Deep tunnel rockburst prediction method and device, computer equipment and storage medium

Abstract

The invention relates to a deep tunnel rockburst prediction method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining the engineering geological conditions of a tunnel site area, carrying out a field ground stress test, and carrying out the ground stress inversion through a multiple regression method; establishing and optimizing a tunnel numerical model in combination with the actual condition of the tunnel; based on field monitoring data, surrounding rock test data and a numerical model, a surrounding rock parameter inversion analysis data set is established, a BP neural network is optimized by adopting a particle swarm optimization algorithm to invert surrounding rock parameters, and the surrounding rock parameter inversion precision and efficiency are improved; numerical simulation and surrounding rock parameter inversion are carried out on the current construction section of the tunnel, the potential rockburst risk of the current construction section is analyzed according to the rockburst criterion, and the rockburst grade and the occurrence probability of the current construction section are judged. The method is closely combined with the actual situation of the tunnel site construction section, and the rock burst prediction accuracy is improved; on the premise that the prediction accuracy is guaranteed, a large amount of engineering cost is saved, and reference is provided for rockburst prediction and site construction.

Description

Technical Field

[0001] The present application relates to the technical field of tunnel rockburst prediction, and in particular to a method, device, computer equipment and storage medium for predicting rockburst in a deep buried tunnel. Background Art

[0002] With the development of economy and society and major strategic needs, more and more tunnel projects are gradually developing in the plateau and mountainous areas of central and western China, and higher requirements are put forward for construction. During the construction of deep-buried tunnels in the plateau, due to the influence of active plate tectonics, high altitude, and deep tunnel burial, the high ground stress characteristics of the tunnel site are obvious, which can easily induce frequent rock burst disasters in hard rock tunnels, seriously threatening the lives of construction workers, damaging machinery and equipment, and affecting the construction progress, bringing great challenges to construction.

[0003] Rockburst prediction and forecasting is an important part of rockburst prevention and control in tunnels. Effective rockburst prediction is the key to guiding on-site rockburst prevention and control. At present, the main rockburst prediction methods include theoretical analysis and numerical simulation. In addition, on-site monitoring technologies such as microseismic monitoring can provide an effective reference for guiding rockburst prevention and control. However, most of the current theoretical analysis or numerical simulation methods carry out rockburst prediction work in tunnel sections based on various hypothetical conditions. Affected by changes in geological conditions and construction methods, the obtained rockburst prediction results are often inconsistent with the on-site rockburst situation, and the accuracy of rockburst prediction needs to be improved. Monitoring methods such as microseismic monitoring are time-consuming, labor-intensive, and expensive, which will bring certain pressure to the project cost. Therefore, it is necessary to propose a simple and effective rockburst prediction method based on on-site construction conditions.

[0004] Combining the actual situation at the construction site is an effective way to accurately predict rockburst. Since rockburst is mainly related to factors such as ground stress and surrounding rock parameters, how to more accurately obtain the characteristics of the tunnel ground stress field and surrounding rock parameters is the key to improving the accuracy of rockburst prediction. Although indoor or field test methods can obtain more accurate data, the test methods are often time-consuming and labor-intensive, and the cost is high. It is impossible to conduct real-time testing on all construction sections of the tunnel, which will affect the accuracy and efficiency of rockburst prediction. Therefore, it is necessary to combine field data for efficient and accurate rockburst prediction. Summary of the invention

[0005] Based on this, it is necessary to provide a deep buried tunnel rock burst prediction method, device, computer equipment and storage medium for the above technical problems. This method uses geostress inversion to obtain the three-dimensional geostress field characteristics of the tunnel site area, and then combines field test data and monitoring data to use particle swarm algorithm to optimize BP neural network for inversion to obtain tunnel surrounding rock parameters, and combines rock burst criteria to predict and analyze the tunnel rock burst level, in order to provide a reference for deep buried tunnel rock burst prediction and forecast, and guide on-site rock burst prevention and control construction.

[0006] A method for predicting rockburst in a deep tunnel, the method comprising:

[0007] According to the engineering geological conditions of the tunnel site and the on-site geostress test results, the multivariate regression method was used to perform geostress inversion and the inversion results of the geostress field in the tunnel site were obtained.

[0008] A tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods.

[0009] Based on the field monitoring data, surrounding rock test data and tunnel numerical simulation model, the data set required for the inversion analysis of surrounding rock parameters is obtained; based on the data set, the particle swarm algorithm is used to optimize the BP neural network to obtain the particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters, and the inversion results of the surrounding rock parameters of each rock burst section of the tunnel are obtained.

[0010] According to the inversion results of the geostress field and the surrounding rock parameters of each rockburst section in the constructed section of the tunnel, the rockburst criterion applicable to the tunnel is established.

[0011] The tunnel numerical simulation model is used to perform numerical simulation on the current construction section of the tunnel, and the BP neural network prediction model with particle swarm optimization is used to invert the surrounding rock parameters. According to the rock burst criterion and the inversion results of the surrounding rock parameters of the current construction section of the tunnel, the rock burst grade and occurrence probability of the current construction section of the tunnel are determined.

[0012] In one embodiment, based on the engineering geological conditions of the tunnel site and the on-site geostress test results, a multivariate regression method is used to perform geostress inversion to obtain the geostress field inversion results of the tunnel site, including:

[0013] Based on the aerial photography data, elevation data and satellite images in the project site survey data, a three-dimensional refined geological topological model of the tunnel site was established using GIS methods and finite element software.

[0014] According to the on-site geostress test results and geomechanical analysis results, the main factors affecting the formation of the initial geostress field of the tunnel are determined, and the stress values ​​of known measuring points under these factors are obtained.

[0015] The inversion multivariate regression equation of the geostress field in the tunnel site area is established based on the known stress values ​​of the measuring points and the corresponding measured stress values.

