TSP and TEM fusion tunnel water judgment method based on random forest

Abstract

The invention relates to a TSP and TEM fusion tunnel water judgment method based on a random forest. The method comprises the following steps: firstly, carrying out geological detection on an excavated section by adopting a tunnel seismic wave forecasting method and a transient electromagnetic method to obtain TSP data and TEM data; afterwards, the TSP data and the TEM data are subjected to normalization fusion preprocessing, label data are made according to the actual excavation condition, and fusion water burst judgment training data are obtained; then, constructing a random forest model, and training the random forest model by adopting the fusion water burst judgment training data to obtain a fusion tunnel water judgment model; and finally, geological detection is carried out on the unexcavated section of the tunnel, three-dimensional data obtained through detection are input into the fused tunnel water judgment model, and the tunnel water gushing risk level is determined. The method not only improves the recognition precision of unfavorable geologic bodies such as water gushing, but also improves the automation level of data processing, reduces manual intervention, and ensures the safety of tunnel construction.

Description

Technical Field

[0001] The present application relates to the technical field of tunnel water body identification, and in particular to a tunnel water body identification method based on random forest-based TSP and TEM fusion. Background Art

[0002] As a crucial component of infrastructure construction, tunneling projects face numerous challenges in geologically complex areas. The presence of groundwater, karst formations, faults, and other unfavorable geological formations often pose significant risks to tunnel excavation, construction safety, and long-term operation. To ensure the safety and smooth progress of tunnel construction, it is crucial to accurately predict the geological conditions ahead of the tunnel.

[0003] Traditional methods for predicting the geology ahead of tunnels typically rely on ground surveys, drilling, and manual measurements. However, these methods are limited by several factors: First, they cannot directly obtain geological information deeper ahead of the excavation face; second, drilling equipment is inefficient and can interfere with construction; and third, field measurement methods have low resolution, making it difficult to effectively predict small geological anomalies. Furthermore, traditional methods require incremental implementation during the engineering process, which is not only time-consuming but also prone to blind spots and errors. Therefore, overcoming these limitations and accurately identifying and predicting the geological conditions ahead of the tunnel excavation face in advance is a critical issue that needs to be addressed.

[0004] To improve the accuracy and depth of predictions, two advanced geophysical techniques, Tunnel Seismic Prospecting (TSP) and Transient Electromagnetic Method (TEM), have been introduced into tunnel geological forecasting. These technologies enable remote detection of the strata ahead of the tunnel through physical means before tunnel excavation, providing construction personnel with timely geological information and significantly improving forecast accuracy. However, TSP technology has low resolution, making it difficult to accurately identify small, undesirable geological bodies. It is susceptible to interference from construction noise, which degrades signal quality. Furthermore, data processing is susceptible to subjective factors, which can lead to multiple solutions for the same data and increase the uncertainty of prediction results. TEM technology has a "shallow blind spot" near the tunnel face, meaning that it cannot effectively detect areas too close to the excavation face. Data acquisition is susceptible to electromagnetic interference from the tunnel construction site, which affects signal accuracy. Its detection capability is poor in low-resistance environments, making it difficult to provide sufficient depth information. Furthermore, its effective detection range and depth are still somewhat limited.

[0005] Therefore, in the related technology, there is an urgent need for a method that can improve the accuracy and reliability of complex geological body identification. Summary of the Invention

[0006] Based on this, it is necessary to provide a tunnel water detection method based on random forest-based TSP and TEM fusion, which can improve the accuracy and reliability of complex geological body identification, in order to address the above technical problems.

[0007] In a first aspect, the present application provides a tunnel water detection method based on random forest-based TSP and TEM fusion. The method comprises:

[0008] Tunnel seismic wave prediction method and transient electromagnetic method are used to conduct geological survey on the excavated section to obtain TSP data and TEM data;

[0009] The TSP data and TEM data are respectively subjected to normalization fusion preprocessing, and label data are produced according to actual excavation conditions to obtain fused water inrush determination training data;

[0010] Constructing a random forest model, and using the fused water inrush determination training data to train the random forest model to obtain a fused tunnel water inrush determination model;

[0011] Conduct geological surveys on the unexcavated tunnel section, input the three-dimensional data obtained from the survey into the fused tunnel water assessment model, and determine the tunnel water inrush risk level.

