Prediction Method, Device and Product for Creep Parameters of Phyllite Tunnel Surrounding Rock

Abstract

An embodiment of the present invention provides a method, device and product for predicting creep parameters of phyllite tunnel surrounding rock, which relates to the field of rock mechanics. In the embodiment of the present invention, first, the indoor creep test results of the rock are fitted with the creep damage constitutive equation to obtain the initial creep parameters of each monitoring point. Then, based on the above initial creep parameters, multiple groups of surrounding rock creep parameters are generated through an orthogonal experimental scheme. Numerical simulation calculations are respectively performed on the multiple groups of surrounding rock creep parameters and the initial creep parameters of other monitoring points to generate multiple groups of training samples corresponding to each monitoring point. A neural network inversion model of the surrounding rock is established through a neural network model. Finally, the actual monitored deformation amount at the tunnel site is substituted into the inversion model to invert the creep parameters of the surrounding rock. In the embodiment of the present invention, according to the engineering geological characteristics of the surrounding rock at each monitoring point, the control variable method is used to establish the inversion models of the surrounding rock near each monitoring point respectively, so as to improve the accuracy of inversion.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of rock mechanics, and in particular to a method, device and product for predicting creep parameters of surrounding rock in phyllite tunnels. Background Art

[0002] After the tunnel starts to be excavated, the stress state of the surrounding rock changes. Coupled with the complexity and concealment of tunnel engineering, it is difficult to accurately obtain the creep parameters of the surrounding rock after excavation. The parameters obtained only based on theoretical and simple indoor experiments often deviate greatly from the actual surrounding rock parameters and are difficult to meet the requirements. With the development and wide application of intelligent analysis in recent years, more and more researchers use the method of combining on-site monitoring and neural network intelligent analysis to invert the creep parameters of tunnel surrounding rock.

[0003] Traditional neural network methods establish a unified parameter inversion model for the whole tunnel surrounding rock. Usually, a large amount of sample data is required to train an inversion model, but generally empirical values are taken as input parameters, which cannot reflect the real situation of the surrounding rock, resulting in problems such as inconsistent neural network structures, low training efficiency, and low accuracy.

[0004] Therefore, there is an urgent need for a new method for predicting creep parameters of surrounding rock in phyllite tunnels. Summary of the Invention

[0005] The embodiments of the present invention provide a method, device and product for predicting creep parameters of surrounding rock in phyllite tunnels to at least partially solve the problems existing in the related art.

[0006] The first aspect of the embodiments of the present invention provides a method for predicting creep parameters of surrounding rock in phyllite tunnels, the method comprising:

[0007] Monitoring the deformation of the current surrounding rock monitoring point, the deformation including: side wall convergence deformation, right arch waist settlement, crown settlement, left arch waist settlement, arch waist convergence deformation;

[0008] Inputting the actually monitored deformation into the neural network inversion model corresponding to the current surrounding rock monitoring point to obtain the creep parameters of the current surrounding rock monitoring point;

[0009] Wherein, a set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and their corresponding deformations, and the corresponding deformations are obtained by simulating and calculating the creep parameters of the current surrounding rock monitoring point and the initial creep parameters of other surrounding rock monitoring points; the creep parameters of the current surrounding rock monitoring point are obtained through an orthogonal experimental scheme based on the initial creep parameters of the current surrounding rock monitoring point;

[0010] The initial creep parameters of the surrounding rock monitoring points are obtained from the results of the triaxial creep experiment of saturated phyllite under seepage-stress conditions and the creep constitutive equation of the surrounding rock of the phyllite tunnel under seepage-stress conditions:

[0011] ε = Aσ e + BS 11 + CS 11 [1 - e -Ft + MS 11 D n + NS 11 D g ,σ≥σ s2 obtained,

[0012] where, σ e is the effective stress, S 11 is the effective deviator stress, D * represents the effective damage amount, which is determined by the results of the triaxial creep experiment of saturated phyllite under seepage-stress conditions. Without seepage conditions, D * = 0; σ 3 represents the minimum principal stress of the in-situ stress of the surrounding rock of the tunnel, σ 1 represents the maximum principal stress of the in-situ stress of the surrounding rock of the tunnel, P represents the osmotic pressure, σ s2 represents the long-term strength of the rock; A, B, C, F, M, N are creep parameters; D n is the crack initiation damage function, which is a linear evolution law. Let D n = kt, where k is the damage parameter affected by the ratio of the deviator stress to the short-term peak strength; D g is the accelerating damage function, which is a power function evolution law. Let D g = mt n , where m, n are accelerating damage parameters; the damage evolution law curve is obtained from the indoor acoustic emission experiment, and the damage parameters k, m, n are obtained by fitting.

[0013] Optionally, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point include:

[0014] Obtain multiple groups of training data for the neural network inversion model corresponding to the current surrounding rock monitoring point;

[0015] Use the multiple groups of training data to train the BP model to obtain the neural network inversion model corresponding to the current surrounding rock monitoring point.

[0016] Optionally, obtaining multiple groups of training data for the neural network inversion model corresponding to the current surrounding rock monitoring point includes:

[0017] Perform a triaxial creep experiment on the surrounding rock at each surrounding rock monitoring point under seepage-stress conditions for saturated phyllite, and fit the experimental results with the creep constitutive equation of the phyllite tunnel surrounding rock under the seepage-stress conditions to obtain the initial creep parameters of each detection point;

[0018] Fix the initial creep parameters of other monitoring points unchanged, and generate a creep parameter sample set for the current surrounding rock monitoring point based on the initial creep parameters of the current monitoring point and the orthogonal experimental design;

[0019] Input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the numerical model respectively to calculate the corresponding deformation amount, and use each set of creep parameters and the corresponding deformation amount as training data, and the numerical model is the GDEM numerical model.