[0016] The least square method is used to determine the optimal solution of the independent variable coefficient in the inversion multivariate regression equation of the tunnel site area's geostress field, and then the inversion result of the tunnel site area's geostress field is determined using the inversion multivariate regression equation of the tunnel site area's geostress field.

[0017] According to the inversion results of the geostress field, the distribution rules of the maximum principal stress, minimum principal stress and intermediate principal stress on the tunnel axis are determined, and the inversion value and direction of the tunnel geostress are obtained through coordinate transformation.

[0018] According to the error between the tunnel geostress inversion value and the measured geostress result and the preset error threshold, the inversion model is continuously adjusted to ensure the accuracy of the inversion model.

[0019] In one embodiment, a tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data, and corresponding excavation and support methods, including:

[0020] A tunnel numerical simulation model is established based on the tunnel face and surrounding rock conditions, engineering geological conditions of the tunnel site, characteristics of the geostress field, tunnel surrounding rock test data and corresponding excavation and support methods.

[0021] The tunnel is numerically simulated through the tunnel numerical simulation model to obtain numerical simulation results; the numerical simulation results include: tunnel vault settlement, clearance convergence and deformation of key areas.

[0022] The numerical simulation results are compared with the actual monitoring data to verify the accuracy of the tunnel numerical simulation model. If the error between the numerical simulation results and the actual monitoring data is greater than or equal to the preset threshold, the cause is identified and the numerical model is continuously adjusted to ensure the reliability of the numerical model.

[0023] In one embodiment, based on the field monitoring data, the surrounding rock test data and the tunnel numerical simulation model, a data set required for the surrounding rock parameter inversion analysis is obtained, including:

[0024] Based on the on-site monitoring data, surrounding rock test data and tunnel numerical simulation model, the surrounding rock parameters of each rockburst section of the tunnel are extracted.

[0025] The surrounding rock parameters of all rockburst sections of the tunnel are combined into a data set required for the surrounding rock parameter inversion analysis.

[0026] In one embodiment, a BP neural network is optimized by a particle swarm algorithm according to a data set to obtain a BP neural network prediction model optimized by a particle swarm; the BP neural network prediction model optimized by a particle swarm is used to invert the surrounding rock parameters of the tunnel to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel, including:

[0027] Select activation function and loss function, analyze and optimize the model parameters of BP neural network, and obtain the optimal structure of BP neural network.

[0028] Initialize parameters, set the number of particles and iterations, and normalize and standardize variable data.

[0029] The BP neural network prediction model optimized by particle swarm optimization was constructed, and the mean square error (MSE) of the prediction model was used as the fitness index of the particle swarm. 2 As an evaluation indicator of prediction accuracy.

[0030] 20% of the data are randomly selected from the dataset as the test set, and the rest of the data are used as the training set.

[0031] The initial data is imported into the particle swarm optimized BP neural network prediction model, and then the training set is used to connect the particle swarm optimized BP neural network prediction model for training. The test set is used to predict the trained particle swarm optimized BP neural network prediction model, and the predicted values ​​of surrounding rock parameters and the mean square error MSE are calculated. Subsequently, the mean square error MSE is returned to the PSO model as the fitness to continuously update the parameter combination (C, g) until the number of iterations is completed, and the particle swarm optimized BP neural network prediction model is obtained.

[0032] The BP neural network prediction model based on particle swarm optimization is used to invert the surrounding rock parameters of the tunnel, and the inversion results of the surrounding rock parameters of each rock burst section of the tunnel are obtained.

[0033] In one embodiment, based on the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rock burst section of the constructed tunnel, a rock burst criterion applicable to the tunnel is established, including:

[0034] According to the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rockburst section of the constructed section of the tunnel, a functional mapping relationship between the rockburst grade of the constructed section of the tunnel and the geostress and surrounding rock parameters is established; or the prediction results of each rockburst criterion are compared with the actual rockburst situation on site to select the most suitable rockburst criterion for the tunnel.

[0035] In one of the embodiments, the method further includes: according to the progress of tunnel construction, based on the increasing on-site monitoring values ​​and surrounding rock test results, continuously increasing the number of samples for surrounding rock parameter inversion analysis, improving the surrounding rock inversion accuracy, continuously optimizing the rock burst criterion, and improving the rock burst prediction accuracy.

[0036] A deep tunnel rockburst prediction device, the device comprising:

[0037] The geostress field inversion module is used to perform geostress inversion based on the acquired engineering geological conditions of the tunnel site and the on-site geostress test results, using the multivariate regression method to obtain the geostress field inversion results of the tunnel site.

[0038] The tunnel numerical simulation model building module is used to build a tunnel numerical simulation model based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods.

[0039] The surrounding rock parameter inversion module is used to obtain the data set required for the surrounding rock parameter inversion analysis based on the field monitoring data, surrounding rock test data and the tunnel numerical simulation model; the particle swarm algorithm is used to optimize the BP neural network according to the data set to obtain the particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel.

[0040] The rockburst criterion construction module is used to establish the rockburst criterion applicable to the tunnel based on the inversion results of the ground stress field and the inversion results of the surrounding rock parameters of each rockburst section of the constructed section of the tunnel.

[0041] The deep-buried tunnel rockburst prediction module is used to perform numerical simulation on the current construction section of the tunnel using the tunnel numerical simulation model, and to perform surrounding rock parameter inversion using the particle swarm optimized BP neural network prediction model. According to the rockburst criterion and the surrounding rock parameter inversion results of the current construction section of the tunnel, the rockburst grade and occurrence probability of the current construction section of the tunnel are determined.

[0042] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0043] According to the engineering geological conditions of the tunnel site and the on-site geostress test results, the multivariate regression method was used to perform geostress inversion and the inversion results of the geostress field in the tunnel site were obtained.

[0044] A tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods.