[0012] Optionally, in one embodiment of the present application, the normalization fusion preprocessing includes:

[0013] TEM data were expanded using radial basis function interpolation;

[0014] The elastic parameters of the TSP data and the apparent resistivity of the TEM data were respectively subjected to data splicing and linear scaling normalization.

[0015] Optionally, in one embodiment of the present application, the adopting the fused water inrush determination training data to train the random forest model to obtain a fused tunnel water inrush determination model includes:

[0016] Adopting a double randomness mechanism to improve the generalization ability of the model;

[0017] When splitting a node, the Gini impurity minimization splitting criterion is used;

[0018] Hyperparameter tuning was performed using grid search combined with 5-fold cross validation.

[0019] Optionally, in one embodiment of the present application, the step of training the random forest model using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model further comprises:

[0020] The model training termination condition is determined based on the out-of-bag estimation error.

[0021] Optionally, in one embodiment of the present application, the method further includes:

[0022] The accuracy, recall and F1 value are used to verify the model performance.

[0023] In a second aspect, the present application also provides a tunnel water detection device based on random forest-based TSP and TEM fusion. The device comprises:

[0024] The data acquisition module is used to conduct geological surveys on the excavated section using the tunnel seismic wave prediction method and transient electromagnetic method to obtain TSP data and TEM data;

[0025] A data preprocessing module is used to perform normalization fusion preprocessing on the TSP data and TEM data respectively, to produce label data according to the actual excavation situation, and to obtain fused water inrush determination training data;

[0026] A model training module is used to construct a random forest model, and train the random forest model using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model;

[0027] The tunnel water inrush risk assessment module is used to conduct geological surveys on unexcavated tunnel sections, input the three-dimensional data obtained from the surveys into the fused tunnel water inrush assessment model, and determine the tunnel water inrush risk level.

[0028] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program and the processor executes the steps of the method described in each of the above embodiments.

[0029] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in each of the above embodiments.

[0030] The above-mentioned random forest-based TSP and TEM fusion tunnel water inrush detection method first uses the tunnel seismic wave prediction method and transient electromagnetic method to conduct geological exploration on the excavated section to obtain TSP data and TEM data; then, the TSP data and TEM data are respectively subjected to normalization and fusion preprocessing, and label data are produced according to the actual excavation situation to obtain fused water inrush detection training data; then, a random forest model is constructed and trained using the fused water inrush detection training data to obtain a fused tunnel water inrush detection model; finally, geological exploration is conducted on the unexcavated tunnel section, and the three-dimensional data obtained from the exploration is input into the fused tunnel water inrush detection model to determine the tunnel water inrush risk level. In other words, multi-physics field fusion is achieved through spatial alignment and normalization of TSP and TEM data, a unified feature space is constructed, the dual randomness mechanism of random forest (random selection of samples and features) is used to improve the robustness of the model, and the dynamic threshold division based on probability voting balances the risks of false alarms and missed alarms. By combining data fusion with machine learning technology, the key features in the data of the two methods are deeply explored, and their advantages are combined to improve the accuracy of water gushing judgment. This not only improves the recognition accuracy of adverse geological bodies such as water gushing, but also improves the level of automation of data processing, reduces manual intervention, and ensures the safety of tunnel construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 FIG2 is an application environment diagram of a tunnel water detection method based on a random forest-based TSP and TEM fusion method in one embodiment;

[0032] Figure 2 1 is a flow chart of a tunnel water detection method based on a random forest-based TSP and TEM fusion method according to an embodiment;

[0033] Figure 3 A schematic diagram illustrating TSP elasticity parameter identification according to one embodiment;

[0034] Figure 4 A schematic diagram showing TEM identification in one embodiment;

[0035] Figure 5 A schematic diagram of TEM data in one embodiment;

[0036] Figure 6 Schematic diagram of the recognition results of a tunnel water detection method based on a random forest-based TSP and TEM fusion method in one embodiment;

[0037] Figure 7 FIG1 is a structural block diagram of a tunnel water detection device based on random forest-based TSP and TEM fusion in one embodiment;

[0038] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0040] The TSP and TEM fusion tunnel water detection method based on random forest provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the terminal communicates with the server through the network. The data storage system can store data that the server needs to process. The data storage system can be integrated on the server or placed on the cloud or other network servers. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented as a standalone server or a server cluster consisting of multiple servers.