[0020] Optionally, the training step of the neural network inversion model corresponding to the current surrounding rock monitoring point further includes:

[0021] Input the test deformation amount into the neural network inversion model corresponding to the currently trained surrounding rock monitoring point, inversely obtain the corresponding test creep parameters, and input the test creep parameters into the numerical model to obtain the simulated deformation amount;

[0022] Compare the error between the test deformation amount and the simulated deformation amount;

[0023] In the case that the error is greater than the preset threshold, adjust the numerical model, use the adjusted numerical model to re-determine the training data, and re-train the BP model.

[0024] Optionally, inputting each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the numerical model respectively to calculate the corresponding deformation amount includes:

[0025] Embed the creep constitutive equation of the phyllite tunnel surrounding rock under the seepage-stress conditions into the GDEM numerical simulation program;

[0026] Input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the GDEM numerical simulation program respectively to calculate the corresponding deformation amount.

[0027] The second aspect of the embodiments of the present invention provides a device for predicting the creep parameters of the phyllite tunnel surrounding rock, and the device includes:

[0028] A monitoring module for monitoring the deformation amount of the current surrounding rock monitoring point, and the deformation amount includes: side wall convergence deformation amount, right arch waist settlement amount, crown settlement amount, left arch waist settlement amount, arch waist convergence deformation amount;

[0029] A prediction module, configured to input the actually monitored deformation amount into a neural network inversion model corresponding to the current surrounding rock monitoring point, and obtain the creep parameters of the current surrounding rock monitoring point;

[0030] A set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and the corresponding deformation amount, which is obtained by simulating the creep parameters of the current surrounding rock monitoring point and the initial creep parameters of other surrounding rock monitoring points; the creep parameters of the current surrounding rock monitoring point are obtained through an orthogonal experimental scheme based on the initial creep parameters of the current surrounding rock monitoring point;

[0031] The initial creep parameters of the surrounding rock monitoring point are obtained from the triaxial creep experimental results of saturated phyllite under seepage-stress conditions and the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions:

[0032] ε=Aσ e +BS 11 +CS 11 [1-e -Ft +MS 11 D n +NS 11 D g , σ≥σ s2 obtained,

[0033] wherein, σ e is the effective stress, S 11 is the effective deviator stress, D * represents the effective damage amount, which is determined by the triaxial creep experimental results of saturated phyllite under seepage-stress conditions. If there is no seepage condition, then D * =0; σ 3 represents the minimum principal stress of the in-situ stress of the tunnel surrounding rock, σ 1 represents the maximum principal stress of the in-situ stress of the tunnel surrounding rock, P represents the osmotic pressure, σ s2 represents the long-term strength of the rock; A, B, C, F, M, N are creep parameters; D n is the crack initiation damage function, which is a linear evolution law. Let D n =kt, where k is a damage parameter affected by the ratio of deviator stress to short-term peak strength; D g is the accelerated damage function, which is a power function evolution law. Let D g =mt n , where m and n are accelerated damage parameters; the damage evolution law curve is obtained from the indoor acoustic emission experiment, and the damage parameters k, m, and n are obtained by fitting.

[0034] Optionally, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point include:

[0035] Obtain multiple groups of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point;

[0036] Use the multiple groups of training data to train the BP model to obtain a neural network inversion model corresponding to the current surrounding rock monitoring point.

[0037] Optionally, obtaining multiple groups of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes:

[0038] Conduct a seepage-stress condition saturated phyllite triaxial creep experiment on the surrounding rock at each surrounding rock monitoring point, fit the experimental results with the phyllite tunnel surrounding rock creep constitutive equation under the seepage-stress condition to obtain the initial creep parameters of each detection point;

[0039] Fix the initial creep parameters of other monitoring points unchanged, and generate a creep parameter sample set for the current surrounding rock monitoring point based on the initial creep parameters of the current monitoring point and the orthogonal experimental design;

[0040] Input each group of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the numerical model respectively to calculate the corresponding deformation amount, and use each group of creep parameters and the corresponding deformation amount as training data, where the numerical model is the GDEM numerical model.

[0041] Optionally, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point further include:

[0042] Input the test deformation amount into the neural network inversion model corresponding to the current surrounding rock monitoring point obtained by training, inversely obtain the corresponding test creep parameters, and input the test creep parameters into the numerical model to obtain the simulated deformation amount;

[0043] Compare the error between the test deformation amount and the simulated deformation amount;

[0044] In the case where the error is greater than the preset threshold, adjust the numerical model, use the adjusted numerical model to re-determine the training data, and re-train the BP model.

[0045] Optionally, inputting each group of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the numerical model respectively to calculate the corresponding deformation amount includes:

[0046] Embed the phyllite tunnel surrounding rock creep constitutive equation under the seepage-stress condition into the GDEM numerical simulation program;

[0047] Input each group of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the GDEM numerical simulation program respectively, and calculate the corresponding deformation amounts.

[0048] In the third aspect of the embodiments of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the steps in the method described in the first aspect of the present invention are implemented.

[0049] In the fourth aspect of the embodiments of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes, the steps in the method described in the first aspect of the present invention are implemented.

