Surrounding rock design method and system for tunnel impact test

Abstract

The invention provides a surrounding rock design method and system for a tunnel impact test, and relates to an engineering structure test technology, and the method comprises the steps: constructing an impact tunnel numerical model based on pre-test data; simulation is conducted based on the impact tunnel numerical model, and influence results on the surrounding rock under different impact working conditions are obtained; deep learning is carried out based on the impact working condition and the geometric information of the corresponding region influenced by the impact, and a prediction model is constructed; inputting a to-be-measured working condition into the prediction model to obtain prediction influence geometric information; and according to the prediction influence geometric information, obtaining surrounding rock design parameters through a pre-constructed first model. According to the method, through quantitative analysis of simulation and deep learning, subjectivity of traditional experience design is avoided, surrounding rock design is more scientific and targeted, surrounding rock response under tunnel impact can be effectively simulated, and accuracy of a test result is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of engineering structure test, in particular to a tunnel impact test surrounding rock design method and system. BACKGROUND

[0002] With the advancement of transportation construction, the number of tunnels built is increasing, and the speed of trains is also increasing, which greatly increases the probability of derailment and impact, and the loss caused by derailment is more serious. Therefore, it is necessary to study the impact response of tunnels. In the past, the test and research on tunnel impact mainly used numerical simulation method, but the numerical simulation method involves high-speed contact problem, complex numerical modeling and large amount of calculation work. In addition, the process of train impact on shield tunnel has very complex influencing factors. Therefore, tunnel impact test is needed to provide data support for design and research.

[0003] When the tunnel is impacted, the interaction between the surrounding rock and the tunnel will affect the bearing capacity of the tunnel, and the characteristics and boundary parameters of the surrounding rock will have a great influence on the response of the tunnel. However, the existing impact test device only considers the impact when the train is in contact with the tunnel plane, and does not consider the interaction between the tunnel and the surrounding rock during the impact process. Because the surrounding rock model of the test device has boundaries, there is reflection and superposition effect of waves, and setting different surrounding rock parameters in the test process will lead to different response data obtained, that is, the data obtained is deviated from the ideal and real situation. Therefore, in the test of the interaction between the surrounding rock and the tunnel, how to accurately design the surrounding rock to obtain reliable response data has become a problem to be solved. SUMMARY

[0004] The purpose of the present application is to provide a tunnel impact test surrounding rock design method and system to improve the above problems. In order to achieve the above purpose, the technical scheme adopted by the present application is as follows: In a first aspect, the present application provides a tunnel impact test surrounding rock design method, comprising: constructing an impact tunnel numerical model based on pre-test data; based on the impact tunnel numerical model, simulating to obtain the influence result on the surrounding rock under different impact conditions; the influence result includes the geometric information of the area affected by impact; based on the impact condition and the corresponding geometric information of the area affected by impact, performing deep learning to construct a prediction model; inputting the to-be-tested condition into the prediction model to obtain predicted influence geometric information; according to the predicted influence geometric information, obtaining the surrounding rock design parameters through a pre-constructed first model, the surrounding rock design parameters including backboard distance, bedrock height, soil body burial depth and plate shape radian.

[0005] In a second aspect, the application also provides a surrounding rock design system for a tunnel impact test, characterized in that it comprises: A first module for constructing an impact tunnel numerical model based on pre-test data; A second module for simulation based on the impact tunnel numerical model to obtain the impact results on the surrounding rock under different impact conditions; the impact results include the geometric information of the area affected by the impact; A third module for deep learning based on the impact conditions and the corresponding geometric information of the area affected by the impact to construct a prediction model; A fourth module for inputting the test conditions into the prediction model to obtain the predicted impact geometric information; A fifth module for obtaining the surrounding rock design parameters, including the backboard distance, bedrock height, soil burial depth and plate shape arc, from the first model pre-constructed according to the predicted impact geometric information.