[0041] In one embodiment, Figure 2 As shown in the figure, a tunnel water detection method based on random forest fusion of TSP and TEM is provided. Figure 1 The following steps are used as an example to illustrate the server in the example:

[0042] S201: Conduct geological surveys on the excavated section using the tunnel seismic wave prediction method and transient electromagnetic method to obtain TSP data and TEM data.

[0043] In the embodiment of the present application, first, based on the surrounding rock conditions revealed by the excavated tunnel section, the tunnel seismic wave prediction method and the transient electromagnetic method are used to conduct geological exploration respectively, and the corresponding TSP data and TEM data are obtained. Specifically, the tunnel seismic wave prediction method, its basic principle is to set a certain number of excitation holes within a certain range of the side wall behind the tunnel excavation face, and excite seismic waves by detonating explosives in the holes. The excited seismic waves propagate into the tunnel surrounding rock in the form of spherical waves. When the wave impedance of the surrounding rock changes (such as cracks, karst, faults or the interface of rock strata), part of the seismic wave transmission boundary continues to propagate forward, and the other part will be reflected back. The reflected seismic waves are received by a high-precision three-component sensor and transmitted to the host to form a seismic wave record. The collected signals are further analyzed by specialized software to obtain intermediate results such as the time profile of P waves, SH waves and SV waves in front of the tunnel excavation face, depth migration profile, rock reflection layer and physical and mechanical parameters. Based on these data, it is determined whether there is an abnormal geological body in the stratum in front of the excavation face. Transient electromagnetic method (TEM) is a time-domain electromagnetic method that uses an ungrounded return line to transmit a pulsed primary electromagnetic field into the rock mass. The decay of the primary electric field generates eddy currents in the geological mass. A receiving coil then senses the spatial and temporal distribution of the secondary electromagnetic field generated by the decay of the eddy currents induced by the pulsed electromagnetic field, thereby solving geological problems. The result of TEM interpretation is a contour map of the apparent resistivity of the rock mass being surveyed. The apparent resistivity contours, their magnitude, and their coordinates are used to determine the shape and location of undesirable geological bodies. The more groundwater in the surrounding rock mass, the lower the apparent resistivity. Conversely, the less groundwater there is and the drier the rock mass, the higher the apparent resistivity.

[0044] The negative reflection and low-velocity anomaly corresponding to TSP can clearly reflect the changes in the surrounding rock properties and the degree of rock fragmentation. Based on the fluctuation characteristics of the longitudinal and transverse wave velocities, the development of groundwater can be roughly determined. The conditions of the surrounding rock in front of the tunnel face can be predicted from a long distance and the location and approximate scale of the fracture zone can be delineated. The effect of TSP elastic parameters on the judgment of rock cracks is as follows. Figure 3 As shown, it can be seen that the TSP elastic parameters respond to both water-bearing and water-free cracks, but have less response to the changes in lithology. This is because both water-bearing and water-free cracks exhibit relatively low velocities. Therefore, the judgment of water-bearing cracks cannot rely solely on TSP elastic parameters, but also requires the introduction of TEM data. TEM low-resistance anomalies directly reflect the state of the surrounding rock and can determine the characteristics of unfavorable surrounding rock within a certain spatial range in front of the excavation face, especially the distribution and scale of groundwater and karst. It is accurate in judging water-bearing cracks and low-resistance surrounding rocks. The effect of TEM low-resistance anomalies on the judgment of low-resistance bodies is as follows: Figure 4As shown in the figure, it can be seen that the TEM method responds more to low-resistance surrounding rock and water-containing cracks because the resistance of low-resistance surrounding rock is relatively low and the resistance of water-containing cracks is also relatively low. There is no response to the latter half of the cracks without water because their resistance is relatively high. Therefore, the TEM method cannot completely eliminate interference, so it is necessary to combine the two methods to achieve the purpose of more accurate water judgment.

[0045] S203: performing normalization fusion preprocessing on the TSP data and TEM data respectively, generating label data according to actual excavation conditions, and obtaining fused water inrush determination training data.

[0046] In this embodiment, TSP and TEM data were pre-processed by normalization and fusion. Three-dimensional TSP seismic data and two-dimensional TEM sector data were spatially aligned using Cartesian coordinate transformation. The corresponding water inflow recording data was then used to form the fused water inflow determination training data. The following table shows the water inflow determination criteria for various TSP and TEM parameters.