[0050] In the embodiments of the present invention, first, the indoor creep experiment results of the rock are fitted with the creep damage constitutive equation to obtain the initial creep parameters of each monitoring point. Then, based on the above initial creep parameters, multiple groups of surrounding rock creep parameters are generated through an orthogonal experimental scheme. Numerical simulation calculations are respectively performed on the multiple groups of surrounding rock creep parameters and the initial creep parameters of other monitoring points to generate multiple groups of training samples corresponding to each monitoring point. A neural network inversion model of the surrounding rock is established through a neural network model. Finally, the actual monitored deformation amount at the tunnel site is substituted into the inversion model to invert the surrounding rock creep parameters. In the embodiments of the present invention, an inversion model of the surrounding rock near each monitoring point is respectively established by using the control variable method according to the engineering geological characteristics of the surrounding rock at each monitoring point, so as to improve the accuracy of inversion. Description of the Drawings

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0052] Figure 1 is a flowchart of a method for predicting the creep parameters of the surrounding rock of a phyllite tunnel in the embodiments of the present invention;

[0053] Figure 2 is a flowchart of a method for predicting the creep parameters of the surrounding rock of a phyllite tunnel in the embodiments of the present invention;

[0054] Figure 3 is a schematic diagram of the monitored values of the settlement deformation of each monitoring point in an implementation case of the embodiments of the present invention;

[0055] Figure 4 is a schematic diagram of the simulated deformation results of the surrounding rock on the 10th day in an implementation case of the embodiments of the present invention;

[0056] Figure 5 It is a schematic diagram of the simulated deformation result of the surrounding rock on the 30th day in an implementation case of an embodiment of the present invention;

[0057] Figure 6 It is a schematic diagram of the simulated deformation result of the surrounding rock on the 60th day in an implementation case of an embodiment of the present invention;

[0058] Figure 7 It is a structural block diagram of a creep parameter prediction device for the surrounding rock of a phyllite tunnel in an embodiment of the present invention. Specific implementation manner

[0059] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0060] Refer to Figure 1 , which shows a flowchart of a creep parameter prediction method for the surrounding rock of a phyllite tunnel in an embodiment of the present invention. The creep parameter prediction method provided by the embodiment of the present invention may include the following steps:

[0061] S101, monitor the deformation amount of the current surrounding rock monitoring point.

[0062] Among them, the deformation amount includes: the convergence deformation amount of the side wall, the settlement amount of the right arch waist, the settlement amount of the crown, the settlement amount of the left arch waist, and the convergence deformation amount of the arch waist.

[0063] In the embodiment of the present invention, creep parameter prediction is performed separately for each monitoring point of the surrounding rock of the phyllite tunnel.

[0064] In the embodiment of the present invention, for the surrounding rock of the phyllite tunnel, the convergence deformation amount of the side wall, the settlement amount of the right arch waist, the settlement amount of the crown, the settlement amount of the left arch waist, and the convergence deformation amount of the arch waist can be monitored respectively, as the deformation amounts corresponding to the monitoring points at the five positions of the side wall, the right arch waist, the crown, the left arch waist, and the arch waist.

[0065] S102, input the actually monitored deformation amount into the neural network inversion model corresponding to the current surrounding rock monitoring point to obtain the creep parameter of the current surrounding rock monitoring point.

[0066] Among them, a set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameter of the current surrounding rock monitoring point and the deformation amount obtained by simulating and calculating the creep parameter of the current surrounding rock monitoring point.

[0067] Among them, a set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and their corresponding deformation amounts, which are obtained by simulating the creep parameters of the current surrounding rock monitoring point and the initial creep parameters of other surrounding rock monitoring points; the creep parameters of the current surrounding rock monitoring point are obtained through an orthogonal experimental scheme based on the initial creep parameters of the current surrounding rock monitoring point;

[0068] The initial creep parameters of the surrounding rock monitoring point are obtained from the triaxial creep experimental results of saturated phyllite under seepage-stress conditions and the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions:

[0069] ε=Aσ e +BS 11 +CS 11 [1-e -Ft +MS 11 D n +NS 11 D g ,σ≥σ s2 obtained,

[0070] Among them, σ e is the effective stress, S 11 is the effective deviator stress, D * represents the effective damage amount, which is determined by the triaxial creep experimental results of saturated phyllite under seepage-stress conditions. If there is no seepage condition, then D * =0;σ 3 represents the minimum principal stress of the in-situ stress of the tunnel surrounding rock, σ 1 represents the maximum principal stress of the in-situ stress of the tunnel surrounding rock, P represents the osmotic pressure, σ s2 represents the long-term strength of the rock; A, B, C, F, M, N are creep parameters; D n is the crack initiation damage function, which is a linear evolution law. Let D n =kt, where k is the damage parameter affected by the ratio of deviator stress to short-term peak strength; D g is the accelerated damage function, which is a power function evolution law. Let D g =mt n , m, n are accelerated damage parameters; the damage evolution law curve is obtained from the indoor acoustic emission experiment, and the damage parameters k, m, n are obtained by fitting.

[0071] In the embodiment of the present invention, triaxial creep tests are carried out on saturated phyllite, controlling the osmotic pressure and confining pressure to be the current osmotic pressure and current confining pressure. The creep loading still adopts the Chen's loading method, and then 20%, 40%, 60%, and 80% of the short-term peak strength of the rock sample are determined for stepwise loading of differential stress (σ 1 -σ3 ) The holding time for each level of loading is 24 hours. If failure occurs before the last level of loading, the test stops loading. If the specimen does not show instability failure under the action of the last level of loading, the load is increased according to the actual situation until the rock sample fails. During the experiment, the transient method is used to collect the permeability of the rock sample, and the damage events of the rock sample are monitored.

[0072] The main steps of the experiment are as follows:

[0073] ① First, dry the phyllite specimen and then place it in a vacuum saturation device for vacuum saturation.