[0006] Advantages: The method can quickly and accurately output the surrounding rock impact geometric information according to the test conditions by collecting data based on simulation and establishing a prediction model combined with deep learning; the first model pre-constructed can convert the predicted geometric information into specific design parameters such as backboard distance and bedrock height, and the parameters cover the dimensions of soil and plate shape arc; the quantitative analysis of simulation and deep learning avoids the subjectivity of traditional experience design, making the surrounding rock design more scientific and targeted, and effectively simulating the surrounding rock response under tunnel impact to ensure the accuracy of the test results.

[0007] Other features and advantages of the application will be described in the following description, and some will become apparent from the description, or will be understood by those skilled in the art from the description, or will be understood by practicing the embodiments of the application. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed in the embodiments. It should be understood that the following drawings only show some embodiments of the application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0009] Figure 1 The flow chart of the surrounding rock design method for the tunnel impact test in the embodiments; Figure 2 The structural schematic diagram of the impact test device in the embodiments; Figure 3 The structural diagram of the surrounding rock design device for the tunnel impact test in the embodiments.

[0010] Symbol explanation: 1-Air cannon, 2-Acceleration pipe, 3-Railway, 4-Base, 5-Infrared velocimeter, 6-Semi-tunnel model, 7-Backfill box, 8-Top cover, 9-Backfill box base, 800-Surrounding rock design equipment for tunnel impact test, 801-Processor, 802-Memory, 803-Multimedia component, 804-I / O interface, 805-Communication component. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0012] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0013] Example 1: See Figure 1 This embodiment provides a method for designing the surrounding rock for a tunnel impact test, including steps S100, S200, S300, S400 and S500. S100. Construct a numerical model of the impact tunnel based on pre-test data; At the start of the experiment, the position and angle of the impact tunnel need to be set for a pre-loading test. A low-strength test block of the same mass is then impacted at a target velocity onto a force response plate at the tunnel location. This response plate is equipped with an acceleration... ,stress ,pressure Pressure range The relevant sensors are then used to collect data via a high-frequency acquisition system. The collected data is processed according to the following analysis workflow: First, calculate the internal forces of the structure based on stress. : ; By measuring acceleration The inertial force of the structure itself during the impact was calculated. : ; Then based on pressure With pressure range Receive the impact force of the train ; Based on the dynamics formula, the external forces on the structure are obtained. : ; Then, based on the formula for the resultant force of external forces, the impact force is corrected to obtain the impact force acting on the soil: ; Based on Newton's second law, a numerical analysis model is established to solve the equations of motion of the continuum and to establish a database of surrounding rock design parameters. ; : Quality matrix; : Acceleration vector of surrounding rock nodes; : Surrounding rock damping matrix; : Velocity vector of surrounding rock nodes; Rock stiffness matrix; : Displacement vector of surrounding rock nodes; : External force vector at the surrounding rock node.

[0014] Based on the above equations, a finite element numerical model of a semi-infinite soil is established. S200. Simulations are performed based on a numerical model of an impact tunnel to obtain the impact results on the surrounding rock under different impact conditions; the impact results include the geometric information of the impact-affected area. Simulations were conducted under different impact conditions to obtain tunnel damage data and the impact on the surrounding rock. The impact results included geometric information of the impact-affected area, such as the diffusion range of the stress wave generated by the impact in the soil; specifically as follows: Based on the peak stress of each grid in the simulation results, the affected region is identified, and the surface triangular mesh of the affected region is extracted. In order to accurately identify the affected region, in addition to peak stress, data such as plastic strain, peak particle velocity, and energy density contour map can also be considered. The entire grid space is binarized according to a preset threshold to obtain the affected region, and then the surface triangular mesh is extracted. Uniformly resample the surface triangular mesh to generate a preset number of three-dimensional points, thus obtaining a point cloud; Each point cloud is rotated and aligned according to the tunnel axis, the point cloud is translated so that its centroid is at the preset origin, and the point cloud is made to fall within the preset unit sphere through scaling.