[0047]

[0048] Among them, v0 is the initial Poisson's ratio, k v is the Poisson's ratio variation coefficient, R is the reflection coefficient, V is the velocity anomaly, V k is the speed abnormality threshold, is the initial longitudinal wave velocity, is the mean value of the measured longitudinal wave velocity, v k is the anomaly threshold of the P-wave velocity, ρ0 is the initial apparent resistivity, ρ min is the minimum value of apparent resistivity, is the average value of the measured apparent resistivity, is the apparent resistivity variation coefficient.

[0049] At the same time, the longitudinal wave velocity (Vp), shear wave velocity (Vs), and Poisson's ratio (V) were extracted from the TSP data, and the apparent resistivity (ρ) was extracted from the TEM data. A four-dimensional feature space was constructed, and label data was produced according to the actual excavation situation. Specifically, each three-dimensional grid cell was labeled with two categories. If the water inflow of the corresponding excavation section was ≥5m 3 / h, marked as a “high-risk” positive sample; if the water inflow is less than 5m 3 / h, marked as "low-risk" negative samples. Label mapping ensures the spatial consistency of the training data by strictly corresponding the geological data of the actual excavation section to the 3D grid coordinates.

[0050] Specifically, in one embodiment of the present application, the normalization fusion preprocessing includes:

[0051] S301: Extending TEM data using radial basis function interpolation.

[0052] S303: performing data splicing and linear scaling normalization on the elastic parameters of the TSP data and the apparent resistivity of the TEM data respectively.

[0053] In one embodiment of the present application, TSP data is three-dimensional grid point data, generally a segment size is 150m*100m*100m, 1m or 1.5m is a grid, and TEM data is generally a plurality of two-dimensional sector data, such as Figure 5 As shown, it covers part of the area in front of the tunnel. Due to the need for spatial alignment, three-dimensional data is required, and the radial basis function (RBF) interpolation method is used to interpolate it. The formula is shown below.

[0054]

[0055] Among them, λ j is the weight coefficient to be determined, φ(||xx j ||) is the radial basis function, which only depends on the distance between points ||xx j ||, p(x) is a low-order polynomial term, usually linear or quadratic, used to capture global trends.

[0056] The elastic parameters of the TSP (P-wave velocity Vp, S-wave velocity Vs, Poisson's ratio V) and the apparent resistivity (ρ) of the TEM are then linearly scaled and normalized, and the data are spliced ​​to eliminate the differences in physical dimensions and form a unified 3D grid dataset. Specifically, the three elastic parameters of the TSP data (Vp, Vs, Poisson's ratio) are jointly interpreted to obtain a threshold, and then linearly normalized using the formula:

[0057]

[0058] Where x is the original parameter value (such as elastic parameter or apparent resistivity), x min is the minimum value of the parameter in the data set, x max is the maximum value of the parameter in the data set.

[0059] The linear weighted average method is then used to smoothly splice the data. The specific method is to change the weight on the left from 1 to 0 and on the right from 0 to 1 linearly, and then perform weighted averaging.

[0060] S205: Constructing a random forest model, and using the fused water inrush determination training data to train the random forest model to obtain a fused tunnel water inrush determination model.

[0061] In an embodiment of the present application, a random forest model is constructed based on a decision tree, and the model is trained using fused water gushing judgment training data. The longitudinal wave velocity (Vp), shear wave velocity (Vs), Poisson's ratio (V) of TSP data and the apparent resistivity of TEM data are used as inputs, and the water gushing label is used as a reference standard. By continuously adjusting the model parameters, the model prediction output is made closer to the water gushing label, and finally a fused tunnel water gushing judgment model is obtained.

[0062] Specifically, in one embodiment of the present application, the random forest model is trained using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model, including:

[0063] S401: Use a dual randomness mechanism to improve model generalization capabilities.

[0064] S403: When splitting a node, the Gini impurity minimization splitting criterion is adopted.

[0065] S405: Hyperparameter tuning is performed through grid search combined with 5-fold cross validation.

