the slope failure model and the tunnel simulation test data group within the slope failure model and support failure model, respectively. The slope simulation test data group is marked as... The corresponding experimental data information is marked as , … The tunnel simulation test data set was marked as The corresponding experimental data information is marked as , … The corresponding risk feature subsequences are labeled as follows: and ; The obtained risk feature subsequences are combined horizontally, vertically, and comprehensively, respectively. The horizontal combination refers to the combination of test data information corresponding to each risk feature subsequence within each risk feature sequence in the corresponding tunnel portal slice image under different working conditions. The corresponding combination results are marked as follows: and And label the corresponding combined datasets as ; The longitudinal combination refers to the combination of experimental data information corresponding to each risk feature subsequence within the risk feature sequence at corresponding positions in different tunnel portal slice images under different working conditions. The corresponding combination results are marked as follows: And label the corresponding combined datasets as ; The comprehensive combination refers to the combination of experimental data information corresponding to each risk feature subsequence within the risk feature sequence at different locations within different tunnel portal slice images under different working conditions. The corresponding combination results are marked as follows: And label the corresponding combined datasets as ; The combined datasets corresponding to the slope simulation test data sets and tunnel simulation test data sets within the horizontal combination, vertical combination, and comprehensive combination are jointly analyzed to obtain the joint influence coefficients between the risk event data corresponding to different working conditions at the corresponding locations. The corresponding joint impact coefficient is denoted as F, where: ,in, and These are the weighting coefficients for the corresponding test data information within the combined datasets of the slope simulation test data group and the tunnel simulation test data group, respectively, within the horizontal combination. The error term coefficients are the corresponding test data information within the combined dataset of the slope simulation test data group and the tunnel simulation test data group corresponding to the horizontal combination. ,in These are the weighting coefficients for the corresponding test data information within the combined dataset corresponding to the slope simulation test data group and the tunnel simulation test data group within the longitudinal combination. The error term coefficients are the corresponding test data information within the combined dataset of the slope simulation test data group and the tunnel simulation test data group corresponding to the longitudinal combination. ,in , and These are the weight coefficients for the corresponding combined datasets. These are the coefficients of the error term for the corresponding combined dataset; Based on the obtained joint impact coefficients, risk impact data tables are set up according to the corresponding risk characteristic sequences in the real-time three-dimensional construction model. The risk impact data tables within each risk characteristic sequence are integrated to construct a risk impact database; The slope risk identification and analysis module is used to traverse the corresponding three-dimensional construction model according to the construction monitoring data, compare and analyze the risk impact database corresponding to the corresponding risk feature sequence, verify the analysis results of each risk feature sequence, and obtain risk identification data.

2. The tunnel entrance slope risk identification and analysis system based on artificial intelligence according to claim 1, characterized in that, The process of collecting geological feature data and construction monitoring data corresponding to the tunnel construction area includes: A monitoring and acquisition unit is set up, which includes a geological monitoring terminal and multiple monitoring and acquisition terminals. Geological feature data corresponding to the tunnel construction area are collected through geological monitoring terminals; Construction monitoring data corresponding to the tunnel construction area is collected through the monitoring and acquisition terminal.

3. The tunnel entrance slope risk identification and analysis system based on artificial intelligence according to claim 2, characterized in that, The process of collecting risk factor data and risk event data for the corresponding tunnel portal slope includes: Set up survey and data collection units to collect survey data within the corresponding tunnel engineering area, and set up risk factor collection terminals and risk event collection terminals; The risk factor collection terminal analyzes and processes the survey data based on artificial intelligence algorithms to obtain corresponding feature keywords. It then uses NER technology to extract and identify information from the obtained feature keywords to obtain corresponding risk factor data. The risk event acquisition terminal identifies risk event types based on natural language processing algorithms, obtains feature datasets corresponding to the respective risk event types, analyzes and processes the obtained feature datasets sequentially using clustering analysis algorithms, obtains risk factor data evaluation indicators corresponding to the respective risk event types, and marks the obtained data as risk event data.

4. The tunnel entrance slope risk identification and analysis system based on artificial intelligence according to claim 3, characterized in that, The process of setting up a 3D construction model includes: Set up a model analysis unit; obtain relevant geological feature data through the model analysis unit, construct an initial three-dimensional construction model, map the obtained construction monitoring data into the initial three-dimensional construction model for update processing, obtain the corresponding three-dimensional construction model, map the slope failure model and support failure model into the three-dimensional construction model according to their respective positions, and obtain the real-time three-dimensional construction model.

5. The process of acquiring risk identification data in the tunnel entrance slope risk identification and analysis system based on artificial intelligence according to claim 4 includes: Acquire construction monitoring data and map the latest acquired construction monitoring data into the real-time three-dimensional construction model. The corresponding risk feature sequence in the real-time three-dimensional construction model obtains the corresponding working condition information based on the corresponding construction monitoring data. The obtained working condition information is then input into the corresponding slope failure model and support failure model to obtain the corresponding slope deformation failure feature data and tunnel deformation failure feature data. The obtained feature data is then marked as sequence identification data. Based on the joint impact coefficients in the risk impact database, the risk characteristic data in each risk characteristic sequence are verified and analyzed to obtain the corresponding verification and identification data. The sequence identification data corresponding to each risk feature sequence is compared and analyzed with the verification identification data to obtain the corresponding verification accuracy. Based on the corresponding verification accuracy, the risk identification data corresponding to the corresponding risk feature sequence is obtained.

Images