[0074] ② Installation and fixation of the rock sample. After sealing the saturated specimen with high-performance water-proof rubber, place it in the triaxial pressure chamber, seal the connection between the specimen and the upper and lower permeable plates with waterproof tape, then install sealing rings at the upper and lower ends, install displacement extensometers, and connect and check the water inlet and outlet.

[0075] ③ Lower the outer wall of the MTS triaxial pressure chamber, and evenly arrange 4 ACk-800 type acoustic emission probes at the upper and lower parts of the outer wall. Use vaseline as the coupling agent, turn on the oil pressure pump, and apply confining pressure to the rock sample.

[0076] ④ Push the distilled water in the water inlet chamber through the piston in the seepage system to apply a certain osmotic pressure to the specimen, keep the confining pressure and osmotic pressure unchanged, apply each level of axial load, and then keep the axial load unchanged. Use the transient method to collect the permeability of the rock sample.

[0077] ⑤ Repeat step 4. According to the step-by-step loading scheme, record and collect the permeability data until the rock sample undergoes accelerated creep failure, and finally summarize and organize the experimental data.

[0078] In the embodiment of the present invention, during the creep experiment, an indoor acoustic emission experiment is carried out to monitor the acoustic emission events, obtain the damage evolution law curve, and obtain the damage parameters k, m, and n by fitting.

[0079] In the embodiment of the present invention, first, the indoor creep experiment results of the rock are fitted with the creep damage constitutive equation to obtain the initial creep parameters of each monitoring point. Then, based on the above initial creep parameters, multiple groups of surrounding rock creep parameters are generated through an orthogonal experimental scheme. Numerical simulation calculations are carried out on multiple groups of surrounding rock creep parameters and the initial creep parameters of other monitoring points respectively to generate multiple groups of training samples corresponding to each monitoring point. A neural network inversion model of the surrounding rock is established through a neural network model. Finally, the actual monitored deformation amount at the tunnel site is substituted into the inversion model to invert the surrounding rock creep parameters.

[0080] Due to the complexity of surrounding rock in actual engineering, even at different positions of the same cross-section, the factors affecting the deformation of surrounding rock are different, and at the same time, the amount of data actually obtained is often limited. Therefore, in the embodiments of the present invention, according to the engineering geological characteristics of the surrounding rock at each monitoring point, the control variable method can be used to establish the inversion models of the surrounding rock near each monitoring point respectively, so as to improve the accuracy of inversion.

[0081] Refer to Figure 2 , which shows the flow chart of the training steps of the neural network inversion model in a method for predicting creep parameters of phyllite tunnel surrounding rock according to an embodiment of the present invention. The training steps of the neural network inversion model provided by the embodiment of the present invention may include:

[0082] S201, establish a BP model using the matlab neural network toolbox. The input nodes of the neural network inversion model are 5, the number of hidden layer nodes is 13, and the output nodes are 6.

[0083] In the embodiments of the present invention, using the BP neural network toolbox in MATLAB, there are 5 input layer nodes (corresponding to the deformation index of the surrounding rock), 6 output layer nodes (corresponding to each surrounding rock creep parameter), and the number of intermediate hidden layer nodes is estimated by the following empirical formula:

[0084]

[0085] In formula (1-1), a is the number of input layer nodes, n is the number of output layer nodes, and i is an integer between 1 and 10. The optimal number is calculated by trial according to the specific neural network fitting result. After trial calculation, taking the number of hidden layer nodes as 13 can make the fitting result reach the best.

[0086] S202, obtain multiple groups of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point.

[0087] Specifically, the step S202 includes the following sub-steps:

[0088] S2021, conduct a saturated phyllite triaxial creep experiment on the surrounding rock at each surrounding rock monitoring point under seepage-stress conditions, and fit the experimental results with the creep constitutive equation of the phyllite tunnel surrounding rock under seepage-stress conditions to obtain the initial creep parameters of each detection point.

[0089] In the embodiments of the present invention, a creep-acoustic emission experiment on saturated phyllite can be conducted on the surrounding rock at the surrounding rock monitoring point under a confining pressure of 15 MPa, and the experimental results are fitted with the creep constitutive equation of the phyllite tunnel surrounding rock to obtain the creep damage parameters of the rock samples at each measuring point as the initial creep parameters.

[0090] In the embodiments of the present invention, through theoretical derivation and experimental analysis, the three-dimensional creep damage constitutive equation of phyllite under seepage-stress conditions is determined as follows:

[0091]

[0092] The undetermined coefficients included in Equation (2-1) are K, P, D * , G 0 , G 1 , η 1 , η 2 , η 3 , M, H, where P represents the osmotic pressure; represents the effective damage amount, which is affected by seepage. If there is no seepage condition, then D * = 0; α represents the load level strength parameter. The creep constitutive equation of rock can be divided into four parts:

[0093] (1) Instantaneous elastic strain: The instantaneous elastic deformation of rock is mainly related to its own elastic modulus and the effective stress it receives. Generally, it is the initial value of the creep curve, corresponding to the first and second terms of Equation (2-1).

[0094] (2) Decay creep: It is related to its own viscoelastic modulus, viscoelastic coefficient, and the deviatoric stress it receives. The strain growth curve is a convex curve with a decreasing slope, corresponding to the third term of Equation (2-1).

[0095] (3) Crack initiation (constant velocity) damage creep: It is related to the deviatoric stress it receives, its own viscosity coefficient, and the degree of crack initiation damage. The strain growth curve is a linear function, corresponding to the fourth term of Equation (2-1).

[0096] (4) Accelerated damage creep: It is related to the deviatoric stress it receives, the viscosity coefficient, and the degree of accelerated damage. The strain growth curve is a concave curve with an increasing slope, which can be characterized by an exponential function, corresponding to the fifth term of Equation (2-1).