[0015] S300: Based on the impact conditions and the geometric information of the corresponding impact-affected areas, a prediction model is constructed using deep learning. The region's geometric information includes three-dimensional point cloud information; this step specifically includes: Input the impact condition into a preset encoder and output the target potential vector; The impact conditions include impact force, area, velocity, impact angle, impact location, rock mass type, tunnel thickness, and longitudinal curvature. These parameters are encoded and concatenated to obtain a parameter vector. A fully connected layer is used as an encoder, inputting the parameter vector and progressively increasing its dimensionality through a series of fully connected layers, introducing nonlinearity. The final fully connected layer maps the dimension to a preset latent space size, and the output is the target latent vector. The decoder works as follows: Create a uniform 2D mesh; for example, generate an (H, 2) matrix for a point cloud with H=2048 points. The target potential vector is copied multiple times to construct a potential tensor. This tensor is then concatenated with a two-dimensional grid to obtain the first tensor. The number of copies is determined by the size of the two-dimensional grid. If the size of the two-dimensional grid is (H, 2), the target potential vector is copied H times. The first tensor represents that each two-dimensional grid point is assigned global working condition information. The first tensor is processed by the first multilayer perceptron to obtain an estimated 3D point cloud; the first 3D point cloud obtained is relatively coarse, with only a rough shape estimate. The target potential vector is concatenated with the estimated 3D point cloud to obtain the second tensor; The second tensor is processed using a second multilayer perceptron to obtain the predicted 3D point cloud.

[0016] After constructing the encoder-decoder architecture described above, the encoder and decoder need to be trained using known sample data. This application employs two methods: firstly, it calculates the Chamfer Distance as a loss term to measure the similarity between the actual point cloud and the predicted point cloud (even if the point order is different, it doesn't matter); secondly, since there may be cases where the shapes and positions are similar but the distribution directions are different, a loss term measuring directional consistency also needs to be introduced, specifically constructed as follows: The covariance matrix of the real sample point cloud is calculated based on the 3D point cloud information, and the covariance matrix is ​​decomposed into features to obtain the real principal component vector. First, the point cloud needs to be translated to the centroid. Then, the variance matrix is ​​calculated by treating the three coordinate axes of the point cloud as three random variables, measuring the linear correlation between them, and giving the principal direction (the direction of maximum variance) and the degree of anisotropy.

[0017] Centralized point cloud representation ; ; yes transpose; ; The diagonal elements represent the variance of the point cloud in the x, y, and z directions, while the off-diagonal elements represent the covariance, indicating the cooperative variation in different directions. Perform eigenvalue decomposition on the covariance matrix: ; Λ is a 3x3 diagonal matrix, and the three values ​​on its diagonal are the eigenvalues ​​(λ1, λ2, λ3), and it is usually assumed that λ1 ≥ λ2 ≥ λ3. V Each column is the corresponding feature vector (v1, v2, v3), which is the true principal component vector in the three-axis direction.

[0018] The predicted 3D point cloud generated by the second multilayer perceptron is obtained, the covariance matrix of the predicted 3D point cloud is calculated, and feature decomposition is performed to obtain the predicted principal component vector. The predicted principal component vectors are calculated using the same method described above. The principal direction loss term is constructed by calculating the angle between the true principal component vector and the predicted principal component vector. ; The position loss term is constructed by calculating the chamfer distance between the real sample point cloud and the predicted 3D point cloud; ; in, This is the location loss value. The smaller the value, the more similar the two point clouds are in shape and spatial location; P For real sample point clouds, p It is a point in a real sample point cloud; Q To predict 3D point clouds, q To predict a point in a 3D point cloud; The square of the Euclidean distance; The total number of points in the real sample point cloud; To predict the total number of points in a 3D point cloud; The loss function of the prediction model is constructed based on the main direction loss term and the position loss term. According to the loss function, the encoder is updated through backpropagation, and the parameters of the decoder are updated simultaneously. ; S400. Input the working condition to be measured into the prediction model to obtain the predicted influence geometric information; Input the working condition parameters to be tested into the model to obtain the predicted three-dimensional point cloud, which represents the spatial range of the surrounding rock that may be affected by the impact under the working condition. Based on this spatial range, select the surrounding rock design parameters to minimize the impact of unreasonable surrounding rock structure design on the test results.