[0045] Based on the field monitoring data, surrounding rock test data and tunnel numerical simulation model, the data set required for the inversion analysis of surrounding rock parameters is obtained; based on the data set, the particle swarm algorithm is used to optimize the BP neural network to obtain the particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters, and the inversion results of the surrounding rock parameters of each rock burst section of the tunnel are obtained.

[0046] According to the inversion results of the geostress field and the surrounding rock parameters of each rockburst section in the constructed section of the tunnel, the rockburst criterion applicable to the tunnel is established.

[0047] The tunnel numerical simulation model is used to perform numerical simulation on the current construction section of the tunnel, and the BP neural network prediction model with particle swarm optimization is used to invert the surrounding rock parameters. According to the rock burst criterion and the inversion results of the surrounding rock parameters of the current construction section of the tunnel, the rock burst grade and occurrence probability of the current construction section of the tunnel are determined.

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0049] According to the engineering geological conditions of the tunnel site and the on-site geostress test results, the multivariate regression method was used to perform geostress inversion and the inversion results of the geostress field in the tunnel site were obtained.

[0050] A tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods.

[0051] Based on the field monitoring data, surrounding rock test data and tunnel numerical simulation model, the data set required for the inversion analysis of surrounding rock parameters is obtained; based on the data set, the particle swarm algorithm is used to optimize the BP neural network to obtain the particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters, and the inversion results of the surrounding rock parameters of each rock burst section of the tunnel are obtained.

[0052] According to the inversion results of the geostress field and the surrounding rock parameters of each rockburst section in the constructed section of the tunnel, the rockburst criterion applicable to the tunnel is established.

[0053] The tunnel numerical simulation model is used to perform numerical simulation on the current construction section of the tunnel, and the BP neural network prediction model with particle swarm optimization is used to invert the surrounding rock parameters. According to the rock burst criterion and the inversion results of the surrounding rock parameters of the current construction section of the tunnel, the rock burst grade and occurrence probability of the current construction section of the tunnel are determined.

[0054] The above-mentioned deep-buried tunnel rockburst prediction method, device, computer equipment and storage medium, the method includes: obtaining the engineering geological conditions of the tunnel site, conducting on-site geostress testing, and using the multivariate regression method to perform geostress inversion; combining the actual situation of the tunnel, establishing and optimizing the tunnel numerical model; establishing a surrounding rock parameter inversion analysis data set based on on-site monitoring data, surrounding rock test data and numerical model, using particle swarm algorithm to optimize BP neural network to invert surrounding rock parameters, and improving the accuracy and efficiency of surrounding rock parameter inversion; conducting numerical simulation and surrounding rock parameter inversion on the current construction section of the tunnel, using rockburst criteria to analyze the potential rockburst risk of the current construction section, and judging the rockburst level and occurrence probability of the current construction section. This method is closely combined with the actual situation of the on-site construction section of the tunnel, and improves the accuracy of rockburst prediction; on the premise of ensuring the accuracy of prediction, it saves a lot of engineering costs and provides a reference for rockburst prediction and on-site construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of a process for predicting rockburst in a deep tunnel in one embodiment;

[0056] Figure 2It is a schematic diagram of the tunnel surrounding rock parameter inversion process based on the particle swarm algorithm to optimize the BP neural network in another embodiment;

[0057] Figure 3 is a diagram showing changes in principal stress along the tunnel axis in another embodiment;

[0058] Figure 4 A comparison diagram of rock burst prediction results and actual rock burst conditions on site in another embodiment;

[0059] Figure 5 is a structural block diagram of a deep tunnel rockburst prediction device in one embodiment;

[0060] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] In one embodiment, Figure 1 As shown, a deep tunnel rockburst prediction method is provided, the method comprising the following steps:

[0063] Step 100: Based on the acquired engineering geological conditions of the tunnel site and the on-site geostress test results, a multivariate regression method is used to perform geostress inversion to obtain the geostress field inversion results of the tunnel site.

[0064] Specifically, conduct on-site geological surveys on the selected tunnel site to obtain the engineering geological conditions of the tunnel site, carry out on-site geostress tests, obtain the magnitude and direction of geostress at no less than 6 locations, use finite element analysis software to establish a refined three-dimensional geological model of the tunnel site, establish the initial geostress field inversion regression equation based on the multiple regression method, and calculate the optimal solution of each variable coefficient, then determine the geostress field of the tunnel site, and analyze the distribution law of the geostress field in the tunnel site.

[0065] Step 102: A tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data, and corresponding excavation and support methods.

[0066] Specifically, a tunnel numerical simulation model is established based on the tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods. The tunnel settlement and deformation are analyzed through numerical simulation and compared with the field monitoring data to verify the reliability of the numerical model. The error between the numerical simulation results and the field monitoring data should be less than 10%.

[0067] Step 104: According to the field monitoring data, the surrounding rock test data and the tunnel numerical simulation model, a data set required for the surrounding rock parameter inversion analysis is obtained; according to the data set, a particle swarm algorithm is used to optimize the BP neural network to obtain a particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel.

[0068] Specifically, based on the field monitoring data, surrounding rock test data and numerical model, the surrounding rock parameters of each rockburst section of the tunnel are extracted and the data set required for the surrounding rock parameter inversion analysis is formed. The BP neural network is used to invert the surrounding rock parameters, and the particle swarm algorithm is used to optimize the BP neural network to improve the accuracy and efficiency of the surrounding rock parameter inversion. The field monitoring data and test data are used as the basis to verify the accuracy of the surrounding rock parameter inversion using the particle swarm algorithm to optimize the BP neural network.

[0069] Step 106: Based on the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rock burst section of the constructed tunnel, a rock burst criterion applicable to the tunnel is established.

[0070] Specifically, according to the rockburst situation of the constructed section of the tunnel, based on the inversion results of the in-situ stress and surrounding rock parameters of each rockburst section, a mapping relationship between the on-site rockburst grade and the in-situ stress and surrounding rock parameters is established, and the rockburst criterion applicable to the tunnel is established.