[0097] In the embodiments of the present invention, it is explored and determined that the strength of rock materials deteriorates according to the damage evolution law. When the rock strength decreases, the creep strain increases accordingly. Therefore, the creep strain law and the damage evolution law are highly positively correlated, and the strain can be approximately characterized by a damage function. After combining and arranging the parameters in Equation (2-1), the creep constitutive equation of the phyllite tunnel surrounding rock is obtained:

[0098] ε = Aσ e + BS 11 + CS 11 [1 - e -Ft + MS 11 D n + NS 11 D g , σ ≥ σ s2(2-2)

[0099] Among them, σ e is the effective stress, S 11 is the effective deviator stress, D * represents the effective damage amount, which is determined by indoor seepage experiments. If there is no seepage condition, then D * = 0; σ s2 represents the long-term strength of the rock; A, B, C, F, M, and N are all undetermined parameters (i.e., the creep parameters that need to be predicted in the embodiments of the present invention). A and B are determined by the instantaneous elastic stage, C and F are determined by the attenuation creep strain curve, M is determined by the isochronous creep strain curve, and N is determined by the accelerated damage creep strain curve; D n is the crack initiation damage function, which is a linear evolution law. Let D n = kt, where k is a damage parameter affected by the ratio of deviator stress to short-term peak strength; D g is the accelerated damage function, which is a power function evolution law. Let D g = mt n , where m and n are accelerated damage parameters; the damage evolution law curve is obtained from indoor acoustic emission experiments, and the damage parameters k, m, and n are obtained by fitting.

[0100] S2022, keeping the initial creep parameters of other monitoring points unchanged, a creep parameter sample set of the current surrounding rock monitoring point is generated based on the initial creep parameters of the current monitoring point and the orthogonal experimental design.

[0101] In the embodiments of the present invention, the surrounding rock is regarded as a rock that has undergone creep damage. The parameter range of the indoor experiment covers the parameters of the surrounding rock. Based on this assumption, through the orthogonal design scheme, training samples of the creep parameters of the surrounding rock are generated.

[0102] Specifically, in order to generate the sample parameters corresponding to the surrounding rock of each monitoring point, the embodiments of the present invention use the equal division of the magnitude difference as the parameter floating range, extract the maximum and minimum values of each initial creep parameter of each monitoring point to obtain the parameter floating range, and then obtain its creep parameter sample based on the initial creep parameters of the current monitoring point.

[0103] For example: among 5 monitoring points, the maximum and minimum values of the initial creep parameter corresponding to the creep parameter A are 0.018 and 0.008 respectively. After subtracting the two and dividing by 5 (there are 5 groups of data in total), 0.005 is obtained. Among them, the parameter data A of the 4th monitoring point is 0.008. Then the parameter sample takes values within the range of 0.008 ± 0.0025 (0.005 divided by 2). By analogy, the creep parameter sample of the surrounding rock of the 4th measuring point is obtained, and each parameter takes 5 values.

[0104] In 2023, each set of creep parameters in the creep parameter sample set is respectively input into the numerical model to calculate the corresponding deformation amount, and each set of creep parameters and the corresponding deformation amount are used as training data.

[0105] In the embodiment of the present invention, the training data of the inversion model for each monitoring point is constructed by the method of controlling variables, specifically: keeping the initial creep parameters of other monitoring points unchanged, generating multiple sets of creep parameters for the current monitoring point based on the initial creep parameters of the current monitoring point and the orthogonal experimental design, and performing numerical simulation based on each set of creep parameters and the initial creep parameters of other monitoring points to obtain the training data of the inversion model for the current monitoring point.

[0106] Specifically, the step S2023 includes the following sub-steps:

[0107] S20231, embed the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions in the GDEM numerical simulation program.

[0108] GDEM is a continuous-discontinuous coupling analysis software based on multi-core CPU parallelism developed by the Institute of Mechanics, Chinese Academy of Sciences. This software can perform finite element analysis on unit blocks, and can also realize progressive failure analysis between blocks, achieving coupled calculation of materials in continuous and discontinuous states. GDEM integrates finite element, discrete element and particle discrete element, expanding its application fields and making the numerical results more in line with reality. It can simultaneously support interactive coupling of seepage, temperature, stress, fracture, etc., and can simulate the fracture and fragmentation process of materials under multi-field coupling and the collision and accumulation process of fragment groups. It can be widely applied in many fields such as geotechnical engineering, geological disasters, mining, tunnels, blasting, hydropower, energy, etc. The GDEM software is mainly divided into two software modules: GDEM-BlockDyna block analysis software and GDEM-PDyna particle discrete element analysis.

[0109] The interface contact model built in GDEM is set according to brittle fracture characteristics. When the interface spring force reaches the tensile or shear threshold, the program will simultaneously clear the cohesion C and tensile strength T of the interface spring. The specific settings are as follows:

[0110] (1) Tensile failure: When the normal stress (assuming the tensile stress is negative), then set the spring normal stress = shear stress C = 0, T = 0.