[0019] S500. Based on the predicted influence geometry information, the surrounding rock design parameters are obtained through a pre-constructed first model. These parameters include backplate distance, bedrock height, soil depth, and plate shape curvature. The process includes the following steps: First, features are extracted from the 3D point cloud: The total volume and surface area of ​​the predicted 3D point cloud are extracted by convex hull estimation, and the ratio of volume to surface area is calculated to obtain the basic geometric features. The maximum and minimum values ​​of the point cloud along the three axes of the tunnel are extracted, and the span of the point cloud along the three axes of the tunnel is calculated to obtain the spatial distribution characteristics; these characteristics are used to determine the absolute location and extent of the affected area in space. Calculate the minimum distance and direction from the centroid of the point cloud to the inner contour surface of the tunnel, and calculate the depth of the point cloud in the directions of the arch crown, sidewall, and invert arch respectively to obtain the positional features; Shape features were extracted using principal component analysis. Principal component analysis first calculates the eigenvalues ​​λ1, λ2, λ3 of each principal component in the same manner as in step S300 above; Then, features such as linearity, flatness, sphericity, and isotropy are calculated based on the eigenvalues. The extracted features are spliced ​​together to obtain point cloud feature vectors, which are then input into the first model to obtain the surrounding rock design parameters.

[0020] The tunnel surrounding rock model tested in this application uses a semi-tunnel model 6, which is installed on one side of a backfill box 7 filled with artificial soil. The back plate is the plate opposite the backfill box 7 and the semi-tunnel model 6. The back plate distance refers to the distance between the back plate and the semi-tunnel model 6. In addition, the distance between the two side plates also needs to be designed. The bedrock height refers to the height of the semi-tunnel model 6 from the bottom of the backfill box 7. The soil burial depth refers to the depth of the soil filling. The curvature of the plate means that the three panels of the backfill box 7 can be designed as curved surfaces. Changes in the curvature of the curved surfaces will also cause differences in the obtained response data, which is a parameter that needs to be precisely designed.

[0021] The construction process of the first model is as follows: Obtain test samples, which include historical surrounding rock design parameters and corresponding historical test data, including impact condition data and actual response data; Simulations were performed based on impact condition data and historical surrounding rock design parameters to obtain simulation response data. The actual response data is compared with the simulated response data, for example, by calculating the cosine similarity. The reliability of the historical surrounding rock design parameters is determined based on the comparison results. When the actual response data is similar to the simulated response data, it indicates that the surrounding rock design parameters used under this working condition are reasonable, and this test can be used as a training sample for parameter learning.

[0022] The test samples were screened based on the reliability of historical surrounding rock design parameters to obtain training samples; The gradient boosting tree is trained using training samples to obtain the first model. Specifically, LightGBM or XGBoost models can be used. The relationship between the design parameters and geometric features is not a simple linear relationship. Gradient boosting trees can automatically learn these complex and non-linear mapping relationships by combining multiple weak learners, which is usually better than other models.

[0023] This embodiment also provides an impact testing apparatus for implementing the above method, see [link to relevant documentation]. Figure 2 It includes: 1. Air cannon, 2. Acceleration pipe, 3. Rail, 4. Base, 5. Infrared velocimeter, 6. Semi-tunnel model, 7. Soil filling box, 8. Top cover and 9. Soil filling box base; The air cannon 1 and the acceleration pipe 2 form a launching device. Opening the control valve launches the train model on the track 3 from the starting position of the acceleration pipe. The train model completes its acceleration process within the pipe, and then continues moving due to inertia. The track 3 is a steel structure; the train is held in place on the track and moves forward, ensuring its trajectory.

[0024] The base 4 has both lifting and rotating functions. The lifting function ensures the normal operation of the train. The rotating function of the base 4 allows the train to move at an angle. It overturned and crashed into the tunnel.

[0025] Infrared velocimeter 5 is installed at the end of the transmitting device to monitor the train's impact speed. The tunnel model is made of tiled concrete. Based on the tunnel test requirements, the tunnel diameter, thickness, and longitudinal curvature are determined, and a semi-tunnel model 6 is constructed.