[0066] In one embodiment of the present application, first, the model generalization ability is improved through a double randomness mechanism, that is, each decision tree randomly extracts 63.2% of the samples from the original data set (bootstrap sampling method with replacement), and the remaining 36.8% of the samples are used as out-of-bag (OOB) data for real-time error estimation. Each time a node is split, a random selection is made from the four-dimensional features (Vp, Vs, V, ρ) The features are split to reduce the correlation between features. The splitting criterion uses Gini impurity minimization. The Gini index calculation formula is:

[0067]

[0068] Among them, p i Based on the proportion of samples in the i-th category in a node, the feature and threshold that results in the largest decrease in the Gini index are selected as the node splitting step until the preset tree depth or node purity requirements are met. Hyperparameter tuning is performed using a grid search combined with 5-fold cross-validation, ultimately determining a maximum tree depth of 15 layers (to limit model complexity and prevent overfitting) and a maximum number of decision trees of 200 (to balance computational efficiency and model performance).

[0069] In one embodiment of the present application, the step of training the random forest model using the fused water inrush determination training data to obtain the fused tunnel water inrush determination model further includes:

[0070] The model training termination condition is determined based on the out-of-bag estimation error.

[0071] In one embodiment of the present application, the out-of-bag (OOB) error is monitored in real time during training, and training is terminated when the OOB error stabilizes at 13.1%. The OOB error is the average of the prediction errors of all out-of-bag samples and can be used to evaluate the overall performance of the model.

[0072] S207: Conducting geological survey on the unexcavated tunnel section, inputting the three-dimensional data obtained from the survey into the fused tunnel water assessment model, and determining the tunnel water inrush risk level.

[0073] In the embodiment of the present application, the tunnel seismic wave prediction method and transient electromagnetic method are used to conduct geological exploration of the unexcavated tunnel. The obtained data are also normalized and fused for preprocessing. The obtained three-dimensional data are input into the trained random forest model, and the probability of water inrush risk of each grid unit is output through probability voting. The specific judgment rule is: if the probability is ≥ 0.7, it is judged as a "high risk" area; if the probability is < 0.7, it is judged as a "low risk" area. Figure 6 The following is the recognition result diagram.

[0074] In one embodiment of the present application, the method further includes:

[0075] The accuracy, recall and F1 value are used to verify the model performance.

[0076] In one embodiment of the present application, the model performance is verified by accuracy (≥92%), recall (≥90%), and F1 value (≥0.88) to ensure that the false negative rate of high-risk areas is ≤10%.

[0077] In the above-mentioned random forest-based TSP and TEM fusion tunnel water inrush detection method, first, geological exploration is carried out on the excavated section using the tunnel seismic wave prediction method and the transient electromagnetic method to obtain TSP data and TEM data; then, the TSP data and TEM data are respectively subjected to normalization and fusion preprocessing, and label data are produced according to the actual excavation situation to obtain fused water inrush detection training data; then, a random forest model is constructed and trained using the fused water inrush detection training data to obtain a fused tunnel water inrush detection model; finally, geological exploration is carried out on the unexcavated tunnel section, and the three-dimensional data obtained from the exploration is input into the fused tunnel water inrush detection model to determine the tunnel water inrush risk level. In other words, multi-physics field fusion is achieved through spatial alignment and normalization of TSP and TEM data, a unified feature space is constructed, the dual randomness mechanism of random forest (random selection of samples and features) is used to improve the robustness of the model, and the dynamic threshold division based on probability voting balances the risks of false alarms and missed alarms. By combining data fusion with machine learning technology, the key features in the data of the two methods are deeply explored, and their advantages are combined to improve the accuracy of water gushing judgment. This not only improves the recognition accuracy of adverse geological bodies such as water gushing, but also improves the level of automation of data processing, reduces manual intervention, and ensures the safety of tunnel construction.

[0078] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0079] Based on the same inventive concept, the embodiments of the present application also provide a random forest-based TSP and TEM fusion tunnel water detection device for implementing the above-mentioned random forest-based TSP and TEM fusion tunnel water detection method. The implementation solution provided by this device is similar to the implementation solution described in the above-mentioned method. Therefore, the specific limitations of one or more random forest-based TSP and TEM fusion tunnel water detection device embodiments provided below can be referred to the limitations of the random forest-based TSP and TEM fusion tunnel water detection method above, and will not be repeated here.

[0080] In one embodiment, Figure 7As shown, a tunnel water inrush risk determination device 700 based on random forest TSP and TEM fusion is provided, comprising: a data acquisition module 701, a data preprocessing module 703, a model training module 705 and a tunnel water inrush risk determination module 707, wherein:

[0081] The data acquisition module 701 is used to perform geological survey on the excavated section using the tunnel seismic wave prediction method and the transient electromagnetic method to obtain TSP data and TEM data.