[0111] (2) Shear failure: When the shear stress Then set

[0112] To achieve the two damage mechanisms during the creep process, it can be realized by changing the constitutive model of the contact surface between simulation units. In the embodiments of the present invention, the crack initiation damage function is set to a linear law related to time, and the accelerated propagation damage function is set to a power function law related to time. Therefore, during the creep simulation process, when in the crack initiation (tensile) damage stage, the tensile strength T of the spring unit between the contact surfaces linearly deteriorates from the initial strength to 0; when in the propagation (shear) damage stage, the shear strength C of the spring unit non-linearly decays from the initial strength to 0, specifically as follows:

[0113] (1) In the crack initiation (tensile) damage stage: The crack initiation damage has a linear relationship with time, D n = kt, where k is a damage parameter affected by the ratio of deviator stress to short-term peak strength, mainly related to the tensile strength. When the normal stress on the contact surface (assuming the tensile stress is negative), the function of the tensile strength T value is:

[0114] T = (1 - D n )T 0 = (1 - kt)T 0 (231 - 1)

[0115] In formula (231 - 1), T 0 represents the tensile strength at the initial moment, and C 0 represents the cohesion at the initial moment. When T decays to 0, then:

[0116] (2) In the propagation (shear) damage stage: The propagation damage has a power function relationship with time, D g = mt n , where m and n are accelerated damage parameters, mainly related to the shear strength. When the shear stress is present, the variation function of the cohesion C value is:

[0117] C = 1 - D g )C 0 = (1 - mt n )C 0 (231 - 2)

[0118] When C decays to 0, then:

[0119] Open the CustomModel.sln program of the GDEM software on the Visual Studio 2008 platform, modify the contact constitutive program CustomModel_Interface therein, incorporate the above-mentioned contact constitutive program (where the creep time is defined as a global variable for convenient calculation), and then embed the creep constitutive equation of phyllite tunnel surrounding rock in the block constitutive program CustomModel_Element.

[0120] After successfully compiling the program using the VS platform, place the.dll file in the numerical model folder. In the GDEM software, use the command blkdyn.LoadUDF(“CustomModel.dll”) to load the custom constitutive dynamic link library. Set the basic material parameters through blkdyn.SetMatByGroup(<>) and set the parameters of the user-defined constitutive through blkdyn.SetUDFValue(<>) Then, use the command blkdyn.SetIModel("Custom") to call the custom contact constitutive model, and use the command blkdyn.SetModel("Custom") to call the custom block constitutive model. In this way, during the core calculation, BlockDyna will automatically call the user-defined contact constitutive and block constitutive for calculation.

[0121] S20232, input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the GDEM numerical simulation program respectively, and calculate the corresponding deformation amounts.

[0122] S203, use the multi-group of training data to train the BP model to obtain a neural network inversion model corresponding to the current surrounding rock monitoring point.

[0123] In the embodiment of the present invention, 70% of the training data can be taken as the training samples, 15% of the data can be taken as the validation sample set, and 15% can be taken as the test sample set to train the neural network inversion model to obtain a neural network inversion model corresponding to the current surrounding rock monitoring point.

[0124] In the embodiment of the present invention, the training samples of each surrounding rock monitoring point can be generated in sequence through the above steps S201~S203, and the neural network inversion model corresponding to each surrounding rock monitoring point can be trained, so that the actual monitored deformation amount can be input into the neural network inversion model corresponding to each surrounding rock monitoring point respectively to obtain the creep parameters of the surrounding rock at each point.

[0125] In the embodiments of the present invention, for each surrounding rock monitoring point, based on the indoor creep test results of the rock at this monitoring point, the initial creep parameters are determined. Then, based on the initial creep parameters, the orthogonal experimental scheme, and numerical simulation calculations, a plurality of training samples corresponding to this monitoring point are generated. Then, based on these training samples, a neural network inversion model corresponding to this monitoring point is trained. Thus, based on this neural network inversion model, during the tunnel excavation process, based on the monitored deformation amount, the creep parameters corresponding to this monitoring point can be accurately predicted. Thus, the actual creep parameters during the excavation process can be obtained, and based on these actual creep parameters, a basis can be provided for the creep prediction and analysis of the tunnel surrounding rock.

[0126] For ease of understanding, the following further explains a method for predicting the creep parameters of the surrounding rock of a phyllite tunnel provided by the embodiments of the present invention through a specific embodiment. It can be understood that this embodiment is only an example:

[0127] Taking the large deformation mileage section of the Shiziping Tunnel project on the Wenchuan - Barkam Expressway as an example for analysis. This tunnel is an extra - long left - right separated tunnel project (the left tunnel is 5695 m long, and the right tunnel is 5672 m long). Large deformations occurred in many sections, resulting in circumferential cracks at multiple points along the initial support arch, and local cracks and spalling on the side walls and the top of the tunnel.

[0128] Select the surrounding rock (taking the section ZK145 + 620 as an example) near ZK145 + 630 in the large deformation mileage section of the Shiziping Tunnel for inversion section to predict and analyze the deformation, and at the same time verify the applicability of the neural network inversion model established in the embodiments of the present invention. The geological conditions of the ZK145 + 620 section are basically the same as those of the inversion section. The results of each monitoring point are shown in Figure 3 . Figure 3 Shows the monitored values of the settlement deformation of each monitoring point in the ZK145 + 620 section after excavation. Among them, data1 represents the convergence deformation of the side wall; data2 represents the settlement amount of the right arch waist; data3 represents the settlement amount of the crown; data4 represents the settlement amount of the left arch waist; data5 represents the convergence deformation amount of the arch waist.

[0129] Input the cumulative deformation amount monitored on the 30th day into the inversion models corresponding to each measuring point, obtain the creep parameters of the surrounding rock at each measuring point of this section, and then assign the parameters of each point to the numerical model of the corresponding area for calculation to obtain the time - history displacement nephogram of the surrounding rock ( Figures 4 to 6 ), Figure 4 Shows the simulated deformation results of the surrounding rock on the 10th day; Figure 5 Shows the simulated deformation results of the surrounding rock on the 30th day; Figure 6 Shows the simulated deformation results of the surrounding rock on the 60th day.