[0026] Example 2: This embodiment provides a surrounding rock design system for tunnel impact tests, including: The first module is used to build a numerical model of the impact tunnel based on pre-experimental data; The second module is used to perform simulations based on the numerical model of the impact tunnel to obtain the impact results on the surrounding rock under different impact conditions; the impact results include the geometric information of the impacted area. The third module is used to perform deep learning based on the impact conditions and the geometric information of the corresponding impact-affected areas to build a prediction model. The fourth module is used to input the working condition to be measured into the prediction model to obtain the predicted influence geometric information; The fifth module is used to obtain the surrounding rock design parameters through a pre-constructed first model based on the predicted influence geometry information. The surrounding rock design parameters include the back plate distance, bedrock height, soil burial depth, and plate shape curvature.

[0027] As an optional implementation, the region geometric information includes three-dimensional point cloud information; the third module includes: The first unit is used to input the impact condition into a preset encoder and output the target potential vector; The second unit is used to create a uniform two-dimensional mesh; The third unit is used to copy the target potential vector multiple times to construct a potential tensor, which is then concatenated with a two-dimensional mesh to obtain the first tensor. The number of copies is determined according to the size of the two-dimensional mesh. The fourth unit is used to process the first tensor using the first multilayer perceptron to obtain an estimated 3D point cloud; The fifth unit is used to concatenate the target latent vector with the estimated 3D point cloud to obtain the second tensor; The sixth unit is used to process the second tensor using a second multilayer perceptron to obtain a predicted 3D point cloud.

[0028] As an optional implementation, the third module further includes: The seventh unit is used to calculate the covariance matrix of the real sample point cloud based on the 3D point cloud information, and to perform eigenvalue decomposition on the covariance matrix to obtain the real principal component vector. The eighth unit is used to acquire the predicted 3D point cloud generated by the second multilayer perceptron, calculate the covariance matrix of the predicted 3D point cloud, and perform feature decomposition to obtain the predicted principal component vector. Unit 9 is used to construct the principal direction loss term by calculating the angle between the true principal component vector and the predicted principal component vector; The tenth unit is used to construct the position loss term by calculating the chamfer distance between the real sample point cloud and the predicted 3D point cloud; The eleventh unit is used to construct the loss function of the prediction model based on the main direction loss term and the position loss term, and to update the encoder through backpropagation according to the loss function.

[0029] As an optional implementation, the fifth module includes: The twelfth unit is used to extract the total volume and surface area of ​​the predicted 3D point cloud through convex hull estimation, and to calculate the ratio of volume to surface area to obtain basic geometric features. The thirteenth unit is used to extract the maximum and minimum values ​​of the point cloud in the three axes of the tunnel, and to calculate the span of the point cloud in the three axes of the tunnel to obtain the spatial distribution characteristics. The fourteenth unit is used to calculate the minimum distance and direction from the centroid of the point cloud to the inner contour surface of the tunnel, and to calculate the depth of the point cloud in the directions of the arch, sidewall, and invert, respectively, to obtain the positional features. Unit 15 is used to extract shape features through principal component analysis; The sixteenth unit is used to stitch together the extracted features to obtain a point cloud feature vector, and then input the point cloud feature vector into the first model to obtain the surrounding rock design parameters.

[0030] Example 3: Corresponding to the above method embodiments, this embodiment also provides a surrounding rock design device for tunnel impact tests. The surrounding rock design device for tunnel impact tests described below and the surrounding rock design method for tunnel impact tests described above can be referred to in correspondence.

[0031] Figure 3 This is a block diagram illustrating a surrounding rock design device 800 for a tunnel impact test according to an exemplary embodiment. Figure 3As shown, the surrounding rock design device 800 for tunnel impact testing includes a processor 801 and a memory 802. The device 800 may also include one or more of the following: a multimedia component 803, an input / output (I / O) interface 804, and a communication component 805. The processor 801 controls the overall operation of the tunnel impact test surrounding rock design device 800 to complete all or part of the steps in the aforementioned tunnel impact test surrounding rock design method. The memory 802 stores various types of data to support the operation of the tunnel impact test surrounding rock design device 800. This data may include, for example, commands for any application or method operating on the tunnel impact test surrounding rock design device 800, and application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0032] Multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals.