[0082] The data preprocessing module 703 is used to perform normalization fusion preprocessing on the TSP data and TEM data respectively, generate label data according to the actual excavation situation, and obtain fused water inrush determination training data.

[0083] The model training module 705 is used to construct a random forest model, and use the fused water inrush determination training data to train the random forest model to obtain a fused tunnel water inrush determination model.

[0084] The tunnel water inrush risk determination module 707 is used to perform geological survey on the unexcavated tunnel section, input the three-dimensional data obtained by the survey into the fused tunnel water inrush risk determination model, and determine the tunnel water inrush risk level.

[0085] In one embodiment of the present application, the data preprocessing module is further configured to:

[0086] TEM data were expanded using radial basis function interpolation;

[0087] The elastic parameters of the TSP data and the apparent resistivity of the TEM data were respectively subjected to data splicing and linear scaling normalization.

[0088] In one embodiment of the present application, the model training module is further used to:

[0089] Adopting a double randomness mechanism to improve the generalization ability of the model;

[0090] When splitting a node, the Gini impurity minimization splitting criterion is used;

[0091] Hyperparameter tuning was performed using grid search combined with 5-fold cross validation.

[0092] In one embodiment of the present application, the model training module is further used to:

[0093] The model training termination condition is determined based on the out-of-bag estimation error.

[0094] In one embodiment of the present application, the method further includes:

[0095] The accuracy, recall and F1 value are used to verify the model performance.

[0096] Each module in the random forest-based TSP and TEM fusion tunnel water detection device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0097] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, memory, communication interface, display screen and input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a TSP and TEM fusion tunnel water judgment method based on random forest is implemented. The display screen of the computer device can be a liquid crystal display or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0098] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0099] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0100] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0101] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0102] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0103] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0104] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0105] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A tunnel water detection method based on random forest-based TSP and TEM fusion, characterized in that: The method comprises: Tunnel seismic wave prediction method and transient electromagnetic method are used to conduct geological survey on the excavated section to obtain TSP data and TEM data; The TSP data and TEM data are respectively subjected to normalization fusion preprocessing, and label data are produced according to actual excavation conditions to obtain fused water inrush determination training data; Constructing a random forest model, and using the fused water inrush determination training data to train the random forest model to obtain a fused tunnel water inrush determination model; Conduct geological surveys on the unexcavated tunnel section, input the three-dimensional data obtained from the survey into the fused tunnel water assessment model, and determine the tunnel water inrush risk level.

2. The tunnel water detection method based on random forest TSP and TEM fusion according to claim 1 is characterized in that: The normalization fusion preprocessing includes: TEM data were expanded using radial basis function interpolation; The elastic parameters of the TSP data and the apparent resistivity of the TEM data were respectively subjected to data splicing and linear scaling normalization.

3. The tunnel water detection method based on random forest TSP and TEM fusion according to claim 1 is characterized in that: The random forest model is trained using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model, comprising: Adopting a double randomness mechanism to improve the generalization ability of the model; When splitting a node, the Gini impurity minimization splitting criterion is used; Hyperparameter tuning was performed using grid search combined with 5-fold cross validation.

4. The tunnel water detection method based on random forest TSP and TEM fusion according to claim 3 is characterized in that: The step of training the random forest model using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model further comprises: The model training termination condition is determined based on the out-of-bag estimation error.

5. The tunnel water detection method based on random forest TSP and TEM fusion according to claim 1 is characterized in that: The method further comprises: The accuracy, recall and F1 value are used to verify the model performance.

6. A tunnel water detection device based on random forest-based TSP and TEM fusion, characterized in that: The device comprises: The data acquisition module is used to conduct geological surveys on the excavated section using the tunnel seismic wave prediction method and transient electromagnetic method to obtain TSP data and TEM data; A data preprocessing module is used to perform normalization fusion preprocessing on the TSP data and TEM data respectively, to produce label data according to the actual excavation situation, and to obtain fused water inrush determination training data; A model training module is used to construct a random forest model, and train the random forest model using the fused water inrush determination training data to obtain a fused tunnel water inrush determination model; The tunnel water inrush risk assessment module is used to conduct geological surveys on unexcavated tunnel sections, input the three-dimensional data obtained from the surveys into the fused tunnel water inrush assessment model, and determine the tunnel water inrush risk level.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images