[0071] Step 108: numerically simulate the current construction section of the tunnel using the tunnel numerical simulation model, and use the particle swarm optimized BP neural network prediction model to invert the surrounding rock parameters. According to the rock burst criterion and the obtained inversion results of the surrounding rock parameters of the current construction section of the tunnel, determine the rock burst level and occurrence probability of the current construction section of the tunnel.

[0072] Specifically, based on the inversion results of geostress and surrounding rock parameters, a functional mapping relationship between the rockburst level of the constructed section of the tunnel and the geostress and surrounding rock parameters is established, or the prediction results of each rockburst criterion are compared with the actual rockburst situation on site to select the most suitable rockburst criterion for the tunnel; then, based on the inversion results of the geostress field and surrounding rock parameters of the current construction section of the tunnel, the potential rockburst risk level and occurrence probability of the current construction section are judged.

[0073] In the above-mentioned deep-buried tunnel rockburst prediction method, the method includes: obtaining the engineering geological conditions of the tunnel site, conducting on-site geostress testing, and using the multivariate regression method to perform geostress inversion; combining the actual situation of the tunnel, establishing and optimizing the tunnel numerical model; establishing a surrounding rock parameter inversion analysis data set based on on-site monitoring data, surrounding rock test data and numerical models, using the particle swarm algorithm to optimize the BP neural network to invert the surrounding rock parameters, and improving the accuracy and efficiency of surrounding rock parameter inversion; conducting numerical simulation and surrounding rock parameter inversion on the current construction section of the tunnel, using rockburst criteria to analyze the potential rockburst risk of the current construction section, and judging the rockburst level and occurrence probability of the current construction section. This method is closely combined with the actual situation of the on-site construction section of the tunnel, and improves the accuracy of rockburst prediction; on the premise of ensuring the accuracy of the prediction, it saves a lot of engineering costs and provides a reference for rockburst prediction and on-site construction.

[0074] In one embodiment, step 100 includes: establishing a three-dimensional refined geological topological model of the tunnel site area using GIS methods and finite element software based on aerial data, elevation data and satellite images in the project address survey data; determining the main factors affecting the formation of the initial geostress field of the tunnel based on the on-site geostress test results and geomechanical analysis results, and obtaining the stress values ​​of known measuring points under the factors; establishing a geostress field inversion multivariate regression equation for the tunnel site area based on the known measuring point stress values ​​and the corresponding measured stress values; determining the optimal solution of the independent variable coefficients in the geostress field inversion multivariate regression equation for the tunnel site area using the least squares method, and then determining the geostress field inversion result of the tunnel site area using the geostress field inversion multivariate regression equation; determining the distribution law of the maximum principal stress, the minimum principal stress and the intermediate principal stress of the tunnel axis based on the geostress field inversion result, and obtaining the tunnel geostress inversion value and direction through coordinate transformation; and continuously adjusting the inversion model based on the error between the tunnel geostress inversion value and the geostress measured result and the preset error threshold to ensure the accuracy of the inversion model.

[0075] Specifically, the specific steps of using the multivariate regression method for geostress inversion include:

[0076] Step S1-1, based on engineering geological survey data, aerial data, elevation data, satellite photos and other data, a 3D refined geological topological model of the tunnel site is established using GIS method and finite element software;

[0077] Step S1-2, based on the on-site geostress test results and geomechanical analysis results, determine the main factors affecting the formation of the initial geostress field of the tunnel, obtain the stress values ​​of the known measuring points under the factors, and establish a regression equation with the corresponding measured stress values;

[0078] Step S1-3, using the least squares method principle to obtain the optimal solution of the coefficients of the individual variables (undetermined factors) in the multivariate regression equation, and then determine the geostress field of the tunnel site;

[0079] Step S1-4, according to the inversion results of the geostress field, determine the distribution law of the maximum principal stress, minimum principal stress, and intermediate principal stress of the tunnel axis, obtain the inversion value and direction of the tunnel geostress through coordinate transformation, and compare the measured results of the geostress. The error between the inversion value of the geostress and the measured value should be less than 10%; if it is greater than 10%, the cause should be found out and the inversion model should be continuously adjusted to ensure the accuracy of the inversion model.

[0080] In one of the embodiments, step 102 includes: establishing a tunnel numerical simulation model according to the tunnel face and surrounding rock conditions, engineering geological conditions of the tunnel site, characteristics of the ground stress field, tunnel surrounding rock test data and corresponding excavation and support methods; numerically simulating the tunnel through the tunnel numerical simulation model to obtain numerical simulation results; wherein the numerical simulation results include: tunnel vault settlement, clearance convergence and deformation of key areas; comparing the numerical simulation results with actual monitoring data to verify the accuracy of the tunnel numerical simulation model, and if the error between the numerical simulation results and the actual monitoring data is greater than or equal to a preset threshold, finding out the cause and continuously adjusting the numerical model to ensure the reliability of the numerical model.

[0081] Specifically, numerical simulation should be modeled in combination with the actual rock structure and structural surface distribution. The vault settlement, clearance convergence and deformation of key areas of the tunnel should be obtained by comparing the numerical simulation with the actual monitoring data to verify the accuracy of the numerical model. The error between the simulation results and the monitoring data should be less than 10%. Otherwise, the cause should be identified and the numerical model should be continuously adjusted to ensure its reliability.

[0082] In one of the embodiments, a data set required for surrounding rock parameter inversion analysis is obtained based on field monitoring data, surrounding rock test data and tunnel numerical simulation model, including: extracting surrounding rock parameters of each rock burst section tunnel based on field monitoring data, surrounding rock test data and tunnel numerical simulation model; combining the surrounding rock parameters of all rock burst sections tunnel into a data set required for surrounding rock parameter inversion analysis.