[0033] The received audio signal can be further stored in memory 802 or transmitted via communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the surrounding rock design equipment 800 for the tunnel impact test and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof, is used; therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0034] Example 4: Corresponding to the above embodiment of the surrounding rock design method for tunnel impact test, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the surrounding rock design method for tunnel impact test described above.

[0035] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described embodiment of the surrounding rock design method for tunnel impact testing.

[0036] Specifically, the readable storage medium can be a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or any other readable storage medium capable of storing program code.

[0037] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0038] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for designing the surrounding rock for tunnel impact tests, characterized in that, include: A numerical model of the impact tunnel was constructed based on preliminary test data; Simulations were conducted based on a numerical model of an impact tunnel to obtain the impact results on the surrounding rock under different impact conditions. The impact results include geometric information of the region affected by the impact; A prediction model is constructed based on deep learning of the impact conditions and the geometric information of the corresponding impact-affected areas. Input the working condition to be tested into the prediction model to obtain the predicted influence geometric information; Based on the predicted influence geometry information, the surrounding rock design parameters are obtained through a pre-constructed first model. The surrounding rock design parameters include back plate distance, bedrock height, soil burial depth, and plate shape curvature.

2. The surrounding rock design method for tunnel impact tests according to claim 1, characterized in that, The regional geometric information includes three-dimensional point cloud information; deep learning is performed based on the impact condition and the corresponding geometric information of the impact-affected region, including: Input the impact condition into a preset encoder and output the target potential vector; Create a uniform two-dimensional mesh; The target latent vector is copied multiple times to construct a latent tensor, which is then concatenated with a two-dimensional mesh to obtain the first tensor. The number of copies is determined by the size of the two-dimensional mesh. The first tensor is processed using a first multilayer perceptron to obtain an estimated 3D point cloud; The target potential vector is concatenated with the estimated 3D point cloud to obtain the second tensor; The second tensor is processed using a second multilayer perceptron to obtain the predicted 3D point cloud.

3. The surrounding rock design method for tunnel impact tests according to claim 2, characterized in that, The deep learning based on the impact conditions and the geometric information of the corresponding impact-affected region includes: The covariance matrix of the real sample point cloud is calculated based on the 3D point cloud information, and the covariance matrix is ​​decomposed into features to obtain the real principal component vector. The predicted 3D point cloud generated by the second multilayer perceptron is obtained, the covariance matrix of the predicted 3D point cloud is calculated, and feature decomposition is performed to obtain the predicted principal component vector. The principal direction loss term is constructed by calculating the angle between the true principal component vector and the predicted principal component vector. The position loss term is constructed by calculating the chamfer distance between the real sample point cloud and the predicted 3D point cloud; The loss function of the prediction model is constructed based on the main direction loss term and the position loss term. The encoder is then updated through backpropagation according to the loss function.

4. The surrounding rock design method for tunnel impact tests according to claim 1, characterized in that, Based on the predicted impact geometry information, the surrounding rock design parameters are obtained through a pre-constructed first model, including: The total volume and surface area of ​​the predicted 3D point cloud are extracted by convex hull estimation, and the ratio of volume to surface area is calculated to obtain the basic geometric features. The maximum and minimum values ​​of the point cloud in the three axes of the tunnel are extracted respectively, and the span of the point cloud in the three axes of the tunnel is calculated to obtain the spatial distribution characteristics. Calculate the minimum distance and direction from the centroid of the point cloud to the inner contour surface of the tunnel, and calculate the depth of the point cloud in the directions of the arch crown, sidewall, and invert arch respectively to obtain the positional features; Shape features were extracted using principal component analysis. The extracted features are concatenated to obtain point cloud feature vectors, which are then input into the first model to obtain the surrounding rock design parameters.

5. The surrounding rock design method for tunnel impact tests according to claim 1, characterized in that, The method includes: Obtain test samples, which include historical surrounding rock design parameters and corresponding historical test data, including impact condition data and actual response data; Simulations were performed based on impact condition data and historical surrounding rock design parameters to obtain simulation response data. The reliability of historical surrounding rock design parameters is determined by comparing the actual response data with the simulation response data and comparing the results. The test samples were screened based on the reliability of historical surrounding rock design parameters to obtain training samples; The gradient boosting tree is trained using training samples to obtain the first model.

