Granite joint multi-working condition stress drop and micro crack correlation representation method and device, computer equipment and readable storage medium

Abstract

The application provides a method and device for correlatively characterizing stress drop and micro cracks of granite joints under multiple working conditions, computer equipment and a readable storage medium, relates to the technical field of deep hard rock engineering mechanics and geological disaster warning, and the method comprises the following steps: constructing a Voronoi-GBM numerical model; constructing a multi-objective optimization function for optimizing multiple micro mechanical parameters; optimizing the micro mechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function; simulating the operation of multiple granite shear working conditions by using the optimized Voronoi-GBM numerical model; and taking the micro crack characteristic parameters as input samples and the macro stress drop characteristic parameters as output samples, and training a target correlation characterization model. The scheme can effectively mine the correlation characterization of stress drop and micro cracks of granite joints, and realize the prediction of macro stress drop behavior from the micro crack characteristics.

Description

Technical Field

[0001] This application relates to the fields of deep hard rock engineering mechanics and geological disaster early warning technology, and in particular to a method and apparatus, computer equipment and readable storage medium for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions. Background Technology

[0002] Granite, a common surrounding rock in deep mineral resource extraction, water conservancy and hydropower projects, and underground space construction, has a controlling impact on engineering stability due to its internal weak structures such as joints and fissures. The shear mechanics behavior of joint surfaces, especially the stress drop phenomenon during shearing, is a key precursor and core indicator of dynamic disasters such as the sudden release of accumulated energy in the rock mass, inducing rockbursts and slippage. Therefore, accurately characterizing the stress drop mechanism and establishing its correlation with internal rock mass damage is crucial for the safety of deep engineering projects.

[0003] Granite exhibits significant heterogeneity at the microscale, composed of interlocking grains of various minerals such as quartz, feldspar, and mica, with numerous microcracks naturally present within the grains and at grain boundaries. Under shear loading, these microscopic defects preferentially initiate, propagate, and connect, ultimately leading to frictional slippage and stress reduction at macroscopic joint surfaces. Different microscopic fracture modes, such as intragranular tensile fracturing and grain boundary shear slip, contribute differently to the macroscopic mechanical response.

[0004] However, existing research methods have significant limitations. Indoor experiments cannot observe the dynamic evolution of cracks within shear zones in real time and without damage. Traditional continuous medium numerical models cannot characterize discontinuous fracture processes, while conventional discrete element models often simplify rock masses into homogeneous aggregates of particles or consider only a single type of contact model. These methods fail to accurately reflect the dual strength characteristics of granite—strong intragranular structures and weak grain boundaries—as well as the mechanical differences in mineral composition. This results in insufficient accuracy in simulating the magnitude, frequency, and microscopic causes of stress drop. In other words, existing numerical simulation methods for granite joint shear behavior cannot accurately construct heterogeneous microscopic models that reflect the differences in granite mineral composition and the weak grain boundary effect. Consequently, they cannot effectively distinguish, monitor, and quantify different types of microscopic cracks, such as intragranular and grain boundary cracks, during shearing. They also fail to reveal the quantitative correlation between macroscopic stress drop and microscopic crack evolution under various working conditions, leading to inaccurate characterization of stress drop phenomena during joint shearing and making it difficult to predict macroscopic instability behavior from the perspective of microscopic damage.

[0005] Therefore, how to construct a characterization method that can reflect the differences in mineral grains and the true joint morphology inside granite, and establish an accurate quantitative correlation between macroscopic stress drop and microscopic cracks, has become an urgent technical problem to be solved. Summary of the Invention

[0006] This application provides a method, apparatus, computer equipment, and readable storage medium for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, aiming to solve the technical problem in related technologies that it is impossible to simultaneously achieve a true characterization of the heterogeneous structure of granite and a cross-scale quantitative correlation between macroscopic and microscopic behaviors.

[0007] In a first aspect, embodiments of this application provide a method for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, including: Acquire multi-condition basic data, wherein the multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under multiple conditions; Based on the multi-condition basic data, a Voronoi-GBM numerical model is constructed, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite. A multi-objective optimization function is constructed to optimize various micromechanical parameters, and the micromechanical parameters of the Voronoi-GBM numerical model are optimized based on the multi-objective optimization function; Based on the optimized Voronoi-GBM numerical model, various granite shearing conditions were simulated and run, and the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite were monitored during the simulation. Using the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples, a target association characterization model is trained, wherein the target association characterization model is used to reflect the correlation between the microscopic crack characteristic parameters and the macroscopic stress drop characteristic parameters in the shearing condition of granite.

[0008] Optionally, in one embodiment of this application, acquiring multi-condition basic data includes: The mineral type and mineral content of the granite were determined by X-ray diffraction testing; The basic mechanical parameters of the granite were obtained through uniaxial compression tests and Brazilian splitting tests. The peak shear strength, residual strength, and stress-displacement curves of the joints in the granite were obtained by direct shear test. The topographic point cloud data of the joint surface of the granite was acquired using a 3D laser scanner, and the roughness coefficient of the joint surface was determined. Micro-CT technology was used to scan the granite before and after shearing to obtain images of the fine crack distribution of the granite.