[0083] In one of the embodiments, a BP neural network is optimized by a particle swarm algorithm according to a data set to obtain a BP neural network prediction model optimized by a particle swarm; the tunnel surrounding rock parameters are inverted by using the BP neural network prediction model optimized by a particle swarm to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel, including: selecting an activation function and a loss function, analyzing and optimizing the model parameters of the BP neural network, and obtaining the optimal structure of the BP neural network; initializing parameters, setting the number of particles and the number of iterations, and normalizing and standardizing the variable data; constructing a BP neural network prediction model optimized by a particle swarm, using the mean square error MSE of the prediction model as the particle swarm fitness index, and using the mean square error MSE and the goodness of fit R 2as the prediction accuracy evaluation index; 20% of the data are randomly selected from the data set as the test set, and the rest of the data are used as the training set; the initial data are imported into the particle swarm optimized BP neural network prediction model, and then the training set is used to connect the particle swarm optimized BP neural network prediction model for training, and the test set is used to predict the trained particle swarm optimized BP neural network prediction model, and the predicted value of the surrounding rock parameters and the mean square error MSE are calculated, and then the mean square error MSE is returned to the PSO model as the fitness to continuously update the parameter combination (C, g), where C∈[0,5000], g∈[0,100], until the number of iterations is completed, and the particle swarm optimized BP neural network prediction model is obtained; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters, and the surrounding rock parameter inversion results of each rock burst section of the tunnel are obtained.

[0084] Specifically, the tunnel surrounding rock parameter inversion process based on particle swarm optimization BP neural network is as follows: Figure 2 shown.

[0085] The specific process of using particle swarm algorithm to optimize BP neural network includes:

[0086] Step S3-1, selecting appropriate activation function and loss function, analyzing and optimizing the model parameters of the BP neural network, such as the number of input layer nodes, the number of output layer nodes, the number of hidden layers, the number of hidden layer nodes and the learning rate, to obtain the optimal structure of the BP neural network;

[0087] Step S3-2, initializing parameters, setting the number of particles and the number of iterations, and normalizing and standardizing the variable data;

[0088] Step S3-3, establish various sub-functions in the Python platform: data normalization and standardization function, BP neural network prediction model of particle swarm optimization, fitness calculation function, particle swarm position and speed function, use the mean square error MSE of the prediction model as the particle swarm fitness index, and use the mean square error MSE and goodness of fit R2 as prediction accuracy evaluation indicators;

[0089] Step S3-4, import the initial data into the BP neural network prediction model of particle swarm optimization, and randomly extract 20% of the data as the test set, and the rest of the data as the training set. After the BP neural network model is trained, the test set is predicted, and the predicted values ​​of the surrounding rock parameters and the mean square error MSE are calculated. Then, the MSE is returned to the PSO model as the fitness to continuously update the parameter combination (C, g), where C∈[0,5000], g∈[0,100], until the number of iterations is completed;

[0090] Step S3-5: Output the optimal parameter combination (C, g) and the predicted values ​​of surrounding rock parameters, MSE and model determination coefficient R under this parameter combination.2 , compared the traditional BP neural network inversion results and measured data, analyzed the optimization ability of the particle swarm algorithm, and continuously improved the prediction accuracy of the particle swarm optimization BP neural network.

[0091] In one embodiment, step 106 includes: establishing a functional mapping relationship between the rockburst level of the constructed section of the tunnel and the in-situ stress and surrounding rock parameters according to the inversion results of the in-situ stress field and the in-situ stress parameters of each rockburst section of the constructed section of the tunnel; or comparing the prediction results of each rockburst criterion with the actual rockburst situation on site, and selecting the rockburst criterion that best suits the tunnel.

[0092] Specifically, the inversion accuracy of surrounding rock parameters based on the particle swarm algorithm to optimize the BP neural network increases with the increase in the number of samples, and accordingly, the accuracy of rockburst prediction also increases. In combination with the progress of tunnel construction, based on the increasing on-site monitoring values ​​and surrounding rock test results, the number of samples for surrounding rock parameter inversion analysis should be continuously increased to improve the surrounding rock inversion accuracy, continuously optimize the rockburst criterion, and improve the rockburst prediction accuracy.

[0093] In one of the embodiments, the method further includes: according to the progress of tunnel construction, based on the increasing on-site monitoring values ​​and surrounding rock test results, continuously increasing the number of samples for surrounding rock parameter inversion analysis, improving the surrounding rock inversion accuracy, continuously optimizing the rock burst criterion, and improving the rock burst prediction accuracy.

[0094] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0095] Specific implementation example: A certain plateau railway tunnel is located in the canyon mountainous area of ​​the western plateau. The tunnel site area is severely eroded, with rugged terrain, large undulations, steep mountains and deep valleys. The tectonic movement in the area is strong and complex, and the tectonic deformation has the characteristics of multi-level, multi-mechanism and multi-scale. The overall tectonic direction is northwest. The total length of the tunnel inclined shaft is about 3.4km, the entrance section is as high as 3800m above sea level, and the maximum burial depth exceeds 1500m. Since the inclined shaft is located in the deep buried section of the active tectonic zone of the plate, it has the characteristics of large burial depth, high altitude and high ground stress. In addition, the inclined shaft body is mainly composed of brittle hard rock such as granite. It faces serious rock burst disaster risks under high ground stress conditions. Therefore, accurate prediction of inclined shaft rock burst is crucial to ensure construction safety.

[0096] A deep buried tunnel rockburst prediction method based on inversion of ground stress and surrounding rock parameters proposed in this application is used to predict and analyze the rockburst level of the inclined shaft. First, a three-dimensional inversion model of ground stress in the tunnel site area is established based on step 100 of the present invention to obtain the change of the principal stress of the inclined shaft axis; the change of the principal stress of the inclined shaft axis is as follows: Figure 3 After coordinate transformation, the error between the inversion result of inclined well geostress and the measured value of on-site geostress is between 3.2% and 8.7%, which meets the engineering needs and error requirements.