[0130] Extract the results of the numerically simulated surrounding rock deformation at different times, and compare them with the actual monitoring values (as shown in the following table) to verify the accuracy of the inverted parameters.

[0131] Comparison between the surrounding rock deformation monitoring data and the predicted values

[0132]

[0133]

[0134] It can be seen from the table that the difference between the monitoring values and the predicted values in the first 20 days is relatively large, with an error of nearly 30%. This is because after the tunnel is excavated, the surrounding rock undergoes rapid deformation in a short period of time, while the inversion model is based on creep, so the difference is large. As time goes by, in the time period from 40 days to 60 days, the error between the monitoring values and the predicted values is within 20%, and the deformation of the surrounding rock can be predicted more accurately, indicating that the prediction model has good applicability in the creep stage of the surrounding rock.

[0135] In subsequent research and analysis, based on this method, a surrounding rock monitoring - inverted parameter system for large - deformation sections can be established to achieve the purpose of dynamically grasping the surrounding rock parameters of the tunnel, providing ideas and methods for the creep prediction and analysis of surrounding rock in complex environments, and providing references for the support design and construction of large - deformation surrounding rock.

[0136] Based on the same inventive concept, an embodiment of the present invention provides a device for predicting the creep parameters of phyllite tunnel surrounding rock, referring to Figure 7 , Figure 7 is a schematic diagram of the device for predicting the creep parameters of phyllite tunnel surrounding rock provided by an embodiment of the present invention. As Figure 7 shown, the device includes:

[0137] A monitoring module 701 for monitoring the deformation of the current surrounding rock monitoring point;

[0138] A prediction module 702 for inputting the actually monitored deformation into the neural network inversion model corresponding to the current surrounding rock monitoring point to obtain the creep parameters of the current surrounding rock monitoring point;

[0139] Among them, a set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and the deformation obtained by simulating and calculating the creep parameters of the current surrounding rock monitoring point.

[0140] Optionally, it further includes:

[0141] A building module for building a neural network inversion model using the matlab neural network toolbox. The input nodes of the neural network inversion model are 5, the number of hidden - layer nodes is 13, and the output nodes are 6.

[0142] Optionally, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point include:

[0143] Obtain multiple groups of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point;

[0144] Use the multiple groups of training data to train the neural network inversion model to obtain the neural network inversion model corresponding to the current surrounding rock monitoring point.

[0145] Optionally, obtaining multiple groups of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes:

[0146] Fit the indoor experimental results corresponding to the surrounding rock at the current surrounding rock monitoring point to obtain the corresponding initial creep parameters;

[0147] Keep the original parameters of other monitoring points unchanged, and generate a creep parameter sample set for the current surrounding rock monitoring point based on the initial creep parameters and orthogonal experimental design;

[0148] Input each group of creep parameters in the creep parameter sample set into the numerical model to calculate the corresponding deformation amount, and use each group of creep parameters and the corresponding deformation amount as training data.

[0149] Optionally, inputting each group of creep parameters in the creep parameter sample set into the numerical model to calculate the corresponding deformation amount includes:

[0150] Embed the creep damage constitutive equation of phyllite surrounding rock into the GDEM numerical simulation program;

[0151] Input each group of creep parameters in the creep parameter sample set into the GDEM numerical simulation program to calculate the corresponding deformation amount.

[0152] For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiment.

[0153] Based on the same inventive concept, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps in the phyllite tunnel surrounding rock creep parameter prediction method described in any of the above embodiments are implemented.

[0154] Based on the same inventive concept, an embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes, the steps in the phyllite tunnel surrounding rock creep parameter prediction method described in any of the above embodiments are implemented.

[0155] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other.

[0156] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, an apparatus, or a computer program product. Therefore, the embodiments of the present invention can take the form of an all-hardware embodiment, an all-software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate a device for realizing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0158] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured product including an instruction device, and the instruction device realizes the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0159] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable terminal device provide steps for realizing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks.

[0160] Although the preferred embodiments of the embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the embodiments of the present invention.

[0161] Finally, it should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or terminal device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or terminal device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising the element.

[0162] The above has introduced in detail a method, device and product for predicting the creep parameters of the surrounding rock of a phyllite tunnel provided by the present invention. Specific examples are used in this text to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for predicting creep parameters of surrounding rock in a phyllite tunnel, characterized in that, the method includes: Monitoring the deformation of the current surrounding rock monitoring points, and the deformation includes: side wall convergence deformation, right arch waist settlement, crown settlement, left arch waist settlement, and arch waist convergence deformation; Input the actually monitored deformation into the neural network inversion model corresponding to the current surrounding rock monitoring point to obtain the creep parameters of the current surrounding rock monitoring point; Among them, a set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and the corresponding deformation, and the corresponding deformation is obtained by simulating and calculating the creep parameters of the current surrounding rock monitoring point and the initial creep parameters of other surrounding rock monitoring points; the creep parameters of the current surrounding rock monitoring point are obtained through an orthogonal experimental scheme based on the initial creep parameters of the current surrounding rock monitoring point; The initial creep parameters of the surrounding rock monitoring points are obtained from the results of the triaxial creep experiment of saturated phyllite under seepage-stress conditions and the creep constitutive equation of the surrounding rock of the phyllite tunnel under seepage-stress conditions: ， Among them, is the effective stress, ; is the effective deviator stress, ; represents the effective damage amount, which is determined by the triaxial creep experiment results of saturated phyllite under seepage-stress conditions. If there is no seepage condition, then ; represents the minimum principal stress of in-situ stress of tunnel surrounding rock, represents the maximum principal stress of in-situ stress of tunnel surrounding rock, P represents the osmotic pressure, represents the long-term strength of the rock; A, B, C, F, M, N are creep parameters; is the crack initiation damage function, which is a linear evolution law. Let , and k is the damage parameter affected by the ratio of deviator stress to short-term peak strength; is the accelerated damage function, which is a power function evolution law. Let , and m, n are accelerated damage parameters; the damage evolution law curve is obtained from the indoor acoustic emission experiment, and the damage parameters are obtained by fitting.