[0009] Optionally, in one embodiment of this application, constructing the Voronoi-GBM numerical model based on the multi-condition basic data includes: A set of crystal nuclei points is generated for each of the mineral types, wherein the ratio of the number of crystal nuclei points corresponding to each of the mineral types is the same as the ratio of the mineral content of the mineral types in the granite; Based on the set of nucleus points for each mineral type, a polygonal grain structure is constructed using the Voronoi diagram partitioning algorithm, wherein each grain in the polygonal grain structure represents a mineral unit; Each grain in the polygonal grain structure is filled with fine particle units, and a set of microscopic parameters corresponding to the mineral type to which the grain belongs is assigned to the grain. A parallel bonding model is introduced between the fine particle units inside the grain, wherein the parallel bonding model is used to simulate intragranular materials; A smooth joint model is introduced between the fine particle units at the adjacent boundaries of every two grains, wherein the smooth joint model is configured with a specified bond strength and a specified friction coefficient to obtain an initial Voronoi-GBM numerical model. The point cloud data of the joint surface morphology of the granite is mapped to the irregular contact surfaces of the hanging wall and footwall of the initial Voronoi-GBM numerical model to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite.

[0010] Optionally, in one embodiment of this application, the multi-objective optimization function is: , Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients, where, , , , , , This represents the peak shear strength obtained from the numerical simulation. σ represents the reference peak strength determined by indoor cyclic shear tests. n The normal stress is represented by JRC, the joint roughness coefficient, and N is the number of shear cycles. This represents the residual shear strength obtained from the numerical simulation. This represents the reference residual strength determined by the indoor cyclic load direct shear test.

[0011] In one embodiment of this application, optionally, the optimization objectives of the multi-objective optimization function include the peak intensity error, stress drop amplitude error, residual strength error of the macroscopic stress-displacement curve, and the similarity between the microscopic crack distribution pattern and the CT image. The optimization of the mesomechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function includes: By iteratively running numerical direct shear simulations, the micromechanical parameters of the Voronoi-GBM numerical model are dynamically adjusted until the error between the macromechanical response of the numerical direct shear output by the multi-objective optimization function and the indoor direct shear test results is less than a predetermined threshold, and the microcrack distribution pattern generated by the simulation matches the CT scan observation results. At this point, the adjustment results of the micromechanical parameters are output.

[0012] In one embodiment of this application, optionally, the macroscopic stress drop characteristic parameters include one or more of the following: single stress drop amplitude, shear displacement when stress drop occurs, stress drop frequency per unit shear displacement, and cumulative stress drop energy release rate. The microcrack characteristic parameters are obtained based on microfracture events recorded in real time during the working condition simulation. The microcrack characteristic parameters include the number, increment, distribution density, and proportion of various types of microcracks in the total number of cracks.

[0013] In one embodiment of this application, optionally, training the target association representation model using the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples includes: Construct a sample dataset with the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples; The sample dataset is randomly divided into a training set and a test set according to a preset ratio; A basic association representation model is configured using regression analysis algorithms or machine learning algorithms, wherein the regression analysis algorithm includes multiple linear regression, and the machine learning algorithm includes support vector machine, random forest, or neural network; Using the microscopic crack feature parameters in the training set as input and the macroscopic stress drop feature parameters in the training set as labels, the basic correlation representation model is iteratively trained. The basic correlation representation model is iteratively updated by minimizing the loss function between the predicted value and the true value. The accuracy of the trained association representation model is verified using the test set, and the prediction error is calculated. If the prediction error exceeds a predetermined effective range, the association representation model is optimized and trained until a target association representation model is obtained where the prediction error does not exceed the predetermined effective range.

[0014] Secondly, embodiments of this application provide a characterization device for the correlation between stress drop and microcracks in granite joints under multiple working conditions, comprising: A multi-condition basic data acquisition unit is used to acquire multi-condition basic data, wherein the multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under multiple conditions. The Voronoi-GBM model construction unit is used to construct a Voronoi-GBM numerical model based on the multi-condition basic data, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite. A multi-objective optimization unit is used to construct a multi-objective optimization function for optimizing various micromechanical parameters, and to optimize the micromechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function; The working condition simulation and monitoring unit is used to simulate various granite shear working conditions based on the optimized Voronoi-GBM numerical model, and to monitor the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite during the working condition simulation process. The correlation characterization model construction unit is used to train the target correlation characterization model with the microcrack characteristic parameters as input samples and the macrostress drop characteristic parameters as output samples. The target correlation characterization model is used to reflect the correlation between the microcrack characteristic parameters and the macrostress drop characteristic parameters in the granite shearing condition.

[0015] In one embodiment of this application, optionally, the multi-condition basic data acquisition unit is used to: determine the mineral type and mineral content of the granite through X-ray diffraction testing; obtain the basic mechanical parameters of the granite through uniaxial compression testing and Brazilian splitting testing; obtain the peak shear strength, residual strength, and stress-displacement curve of the joints of the granite through direct shear testing; acquire the morphological point cloud data of the joint surface of the granite using a three-dimensional laser scanner, and determine the roughness coefficient of the joint surface; and scan the granite before and after shearing using micro-CT technology to obtain a microscopic crack distribution image of the granite.