[0097] Then, based on step 102 of this method, the numerical model of each construction section of the tunnel is established, the geostress field of the corresponding construction section obtained by geostress inversion and the surrounding rock parameters tested on site are applied, and the numerical simulation results are compared with the field monitoring results, as shown in Table 1. The results show that the error between the numerical simulation and the field monitoring results is within 10%, which meets the error requirements. After that, the constructed tunnel sections are simulated, and the data with an error of more than 10% between the simulated value and the measured value are eliminated, and 48 sets of data sets for surrounding rock parameter inversion are obtained.

[0098] Table 1 Comparison between numerical simulation and monitoring data

[0099]

[0100] Then based on step 104 of the present invention, refer to Figure 2 The figure shows the schematic diagram of the tunnel surrounding rock parameter inversion process based on the particle swarm algorithm to optimize the BP neural network. The particle swarm algorithm is used to optimize the BP neural network for surrounding rock parameter inversion, where the number of input layer nodes of the BP neural network is set to 2, the number of output layer nodes is 4, the number of hidden layers is 2, the number of hidden layer nodes is 8, and the learning rate is set to 0.01. At the same time, the Sigmoid function is selected as the activation function, and the Euclidean distance loss function is selected as the loss function to meet the analysis of small sample data.

[0101] After 12 iterations, the fitness value of the PSO algorithm gradually stabilized. At the 23rd iteration, the optimal fitness was MSE = 2.862, and the goodness of fit was R 2 =0.963. The optimal fitness of the traditional BP neural network is MSE = 6.324, and the goodness of fit is R 2 =0.827, it can be seen that the particle swarm algorithm can effectively reduce the number of iterations of searching for the global optimal parameters and avoid falling into the local optimum, thereby improving the prediction results of the BP neural network. The inversion method of the present invention is used to obtain that the surrounding rock strength at 0+810 of the inclined well is about 114.6MPa, and the uniaxial compressive strength of the rock sampled and tested on site is about 118.7MPa. The error between the inversion prediction result and the experimental test result is 3.45%, which meets the requirements.

[0102] Finally, the rockburst level of the entire section of the inclined shaft was predicted; the comparison between the rockburst prediction results and the actual rockburst situation on site is shown in the figure below. Figure 4 As shown. The accuracy of each rockburst prediction method is judged by synthesizing the rockburst prediction results, as shown in Table 2. The results show that the rockburst prediction accuracy of the inclined shaft obtained by inversion of surrounding rock parameters based on ground stress inversion and optimization of BP neural network model by PSO algorithm is 82.9%, while the rockburst prediction accuracy obtained by initial risk assessment of tunnel and Russenes criterion calculation is 51.4% and 45.7% respectively, which is much lower than the prediction accuracy of the present invention.

[0103] Table 2 Comparison of rockburst prediction results in inclined wells

[0104] Rockburst criterion Initial risk assessment PSO-BP Inversion Russenes criterion Accuracy 51.4％ 82.9％ 45.7

[0105] It can be seen that the present invention effectively combines tunnel field tests with monitoring data and engineering geological conditions, and can more effectively predict the rockburst level, providing a reference for on-site rockburst prevention and control construction.

[0106] In one embodiment, Figure 5 As shown, a deep buried tunnel rockburst prediction device is provided, comprising: a ground stress field inversion module, a tunnel numerical simulation model establishment module, a surrounding rock parameter inversion module, a rockburst criterion construction module and a deep buried tunnel rockburst prediction module, wherein:

[0107] The geostress field inversion module is used to perform geostress inversion based on the acquired engineering geological conditions of the tunnel site and the on-site geostress test results, using the multivariate regression method to obtain the geostress field inversion results of the tunnel site.

[0108] The tunnel numerical simulation model building module is used to build a tunnel numerical simulation model based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods.

[0109] The surrounding rock parameter inversion module is used to obtain the data set required for the surrounding rock parameter inversion analysis based on the field monitoring data, surrounding rock test data and the tunnel numerical simulation model; the particle swarm algorithm is used to optimize the BP neural network according to the data set to obtain the particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel.

[0110] The rockburst criterion construction module is used to establish the rockburst criterion applicable to the tunnel based on the inversion results of the ground stress field and the inversion results of the surrounding rock parameters of each rockburst section of the constructed section of the tunnel.

[0111] The deep-buried tunnel rockburst prediction module is used to perform numerical simulation on the current construction section of the tunnel using the tunnel numerical simulation model, and to perform surrounding rock parameter inversion using the particle swarm optimized BP neural network prediction model. According to the rockburst criterion and the surrounding rock parameter inversion results of the current construction section of the tunnel, the rockburst grade and occurrence probability of the current construction section of the tunnel are determined.

[0112] In one of the embodiments, the geostress field inversion module is also used to establish a three-dimensional refined geological topological model of the tunnel site area using GIS methods and finite element software based on aerial data, elevation data and satellite images in the project address survey data; determine the main factors affecting the formation of the initial geostress field of the tunnel based on the on-site geostress test results and geomechanical analysis results, and obtain the stress values ​​of known measuring points under the factors; establish a geostress field inversion multivariate regression equation for the tunnel site area based on the known measuring point stress values ​​and the corresponding measured stress values; use the least squares method to determine the optimal solution of the independent variable coefficients in the geostress field inversion multivariate regression equation for the tunnel site area, and then use the geostress field inversion multivariate regression equation for the tunnel site area to determine the geostress field inversion result of the tunnel site area; based on the geostress field inversion result, determine the distribution law of the maximum principal stress, minimum principal stress and intermediate principal stress of the tunnel axis, and obtain the tunnel geostress inversion value and direction through coordinate transformation; according to the error between the tunnel geostress inversion value and the geostress measured result and the preset error threshold, continuously adjust the inversion model to ensure the accuracy of the inversion model.