6. The surrounding rock design method for tunnel impact tests according to claim 1, characterized in that, Obtain the impact results on the surrounding rock under different impact conditions, including: Based on the peak stress of each mesh in the simulation results, the affected region is identified, and the surface triangular mesh of the affected region is extracted; Uniformly resample the surface triangular mesh to generate a preset number of three-dimensional points, thus obtaining a point cloud; Each point cloud is rotated and aligned according to the tunnel axis, the point cloud is translated so that its centroid is at the preset origin, and the point cloud is made to fall within the preset unit sphere through scaling.

7. A surrounding rock design system for tunnel impact tests, characterized in that, include: The first module is used to build a numerical model of the impact tunnel based on pre-experimental data; The second module is used to perform simulations based on the numerical model of the impact tunnel to obtain the impact results on the surrounding rock under different impact conditions; the impact results include the geometric information of the impacted area. The third module is used to perform deep learning based on the impact conditions and the geometric information of the corresponding impact-affected areas to build a prediction model. The fourth module is used to input the working condition to be measured into the prediction model to obtain the predicted influence geometric information; The fifth module is used to obtain the surrounding rock design parameters through a pre-constructed first model based on the predicted influence geometry information. The surrounding rock design parameters include the back plate distance, bedrock height, soil burial depth, and plate shape curvature.

8. The surrounding rock design system for tunnel impact tests according to claim 7, characterized in that, The region's geometric information includes three-dimensional point cloud information; The third module includes: The first unit is used to input the impact condition into a preset encoder and output the target potential vector; The second unit is used to create a uniform two-dimensional mesh; The third unit is used to copy the target potential vector multiple times to construct a potential tensor, which is then concatenated with a two-dimensional mesh to obtain the first tensor. The number of copies is determined according to the size of the two-dimensional mesh. The fourth unit is used to process the first tensor using the first multilayer perceptron to obtain an estimated 3D point cloud; The fifth unit is used to concatenate the target latent vector with the estimated 3D point cloud to obtain the second tensor; The sixth unit is used to process the second tensor using a second multilayer perceptron to obtain a predicted 3D point cloud.

9. The surrounding rock design system for tunnel impact tests according to claim 8, characterized in that, The third module also includes: The seventh unit is used to calculate the covariance matrix of the real sample point cloud based on the 3D point cloud information, and to perform eigenvalue decomposition on the covariance matrix to obtain the real principal component vector. The eighth unit is used to acquire the predicted 3D point cloud generated by the second multilayer perceptron, calculate the covariance matrix of the predicted 3D point cloud, and perform feature decomposition to obtain the predicted principal component vector. Unit 9 is used to construct the principal direction loss term by calculating the angle between the true principal component vector and the predicted principal component vector; The tenth unit is used to construct the position loss term by calculating the chamfer distance between the real sample point cloud and the predicted 3D point cloud; The eleventh unit is used to construct the loss function of the prediction model based on the main direction loss term and the position loss term, and to update the encoder through backpropagation according to the loss function.

10. The surrounding rock design system for tunnel impact tests according to claim 7, characterized in that, The fifth module includes: The twelfth unit is used to extract the total volume and surface area of ​​the predicted 3D point cloud through convex hull estimation, and to calculate the ratio of volume to surface area to obtain basic geometric features. The thirteenth unit is used to extract the maximum and minimum values ​​of the point cloud in the three axes of the tunnel, and to calculate the span of the point cloud in the three axes of the tunnel to obtain the spatial distribution characteristics. The fourteenth unit is used to calculate the minimum distance and direction from the centroid of the point cloud to the inner contour surface of the tunnel, and to calculate the depth of the point cloud in the directions of the arch, sidewall, and invert, respectively, to obtain the positional features. Unit 15 is used to extract shape features through principal component analysis; The sixteenth unit is used to stitch together the extracted features to obtain a point cloud feature vector, and then input the point cloud feature vector into the first model to obtain the surrounding rock design parameters.

Images