[0016] In one embodiment of this application, optionally, the Voronoi-GBM model construction unit is used to: generate a corresponding set of nucleus points for each mineral type, wherein the ratio of the number of nucleus points corresponding to each of the multiple mineral types is the same as the ratio of the mineral content of the multiple mineral types in the granite; construct a polygonal grain structure based on the set of nucleus points for each mineral type using a Voronoi graph partitioning algorithm, wherein each grain in the polygonal grain structure represents a mineral unit; fill each grain in the polygonal grain structure with fine-grained units, and assign each grain a unit corresponding to the mineral type according to the mineral type to which the grain belongs. The process involves: establishing a set of microscopic parameters; introducing a parallel bonding model between the fine particle units within the grains, wherein the parallel bonding model is used to simulate intragranular materials; introducing a smooth joint model between the fine particle units at the adjacent boundaries of every two grains, wherein the smooth joint model is configured with a specified bonding strength and a specified friction coefficient, to obtain an initial Voronoi-GBM numerical model; mapping the topographic point cloud data of the joint surface of the granite to the irregular contact surfaces of the upper and lower plates of the initial Voronoi-GBM numerical model, to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite.

[0017] Optionally, in one embodiment of this application, the multi-objective optimization function is: , Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients, where, , , , , , This represents the peak shear strength obtained from the numerical simulation. σ represents the reference peak strength determined by indoor cyclic shear tests. n The normal stress is represented by JRC, the joint roughness coefficient, and N is the number of shear cycles. This represents the residual shear strength obtained from the numerical simulation. This represents the reference residual strength determined by the indoor cyclic load direct shear test.

[0018] In one embodiment of this application, optionally, the optimization objectives of the multi-objective optimization function include the peak strength error, stress drop amplitude error, residual strength error, and similarity between the microcrack distribution pattern and the CT image of the macroscopic stress-displacement curve; the multi-objective optimization unit is used to: dynamically adjust the micromechanical parameters of the Voronoi-GBM numerical model by iteratively running numerical direct shear simulation until the error between the macroscopic mechanical response of the numerical direct shear output by the multi-objective optimization function and the indoor direct shear test results is less than a predetermined threshold, and the simulated microcrack distribution pattern matches the CT scan observation results, and then output the adjustment results of the micromechanical parameters.

[0019] In one embodiment of this application, optionally, the macroscopic stress drop characteristic parameters include one or more of the following: single stress drop amplitude, shear displacement at the time of stress drop, stress drop frequency per unit shear displacement, and cumulative stress drop energy release rate; the microscopic crack characteristic parameters are obtained based on microscopic fracture events recorded in real time during the working condition simulation, and the microscopic crack characteristic parameters include the number, increment, distribution density, and proportion of various types of microscopic cracks in the total number of cracks.

[0020] In one embodiment of this application, optionally, the correlation representation model construction unit is configured to: construct a sample dataset with the microcrack feature parameters as input samples and the macrostress drop feature parameters as output samples; randomly divide the sample dataset into a training set and a test set according to a preset ratio; configure a basic correlation representation model using a regression analysis algorithm or a machine learning algorithm, wherein the regression analysis algorithm includes multiple linear regression, and the machine learning algorithm includes support vector machine, random forest, or neural network; iteratively train the basic correlation representation model using the microcrack feature parameters in the training set as input and the macrostress drop feature parameters in the training set as labels, and iteratively update the basic correlation representation model by minimizing the loss function between the predicted value and the true value; verify the accuracy of the trained correlation representation model using the test set, and calculate the prediction error, wherein if the prediction error exceeds a predetermined effective range, the correlation representation model is optimized and trained until a target correlation representation model is obtained in which the prediction error does not exceed the predetermined effective range.

[0021] Thirdly, embodiments of this application provide a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to perform the method described in the first aspect above.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for performing the method described in the first aspect above.

[0023] The above technical solutions address the technical problem that it is impossible to simultaneously achieve a true representation of the heterogeneous structure of granite and a cross-scale quantitative correlation between macroscopic and microscopic behaviors in related technologies. They can effectively explore the correlation between stress drop in granite joints and microscopic cracks, and realize the prediction of macroscopic stress drop behavior from microscopic crack characteristics. This provides an effective quantitative tool at the microscopic level for instability early warning in deep hard rock engineering, thereby improving construction safety. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating a method for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, according to an embodiment of this application, is shown. Figure 2 A schematic diagram is shown of the mineral composition and heterogeneous microstructure of a real rock core reconstructed by discrete particle units from a Voronoi-GBM-based granite PFC2D numerical model according to an embodiment of this application. Figure 3 A schematic diagram of a heterogeneous rock structure containing natural joints, reproduced by a PFC2D direct shear test numerical model based on real granite mineral composition according to an embodiment of this application, is shown. Figure 4 A schematic diagram of crack evolution in a direct shear numerical simulation of granite PFC according to an embodiment of this application is shown; Figure 5 A block diagram of a computer device according to one embodiment of this application is shown; Figure 6 A block diagram of a computer device according to another embodiment of this application is shown. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Figure 1 A flowchart illustrating a method for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, according to an embodiment of this application, is shown. like Figure 1 As shown, a method for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions according to an embodiment of this application includes: Step 102: Obtain basic data for multiple operating conditions.

[0028] The multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under various conditions, and is the basis for characterizing the correlation between stress drop and micro-cracks in granite joints.

[0029] In one possible design, step 102 includes: determining the mineral type and mineral content of the granite through X-ray diffraction testing; obtaining the basic mechanical parameters of the granite through uniaxial compression testing and Brazilian splitting testing; obtaining the peak shear strength, residual strength, and stress-displacement curves of the joints of the granite through direct shear testing; acquiring the morphological point cloud data of the joint surface of the granite using a three-dimensional laser scanner and determining the roughness coefficient of the joint surface; and scanning the granite before and after shearing using micro-CT (Computed Tomography) technology to obtain microscopic crack distribution images of the granite.