[0113] In one of the embodiments, the tunnel numerical simulation model establishment module is also used to establish a tunnel numerical simulation model based on the tunnel face and surrounding rock conditions, engineering geological conditions of the tunnel site, characteristics of the ground stress field, tunnel surrounding rock test data and corresponding excavation and support methods; the tunnel is numerically simulated through the tunnel numerical simulation model to obtain numerical simulation results; wherein the numerical simulation results include: tunnel vault settlement, clearance convergence and deformation of key areas; the numerical simulation results are compared with the actual monitoring data to verify the accuracy of the tunnel numerical simulation model; if the error between the numerical simulation results and the actual monitoring data is greater than or equal to a preset threshold, the cause is identified and the numerical model is continuously adjusted to ensure the reliability of the numerical model.

[0114] In one of the embodiments, the surrounding rock parameter inversion module is also used to extract the surrounding rock parameters of each rock burst section tunnel based on field monitoring data, surrounding rock test data and tunnel numerical simulation model; and the surrounding rock parameters of all rock burst sections tunnel are combined into a data set required for surrounding rock parameter inversion analysis.

[0115] In one embodiment, the surrounding rock parameter inversion module is also used to select appropriate activation function and loss function, analyze and optimize the model parameters of the BP neural network, and obtain the optimal structure of the BP neural network; initialize parameters, set the number of particles and the number of iterations, and normalize and standardize the variable data; construct a BP neural network prediction model optimized by particle swarm optimization, use the mean square error MSE of the prediction model as the particle swarm fitness index, and use the mean square error MSE and the goodness of fit R 2 as the prediction accuracy evaluation index; 20% of the data are randomly selected from the data set as the test set, and the rest of the data are used as the training set; the initial data are imported into the particle swarm optimized BP neural network prediction model, and then the training set is used to connect the particle swarm optimized BP neural network prediction model for training, and the test set is used to predict the trained particle swarm optimized BP neural network prediction model, and the predicted value of the surrounding rock parameters and the mean square error MSE are calculated, and then the mean square error MSE is returned to the PSO model as the fitness to continuously update the parameter combination (C, g) until the number of iterations is completed, and the particle swarm optimized BP neural network prediction model is obtained; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters, and the surrounding rock parameter inversion results of each rock burst section of the tunnel are obtained.

[0116] In one of the embodiments, a rockburst criterion construction module is used to establish a functional mapping relationship between the rockburst grade of the constructed section of the tunnel and the ground stress and surrounding rock parameters based on the ground stress field inversion results and the surrounding rock parameter inversion results of each rockburst section of the constructed section of the tunnel; or to compare the predicted results of each rockburst criterion with the actual rockburst situation on site to select the rockburst criterion most suitable for the tunnel.

[0117] In one of the embodiments, the device also includes an optimization module for continuously increasing the number of samples for surrounding rock parameter inversion analysis according to the progress of tunnel construction, based on the increasing on-site monitoring values ​​and surrounding rock test results, improving the surrounding rock inversion accuracy, continuously optimizing the rock burst criterion, and improving the rock burst prediction accuracy.

[0118] For the specific definition of the deep tunnel rockburst prediction device, please refer to the definition of the deep tunnel rockburst prediction method above, which will not be repeated here. Each module in the above-mentioned deep tunnel rockburst prediction device can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0119] 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 6 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through 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 the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a deep tunnel rockburst prediction method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0120] Those skilled in the art will understand that Figure 6 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 those shown in the figure, or combine certain components, or have a different arrangement of components.

[0121] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above method embodiment when executing the computer program.

[0122] 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 method embodiment are implemented.

[0123] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed 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, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0124] The technical features of the above embodiments may 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.

[0125] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for predicting rockburst in a deep tunnel, characterized in that: The method comprises: Based on the engineering geological conditions of the tunnel site and the on-site geostress test results, the geostress inversion was performed using the multivariate regression method to obtain the inversion results of the geostress field in the tunnel site. A tunnel numerical simulation model is established based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data, and corresponding excavation and support methods; According to the field monitoring data, the surrounding rock test data and the tunnel numerical simulation model, a data set required for the inversion analysis of the surrounding rock parameters is obtained; according to the data set, a particle swarm algorithm is used to optimize the BP neural network to obtain a particle swarm optimized BP neural network prediction model; the particle swarm optimized BP neural network prediction model is used to invert the tunnel surrounding rock parameters to obtain the inversion results of the surrounding rock parameters of each rock burst section of the tunnel; According to the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rockburst section of the tunnel under construction, the rockburst criterion applicable to the tunnel is established; The tunnel numerical simulation model is used to perform numerical simulation on the current construction section of the tunnel, and the BP neural network prediction model with particle swarm optimization is used to perform surrounding rock parameter inversion. According to the rock burst criterion and the obtained surrounding rock parameter inversion result of the current construction section of the tunnel, the rock burst grade and occurrence probability of the current construction section of the tunnel are determined.

2. The deep tunnel rockburst prediction method according to claim 1, characterized in that: Based on the engineering geological conditions of the tunnel site and the on-site geostress test results, the geostress inversion was performed using the multivariate regression method to obtain the inversion results of the geostress field in the tunnel site, including: Based on the aerial data, elevation data and satellite images in the project site survey data, a three-dimensional refined geological topological model of the tunnel site area was established using GIS methods and finite element software; According to the on-site geostress test results and geomechanical analysis results, determine the main factors affecting the formation of the initial geostress field of the tunnel, and obtain the stress values ​​of known measuring points under these factors; Establishing a multivariate regression equation for inverse geostress field in the tunnel site area according to the known stress values ​​of the measuring points and the corresponding measured stress values; The optimal solution of the independent variable coefficient in the inversion multivariate regression equation of the in-situ stress field of the tunnel site area is determined by using the least square method, and then the inversion result of the in-situ stress field of the tunnel site area is determined by using the inversion multivariate regression equation of the in-situ stress field of the tunnel site area; According to the inversion results of the geostress field, the distribution law of the maximum principal stress, minimum principal stress and intermediate principal stress of the tunnel axis is determined, and the inversion value and direction of the tunnel geostress are obtained through coordinate transformation; According to the error between the tunnel geostress inversion value and the measured geostress result and the preset error threshold, the inversion model is continuously adjusted to ensure the accuracy of the inversion model.