2. The method according to claim 1, characterized in that, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point include: Obtaining multiple sets of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point; Using the multiple sets of training data to train the BP model to obtain the neural network inversion model corresponding to the current surrounding rock monitoring point.

3. The method according to claim 2, characterized in that, Obtaining multiple sets of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: Conducting a triaxial creep experiment of saturated phyllite under seepage-stress conditions on the surrounding rock at each surrounding rock monitoring point, fitting the experimental results with the creep constitutive equation of the surrounding rock of the phyllite tunnel under seepage-stress conditions to obtain the initial creep parameters of each detection point; Fixing the initial creep parameters of other monitoring points unchanged, and generating a creep parameter sample set of the current surrounding rock monitoring point based on the initial creep parameters of the current monitoring point and the orthogonal experimental design; Inputting each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the numerical model to calculate the corresponding deformation, and taking each set of creep parameters and the corresponding deformation as training data, and the numerical model is the GDEM numerical model.

4. The method according to claim 3, characterized in that, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point further include: Inputting the test deformation into the neural network inversion model corresponding to the current surrounding rock monitoring point obtained by training, inversely obtaining the corresponding test creep parameters, and inputting the test creep parameters into the numerical model to obtain the simulated deformation; Comparing the error between the test deformation and the simulated deformation; In the case that the error is greater than the preset threshold, adjusting the numerical model, re-determining the training data using the adjusted numerical model, and re-training the BP model.

5. The method according to claim 3, characterized in that, Input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into a numerical model to calculate the corresponding deformation amounts, including: Embed the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions into the GDEM numerical simulation program; Input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into the GDEM numerical simulation program respectively to calculate the corresponding deformation amounts.

6. A device for predicting creep parameters of phyllite tunnel surrounding rock, characterized in that, the device includes: A monitoring module for monitoring the deformation amounts of the current surrounding rock monitoring points, and the deformation amounts include: side wall convergence deformation amount, right arch waist settlement amount, crown settlement amount, left arch waist settlement amount, arch waist convergence deformation amount; A prediction module for inputting the actually monitored deformation amounts into the neural network inversion model corresponding to the current surrounding rock monitoring point to obtain the creep parameters of the current surrounding rock monitoring point; A set of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: the creep parameters of the current surrounding rock monitoring point and their corresponding deformation amounts, and the corresponding deformation amounts are obtained by simulating and calculating the creep parameters of the current surrounding rock monitoring point and the initial creep parameters of other surrounding rock monitoring points; the creep parameters of the current surrounding rock monitoring point are obtained through an orthogonal experimental scheme based on the initial creep parameters of the current surrounding rock monitoring point; The initial creep parameters of the surrounding rock monitoring points are obtained through the results of saturated phyllite triaxial creep experiments under seepage-stress conditions and the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions: ， Among them, is the effective stress, ; is the effective deviator stress, ; represents the effective damage amount, which is determined by the triaxial creep experiment results of saturated phyllite under seepage-stress conditions. If there is no seepage condition, then ; represents the minimum principal stress of the in-situ stress of the tunnel surrounding rock, represents the maximum principal stress of the in-situ stress of the tunnel surrounding rock, P represents the osmotic pressure, represents the long-term strength of the rock; A, B, C, F, M, N are creep parameters; is the crack initiation damage function, which is a linear evolution law. Let , and k is the damage parameter affected by the ratio of deviator stress to short-term peak strength; is the accelerating damage function, which is a power function evolution law. Let , and m, n are accelerating damage parameters; the damage evolution law curve is obtained from the indoor acoustic emission experiment, and the damage parameters are obtained by fitting .

7. The device according to claim 6, characterized in that, the training steps of the neural network inversion model corresponding to the current surrounding rock monitoring point include: Obtain multiple sets of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point; Use the multiple sets of training data to train the BP model to obtain the neural network inversion model corresponding to the current surrounding rock monitoring point.

8. The device according to claim 6, characterized in that, Obtaining multiple sets of training data of the neural network inversion model corresponding to the current surrounding rock monitoring point includes: Conduct saturated phyllite triaxial creep experiments on the surrounding rock at each surrounding rock monitoring point, fit the experimental results with the creep constitutive equation of phyllite tunnel surrounding rock under seepage-stress conditions to obtain the initial creep parameters of each detection point; Keep the initial creep parameters of other monitoring points unchanged, and generate a creep parameter sample set of the current surrounding rock monitoring point based on the initial creep parameters of the current monitoring point and the orthogonal experimental design; Input each set of creep parameters in the creep parameter sample set and the initial creep parameters of other monitoring points into a numerical model to calculate the corresponding deformation amounts, and use each set of creep parameters and the corresponding deformation amounts as training data, and the numerical model is the GDEM numerical model.

9. An electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the prediction method for the creep parameters of the phyllite tunnel surrounding rock according to any one of claims 1-5.

10. A computer-readable storage medium, on which a computer program is stored, characterized in that when the program is executed by a processor, it implements the steps of the prediction method for the creep parameters of the phyllite tunnel surrounding rock according to any one of claims 1-5.

Images