[0030] Mineral type and mineral content reflect the variety and volume percentage of different mineral grains in granite, providing a compositional basis for heterogeneous models. Basic mechanical parameters reflect the strength, or deformation characteristics, of granite at the macroscopic scale, and can be used to calibrate the macroscopic mechanical response of numerical models. Peak shear strength, residual strength, and stress-displacement curves reflect the strength decay law and stress drop characteristics of joint surfaces during shearing, serving as a benchmark for stress drop simulation. Topographic point cloud data reflects the three-dimensional geometry of the joint surface, providing a basis for reconstructing the shape of the actual contact surface. Roughness coefficient reflects the degree of undulation of the joint surface, used to quantify the influence of morphology on shear mechanical behavior. Microscopic crack distribution images reflect the location and density of internal cracks in the rock before and after shearing, serving as a basis for verifying the accuracy of the numerical model's simulation of microscopic fracture modes.

[0031] Therefore, multi-dimensional data reflecting the correlation between stress drop and microcracks in granite joints can be extracted, which can serve as the basis for accurately quantifying this correlation in subsequent steps.

[0032] Step 104: Based on the multi-condition basic data, construct the Voronoi-GBM numerical model.

[0033] The Voronoi-GBM (Voronoi-Gradient Boosting Machine) numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite.

[0034] The Voronoi-GBM numerical model is a discrete element model based on Voronoi diagrams to generate polygonal grain structures. The Voronoi diagram is a method of dividing a plane into multiple regions, each with a seed point. Any location within a region is closer to the seed point than to other seed points. Based on this, the model can be used to characterize the spatial distribution of different mineral grains in granite. It employs a parallel bonding model within the grains to simulate high-strength intragranular materials and incorporates a smooth joint model at grain boundaries to simulate natural weak interfaces, thereby achieving a microscopic reconstruction of the heterogeneous characteristics of granite.

[0035] In one possible design, step 104 includes: generating a corresponding set of nucleus points for each of the mineral types, wherein the proportion of the number of nucleus points corresponding to each of the various mineral types is the same as the proportion of the mineral content of the various mineral types in the granite. The set of nucleus points corresponding to each mineral type reflects the spatial distribution and density relationship of various mineral grains in the granite, wherein the proportion of the number of nucleus points is the same as the proportion of mineral content, thereby ensuring that the area ratio of different mineral grains in the Voronoi-GBM numerical model is consistent with the composition of real rocks, providing a reliable basis for the subsequent construction of polygonal grain structures with real mineral spatial distribution characteristics.

[0036] Step 104 further includes: constructing a polygonal grain structure based on the set of nucleus points for each mineral type using a Voronoi diagram subdivision algorithm. Specifically, the polygonal grain structure reflects the spatial geometry, size distribution, and inter-grain interactions of different mineral grains in granite. Each grain in the polygonal grain structure represents a mineral unit, thereby transforming the irregular mosaic arrangement of minerals such as quartz, feldspar, and mica in real rocks into a computable discrete geometric expression in the numerical model, providing a realistic topological path for simulating crack generation along grain boundaries and its transgranular propagation.

[0037] Step 104 further includes: filling each grain in the polygonal grain structure with fine-grained units, and assigning a set of microscopic parameters corresponding to the mineral type to each grain; introducing a parallel bonding model between the fine-grained units inside the grain, wherein the parallel bonding model is used to simulate intragranular materials. The filled fine-grained units are the basic discrete element carriers constituting the Voronoi-GBM numerical model, which discretizes the continuous polygonal grains into a large number of independently moving particle assemblies, thereby providing microscale computational units for crack initiation, propagation, and contact fracture between particles. At the same time, assigning corresponding microscopic parameter sets to the grains according to the mineral type reflects the essential differences in mechanical properties such as elastic modulus, bond strength, and internal friction angle of different minerals such as quartz, feldspar, and mica, enabling the Voronoi-GBM numerical model to learn and distinguish the different contribution levels of various minerals to the macroscopic mechanical response. Furthermore, a parallel bonding model is introduced between the fine particle units inside the grain. This parallel bonding model reflects the ability of the material inside the whole grain to withstand tensile and shear loads simultaneously. The parallel bonding model simulates the generation and propagation process of tensile cracks and shear cracks inside the grain through the fracture mechanism of the bonding bond, thus providing a micromechanical description of the intragranular fracture behavior that conforms to physical reality.

[0038] Furthermore, step 104 also includes: introducing a smooth joint model between the fine particle units at the adjacent boundaries of every two grains to obtain an initial Voronoi-GBM numerical model; mapping the topographic point cloud data of the joint surface of the granite to the irregular contact surfaces of the hanging wall and the footwall of the initial Voronoi-GBM numerical model to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite.

[0039] The smooth joint model is configured with a specified bond strength and a specified coefficient of friction.

[0040] A smooth joint model is introduced between the fine grain units at the adjacent boundaries of every two grains. This smooth joint model is configured with a specified bond strength and a specified friction coefficient that are lower than the intragranular strength. It reflects the essential characteristic of natural grain boundaries as mechanically weak surfaces, so that cracks preferentially generate and propagate along the grain boundaries during shearing, thus realistically reproducing the dual strength characteristics of granite, which is strong within the grains and weak at the grain boundaries.