3. The deep tunnel rockburst prediction method according to claim 1, characterized in that: According to the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods, a tunnel numerical simulation model is established, including: A tunnel numerical simulation model is established based on the tunnel face and surrounding rock conditions, engineering geological conditions of the tunnel site, characteristics of the geostress field, tunnel surrounding rock test data and corresponding excavation and support methods; The tunnel is numerically simulated by the tunnel numerical simulation model to obtain numerical simulation results; wherein the numerical simulation results include: tunnel vault settlement, clearance convergence and deformation of key areas; The numerical simulation results are compared with the actual monitoring data to verify the accuracy of the tunnel numerical simulation model. If the error between the numerical simulation results and the actual monitoring data is greater than or equal to a preset threshold, the cause is identified and the numerical model is continuously adjusted to ensure the reliability of the numerical model.

4. The deep tunnel rockburst prediction method according to claim 1, characterized in that: According to the field monitoring data, surrounding rock test data and the tunnel numerical simulation model, the data set required for the surrounding rock parameter inversion analysis is obtained, including: Extracting surrounding rock parameters of each rockburst section of the tunnel according to the on-site monitoring data, surrounding rock test data and the tunnel numerical simulation model; The surrounding rock parameters of all rockburst sections of the tunnel are combined into a data set required for the surrounding rock parameter inversion analysis.

5. The deep tunnel rockburst prediction method according to claim 1, characterized in that: According to the data set, a BP neural network is optimized by using a particle swarm algorithm to obtain a BP neural network prediction model optimized by a particle swarm; the tunnel surrounding rock parameters are inverted by using the particle swarm optimized BP neural network prediction model to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel, including: Select activation function and loss function, analyze and optimize the model parameters of BP neural network, and obtain the optimal structure of BP neural network; Initialize parameters, set the number of particles and iterations, and normalize and standardize variable data; The BP neural network prediction model optimized by particle swarm optimization was constructed, and the mean square error (MSE) of the prediction model was used as the fitness index of the particle swarm. 2 As an indicator for evaluating prediction accuracy; Randomly extract 20% of the data from the dataset as a test set, and the rest of the data as a training set; The initial data is imported into the particle swarm optimized BP neural network prediction model, and then the training set is used to connect the particle swarm optimized BP neural network prediction model for training, and the test set is used to predict the trained particle swarm optimized BP neural network prediction model, and the predicted value of the surrounding rock parameters and the mean square error MSE are calculated, and then the mean square error MSE is returned to the PSO model as the fitness to continuously update the parameter combination (C, g) until the number of iterations is completed, and the particle swarm optimized BP neural network prediction model is obtained; The BP neural network prediction model based on particle swarm optimization is used to invert the surrounding rock parameters of the tunnel, and the inversion results of the surrounding rock parameters of each rock burst section of the tunnel are obtained.

6. The deep tunnel rockburst prediction method according to claim 1, characterized in that: Based on the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rockburst section of the tunnel construction section, the rockburst criterion applicable to the tunnel is established, including: According to the inversion results of the geostress field and the inversion results of the surrounding rock parameters of each rockburst section of the constructed section of the tunnel, a functional mapping relationship between the rockburst grade of the constructed section of the tunnel and the geostress and surrounding rock parameters is established; or the prediction results of each rockburst criterion are compared with the actual rockburst situation on site to select the most suitable rockburst criterion for the tunnel.

7. The deep tunnel rockburst prediction method according to claim 1, characterized in that: The method further comprises: According to the progress of tunnel construction, based on the increasing on-site monitoring values ​​and surrounding rock test results, the number of samples for surrounding rock parameter inversion analysis is continuously increased to improve the surrounding rock inversion accuracy, continuously optimize the rock burst judgment criteria, and improve the rock burst prediction accuracy.

8. A deep tunnel rockburst prediction device, characterized in that: The device comprises: The geostress field inversion module is used to perform geostress inversion using the multivariate regression method based on the acquired engineering geological conditions of the tunnel site and the on-site geostress test results, and obtain the inversion results of the geostress field in the tunnel site; The tunnel numerical simulation model building module is used to build a tunnel numerical simulation model based on the actual tunnel face and surrounding rock conditions, engineering geological conditions, geostress field characteristics, tunnel surrounding rock test data and corresponding excavation and support methods; The surrounding rock parameter inversion module is used to obtain the data set required for the surrounding rock parameter inversion analysis according to the field monitoring data, the surrounding rock test data and the tunnel numerical simulation model; optimize the BP neural network using the particle swarm algorithm according to the data set to obtain the particle swarm optimized BP neural network prediction model; use the particle swarm optimized BP neural network prediction model to invert the tunnel surrounding rock parameters to obtain the surrounding rock parameter inversion results of each rock burst section of the tunnel; The rockburst criterion building module is used to establish the rockburst criterion applicable to the tunnel based on the inversion results of the ground stress field and the inversion results of the surrounding rock parameters of each rockburst section of the tunnel construction section; The deep buried tunnel rockburst prediction module is used to perform numerical simulation on the current construction section of the tunnel using the tunnel numerical simulation model, and to perform surrounding rock parameter inversion using the particle swarm optimized BP neural network prediction model. According to the rockburst criterion and the obtained surrounding rock parameter inversion result of the current construction section of the tunnel, the rockburst grade and occurrence probability of the current construction section of the tunnel are determined.

9. 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 deep tunnel rockburst prediction method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the deep tunnel rockburst prediction method according to any one of claims 1 to 7 is implemented.

Images