[0041] Simultaneously, mapping the morphological point cloud data of the granite joint surface to the irregular contact surfaces of the hanging wall and footwall in the initial Voronoi-GBM numerical model reflects the three-dimensional undulation of the actual joint surface, ensuring that the mechanical behavior of the contact surfaces of the hanging wall and footwall is consistent with the actual joint. Thus, a weak interface system conforming to the mechanical differences of mineral boundaries is constructed within the model, and the contact geometry of the actual joint is restored at the model boundaries. This results in a Voronoi-GBM numerical model that can simultaneously reflect the heterogeneous grain structure of granite and the morphology of the actual joint, providing a reliable microstructural basis for accurate prediction of crack paths and realistic reproduction of joint surface friction behavior in subsequent shear simulations.

[0042] Step 106: Construct a multi-objective optimization function for optimizing various micromechanical parameters, and optimize the micromechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function.

[0043] The multi-objective optimization function is: , A multi-objective optimization function is a single comprehensive objective function composed of a weighted combination of individual error indices. For the micromechanical parameters to be optimized, Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients reflect the respective influence levels of each of the three factors on the overall error. + + =1, where, , , , , , This represents the peak shear strength obtained from the numerical simulation. σn represents the reference peak strength determined by the indoor cyclic load direct shear test, JRC is the joint roughness coefficient, and N is the number of shear cycles. This represents the residual shear strength obtained from the numerical simulation. This represents the reference residual strength determined by indoor cyclic shear tests. In other words, The peak shear strength obtained from numerical simulation Reference peak strength determined by indoor cyclic loading direct shear test The relative error between them Residual shear strength obtained from numerical simulation Reference residual strength determined by indoor cyclic shear test The relative error between them The stress drop obtained from numerical simulation The reference stress drop calculated based on the reference values ​​from the aforementioned indoor cyclic load direct shear test. The relative error between them.

[0044] It can be said that the optimization objectives of the multi-objective optimization function include the peak strength error, stress drop amplitude error, residual strength error of the macroscopic stress-displacement curve, and the similarity between the microscopic crack distribution pattern and the CT image. The optimization method includes: dynamically adjusting the microscopic mechanical parameters of the Voronoi-GBM numerical model by iteratively running numerical direct shear simulation until the error between the macroscopic mechanical response of the numerical direct shear output by the multi-objective optimization function and the indoor direct shear test results is less than a predetermined threshold, and the microscopic crack distribution pattern generated by the simulation matches the CT scan observation results, at which point the adjustment result of the microscopic mechanical parameters is output.

[0045] Peak strength error reflects the relative deviation between the peak shear strength obtained from numerical simulation and the reference peak strength determined by indoor cyclic loading direct shear tests. This reference peak strength is calculated using empirical formulas based on normal stress, joint roughness coefficient, and the number of shear cycles, thus quantifying the model's simulation accuracy in terms of ultimate bearing capacity. Residual strength error reflects the relative deviation between the residual shear strength obtained from numerical simulation and the reference residual strength from indoor tests, thus quantifying the model's simulation accuracy in the strength decay behavior after shear failure. Stress drop amplitude error reflects the relative deviation between the stress drop obtained from numerical simulation and the reference stress drop calculated based on experimental reference values, thus quantifying the model's simulation accuracy for sudden energy release. Simultaneously, the multi-objective optimization function integrates the above three individual errors into a single comprehensive objective function through weighted combination, with weight coefficients reflecting the different levels of influence of each error on the overall error. Thus, by iteratively running numerical direct shear simulations, the micromechanical parameters of the intragranular parallel bonding model and the grain boundary smooth joint model are dynamically adjusted until the macroscopic mechanical response error is less than a predetermined threshold and the microscopic failure mode matches the observation. This yields a set of calibration parameters that can simultaneously reflect macroscopic mechanical behavior and microscopic fracture mode, solving the problem of inaccurate parameters caused by traditional calibration methods that rely solely on macroscopic curves while ignoring microscopic verification.

[0046] Step 108: Based on the optimized Voronoi-GBM numerical model, simulate various granite shear conditions and monitor the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite during the simulation.

[0047] The macroscopic stress drop characteristic parameters include one or more of the following: single stress drop amplitude, shear displacement at the time of stress drop, stress drop frequency per unit shear displacement, and cumulative stress drop energy release rate. The microscopic crack characteristic parameters are obtained based on microscopic fracture events recorded in real time during the working condition simulation. The microscopic crack characteristic parameters include the number, increment, distribution density, and proportion of various types of microscopic cracks in the total number of cracks.

[0048] In short, the magnitude of a single stress drop reflects the intensity of a single macroscopic energy release, the shear displacement at the time of stress drop reveals the location of the stress drop, the frequency of stress drop per unit shear displacement characterizes the density of energy release, and the cumulative stress drop energy release rate can quantify the total energy dissipation of the entire shear process. These parameters together constitute a complete description of macroscopic stress drop behavior.

[0049] Simultaneously, based on real-time recorded micro-fracture events, the number, increment, distribution density, and proportion of various micro-cracks in the total crack quantity can be obtained. These parameters reflect the generation and propagation patterns of intragranular and grain boundary cracks, as well as tensile and shear cracks, at different shear stages. Therefore, by simultaneously collecting macroscopic stress drop characteristic parameters and micro-crack characteristic parameters on the basis of the optimized Voronoi-GBM numerical model, a multi-condition, multi-dimensional data foundation can be provided for establishing a quantitative correlation model between the two.

[0050] Step 110: Using the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples, train the target association representation model.

[0051] The target correlation characterization model is used to reflect the correlation between microscopic crack characteristic parameters and macroscopic stress drop characteristic parameters in granite shearing conditions.

[0052] Specifically, during the training process, a sample dataset can be constructed with the microscopic crack feature parameters as input samples and the macroscopic stress drop feature parameters as output samples. The sample dataset is randomly divided into a training set and a test set according to a preset ratio. A basic correlation representation model is configured using a regression analysis algorithm or a machine learning algorithm, wherein the regression analysis algorithm includes multiple linear regression, and the machine learning algorithm includes support vector machine, random forest, or neural network. The basic correlation representation model is iteratively trained using the microscopic crack feature parameters in the training set as input and the macroscopic stress drop feature parameters in the training set as labels. The basic correlation representation model is iteratively updated by minimizing the loss function between the predicted value and the true value. The accuracy of the trained correlation representation model is verified using the test set, and the prediction error is calculated. If the prediction error exceeds a predetermined effective range, the correlation representation model is optimized and trained until a target correlation representation model is obtained where the prediction error does not exceed the predetermined effective range.

[0053] Microscopic crack characteristic parameters, used as input samples, reflect the changing patterns of the number, density, and proportion of various types of cracks, including intragranular and grain boundary cracks, tensile and shear cracks, during shearing, providing a microscopic-level characteristic description for prediction. Macroscopic stress drop characteristic parameters, used as output samples, reflect key indicators such as single stress drop amplitude, frequency, and energy release rate, providing supervisory labels for model training.

[0054] Therefore, by constructing a sample dataset with microcrack features as input and macrostress drop features as output, and using algorithms such as multiple linear regression, support vector machine, random forest, and neural network for iterative training, while verifying the accuracy with a test set, a target correlation representation model that can quantify the relationship between microcracks and macrostress drop can be obtained, thus achieving the goal of predicting macrostress drop behavior from microcrack features.

[0055] The above technical solution first obtains basic data under multiple working conditions through indoor experiments. Then, based on this data, a Voronoi-GBM numerical model reflecting the heterogeneous grain structure and true joint morphology is constructed. Next, micromechanical parameters are calibrated using a multi-objective optimization function, and multi-working-condition shear simulations are conducted, simultaneously monitoring macroscopic stress drop and microscopic crack evolution. Finally, a correlation characterization model is trained using microscopic crack characteristics as input and macroscopic stress drop characteristics as output. Based on this solution, the correlation between stress drop and microscopic cracks in granite joints can be effectively explored, enabling the prediction of macroscopic stress drop behavior from microscopic crack characteristics. This provides an effective quantitative tool at the microscopic level for instability early warning in deep hard rock engineering, improving construction safety.

[0056] Figure 2 This illustration shows a schematic diagram of the mineral composition and heterogeneous microstructure of a real rock core, reconstructed using discrete particle units from a Voronoi-GBM-based numerical model of granite in PFC2D (two-dimensional particle flow numerical simulation software) according to an embodiment of this application. Within a 100mm × 50mm unit size, different colors are used to represent the mineral composition within a 25mm width range of the real rock core, including potassium feldspar, plagioclase, quartz, and mica. This provides a numerical characterization for studying the micromechanical mechanisms of rock mechanics experiments.

[0057] Figure 3 A schematic diagram is shown of a heterogeneous rock structure containing natural joints, reproduced by a PFC2D direct shear test numerical model based on real granite mineral composition, according to an embodiment of this application. The heterogeneous rock structure containing natural joints was reproduced using a 100mm × 100mm standard sample, used for numerical simulation studies of the shear mechanical behavior and microscopic failure mechanism of the rock structure surface, in comparison with real direct shear tests.

[0058] Figure 4 This paper illustrates a schematic diagram of crack evolution in a direct shear numerical simulation of granite PFC (Particle Flow Program) according to an embodiment of this application. The result shows the crack evolution of the direct shear numerical simulation of granite PFC. By superimposing red cracks with mineral components, the paper intuitively reveals the crack initiation and propagation laws during the shearing process and clearly distinguishes the spatial distribution characteristics of intragranular cracks and grain boundary (between different mineral grains) cracks.

[0059] In addition, embodiments of this application provide a characterization device for the correlation between stress drop and microcracks in granite joints under multiple working conditions, comprising: A multi-condition basic data acquisition unit is used to acquire multi-condition basic data, wherein the multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under multiple conditions. The Voronoi-GBM model construction unit is used to construct a Voronoi-GBM numerical model based on the multi-condition basic data, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite. A multi-objective optimization unit is used to construct a multi-objective optimization function for optimizing various micromechanical parameters, and to optimize the micromechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function; The working condition simulation and monitoring unit is used to simulate various granite shear working conditions based on the optimized Voronoi-GBM numerical model, and to monitor the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite during the working condition simulation process. The correlation characterization model construction unit is used to train the target correlation characterization model with the microcrack characteristic parameters as input samples and the macrostress drop characteristic parameters as output samples. The target correlation characterization model is used to reflect the correlation between the microcrack characteristic parameters and the macrostress drop characteristic parameters in the granite shearing condition.

[0060] In one embodiment of this application, optionally, the multi-condition basic data acquisition unit is used to: determine the mineral type and mineral content of the granite through X-ray diffraction testing; obtain the basic mechanical parameters of the granite through uniaxial compression testing and Brazilian splitting testing; obtain the peak shear strength, residual strength, and stress-displacement curve of the joints of the granite through direct shear testing; acquire the morphological point cloud data of the joint surface of the granite using a three-dimensional laser scanner, and determine the roughness coefficient of the joint surface; and scan the granite before and after shearing using micro-CT technology to obtain a microscopic crack distribution image of the granite.

[0061] In one embodiment of this application, optionally, the Voronoi-GBM model construction unit is used to: generate a corresponding set of nucleus points for each mineral type, wherein the ratio of the number of nucleus points corresponding to each of the multiple mineral types is the same as the ratio of the mineral content of the multiple mineral types in the granite; construct a polygonal grain structure based on the set of nucleus points for each mineral type using a Voronoi graph partitioning algorithm, wherein each grain in the polygonal grain structure represents a mineral unit; fill each grain in the polygonal grain structure with fine-grained units, and assign each grain a unit corresponding to the mineral type according to the mineral type to which the grain belongs. The process involves: establishing a set of microscopic parameters; introducing a parallel bonding model between the fine particle units within the grains, wherein the parallel bonding model is used to simulate intragranular materials; introducing a smooth joint model between the fine particle units at the adjacent boundaries of every two grains, wherein the smooth joint model is configured with a specified bonding strength and a specified friction coefficient, to obtain an initial Voronoi-GBM numerical model; mapping the topographic point cloud data of the joint surface of the granite to the irregular contact surfaces of the upper and lower plates of the initial Voronoi-GBM numerical model, to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite.

[0062] Optionally, in one embodiment of this application, the multi-objective optimization function is: , Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients, where, , , , , , This represents the peak shear strength obtained from the numerical simulation. σ represents the reference peak strength determined by indoor cyclic shear tests.n The normal stress is represented by JRC, the joint roughness coefficient, and N is the number of shear cycles. This represents the residual shear strength obtained from the numerical simulation. This represents the reference residual strength determined by the indoor cyclic load direct shear test.

[0063] In one embodiment of this application, optionally, the optimization objectives of the multi-objective optimization function include the peak strength error, stress drop amplitude error, residual strength error, and similarity between the microcrack distribution pattern and the CT image of the macroscopic stress-displacement curve; the multi-objective optimization unit is used to: dynamically adjust the micromechanical parameters of the Voronoi-GBM numerical model by iteratively running numerical direct shear simulation until the error between the macroscopic mechanical response of the numerical direct shear output by the multi-objective optimization function and the indoor direct shear test results is less than a predetermined threshold, and the simulated microcrack distribution pattern matches the CT scan observation results, and then output the adjustment results of the micromechanical parameters.

[0064] In one embodiment of this application, optionally, the macroscopic stress drop characteristic parameters include one or more of the following: single stress drop amplitude, shear displacement at the time of stress drop, stress drop frequency per unit shear displacement, and cumulative stress drop energy release rate; the microscopic crack characteristic parameters are obtained based on microscopic fracture events recorded in real time during the working condition simulation, and the microscopic crack characteristic parameters include the number, increment, distribution density, and proportion of various types of microscopic cracks in the total number of cracks.

[0065] In one embodiment of this application, optionally, the correlation representation model construction unit is configured to: construct a sample dataset with the microcrack feature parameters as input samples and the macrostress drop feature parameters as output samples; randomly divide the sample dataset into a training set and a test set according to a preset ratio; configure a basic correlation representation model using a regression analysis algorithm or a machine learning algorithm, wherein the regression analysis algorithm includes multiple linear regression, and the machine learning algorithm includes support vector machine, random forest, or neural network; iteratively train the basic correlation representation model using the microcrack feature parameters in the training set as input and the macrostress drop feature parameters in the training set as labels, and iteratively update the basic correlation representation model by minimizing the loss function between the predicted value and the true value; verify the accuracy of the trained correlation representation model using the test set, and calculate the prediction error, wherein if the prediction error exceeds a predetermined effective range, the correlation representation model is optimized and trained until a target correlation representation model is obtained in which the prediction error does not exceed the predetermined effective range.

[0066] The device uses the solution described in any one of the above embodiments, and therefore has all the above-mentioned technical effects, which will not be repeated here.

[0067] In another embodiment, this application provides a computer device, which may be a server, and its internal structure diagram may be as follows. Figure 5 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it can implement the methods described in any of the above embodiments.

[0068] In one embodiment, this application also provides a computer device, which can be a client, and its internal structure diagram can be as follows: Figure 6 As shown, the computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it can implement the methods described in any of the above embodiments.

[0069] Any of the computer devices described in the embodiments of this application exist in various forms, including but not limited to: (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include smartphones, multimedia phones, feature phones, and low-end phones.

[0070] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, etc.

[0071] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes: audio and video players, handheld game consoles, e-books, as well as smart toys, wearable devices, and portable car navigation devices.

[0072] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0073] (5) Other electronic devices with data interaction functions.

[0074] In addition, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for performing the following steps: acquiring multi-condition basic data, wherein the multi-condition basic data is used to reflect the multidimensional rock characteristics of granite under multiple conditions; constructing a Voronoi-GBM numerical model based on the multi-condition basic data, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite; constructing a multi-objective optimization function for optimizing various micromechanical parameters, and optimizing the micromechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function; simulating various granite shear conditions based on the optimized Voronoi-GBM numerical model, and monitoring the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite during the simulation; training a target association characterization model using the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples, wherein the target association characterization model is used to reflect the correlation between the microscopic crack characteristic parameters and the macroscopic stress drop characteristic parameters in the granite shear conditions.

[0075] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0076] The technical solution of this application has been described in detail above with reference to the accompanying drawings. Through the technical solution of this application, the correlation between stress drop and microcracks in granite joints can be effectively explored, and macro-stress drop behavior can be predicted from the characteristics of microcracks. This provides an effective quantitative tool at the micro level for instability early warning in deep hard rock engineering and improves construction safety.

[0077] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0078] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0079] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms "a," "the," and "the" as used in the embodiments of this application are also intended to include the plural forms unless the context clearly indicates otherwise.

[0080] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0081] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0082] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0083] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, characterized in that, include: Acquire multi-condition basic data, wherein the multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under multiple conditions; Based on the aforementioned multi-condition basic data, a Voronoi-GBM numerical model is constructed, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite. A multi-objective optimization function is constructed to optimize various micromechanical parameters, and the micromechanical parameters of the Voronoi-GBM numerical model are optimized based on the multi-objective optimization function; Based on the optimized Voronoi-GBM numerical model, various granite shearing conditions were simulated and run, and the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite were monitored during the simulation. Using the microcrack characteristic parameters as input samples and the macrostress drop characteristic parameters as output samples, a target association characterization model is trained, wherein the target association characterization model is used to reflect the correlation between the microcrack characteristic parameters and the macrostress drop characteristic parameters in the shearing condition of granite; The construction of the Voronoi-GBM numerical model based on the multi-condition basic data includes: A set of crystal nuclei points is generated for each mineral type, wherein the ratio of the number of crystal nuclei points corresponding to each of the various mineral types is the same as the ratio of the mineral content of the various mineral types in the granite; Based on the set of nucleus points for each mineral type, a polygonal grain structure is constructed using the Voronoi diagram partitioning algorithm, wherein each grain in the polygonal grain structure represents a mineral unit; Each grain in the polygonal grain structure is filled with fine particle units, and a set of microscopic parameters corresponding to the mineral type to which the grain belongs is assigned to the grain. A parallel bonding model is introduced between the fine particle units inside the grain, wherein the parallel bonding model is used to simulate intragranular materials; A smooth joint model is introduced between the fine particle units at the adjacent boundaries of every two grains, wherein the smooth joint model is configured with a specified bond strength and a specified friction coefficient to obtain an initial Voronoi-GBM numerical model. The point cloud data of the joint surface morphology of the granite is mapped to the irregular contact surfaces of the hanging wall and the footwall of the initial Voronoi-GBM numerical model to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite. The multi-objective optimization function is: ， Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients.

2. The method according to claim 1, characterized in that, The acquisition of multi-condition basic data includes: The mineral type and mineral content of the granite were determined by X-ray diffraction testing; The basic mechanical parameters of the granite were obtained through uniaxial compression tests and Brazilian splitting tests. The peak shear strength, residual strength, and stress-displacement curves of the joints in the granite were obtained by direct shear test. The topographic point cloud data of the joint surface of the granite was obtained using a 3D laser scanner, and the roughness coefficient of the joint surface was determined. Micro-CT technology was used to scan the granite before and after shearing to obtain images of the fine crack distribution of the granite.

3. The method according to claim 1 or 2, characterized in that, ， ， ， ， ， This represents the peak shear strength obtained from the numerical simulation. σ represents the reference peak strength determined by indoor cyclic shear tests. n The normal stress is represented by JRC, the joint roughness coefficient, and N is the number of shear cycles. This represents the residual shear strength obtained from the numerical simulation. This represents the reference residual strength determined by the indoor cyclic load direct shear test.

4. The method according to claim 1, characterized in that, The optimization objectives of the multi-objective optimization function include the peak intensity error, stress reduction amplitude error, residual strength error of the macroscopic stress-displacement curve, and the similarity between the microscopic crack distribution pattern and the CT image. The optimization of the mesomechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function includes: By iteratively running numerical direct shear simulations, the micromechanical parameters of the Voronoi-GBM numerical model are dynamically adjusted until the error between the macromechanical response of the numerical direct shear output by the multi-objective optimization function and the indoor direct shear test results is less than a predetermined threshold, and the microcrack distribution pattern generated by the simulation matches the CT scan observation results. At this point, the adjustment results of the micromechanical parameters are output.

5. The method according to claim 4, characterized in that, The macroscopic stress drop characteristic parameters include one or more of the following: single stress drop amplitude, shear displacement when stress drop occurs, stress drop frequency per unit shear displacement, and cumulative stress drop energy release rate. The microcrack characteristic parameters are obtained based on microfracture events recorded in real time during the working condition simulation. The microcrack characteristic parameters include the number, increment, distribution density, and proportion of various types of microcracks in the total number of cracks.

6. The method according to claim 1, characterized in that, The step of training a target association representation model using the microscopic crack feature parameters as input samples and the macroscopic stress drop feature parameters as output samples includes: Construct a sample dataset with the microscopic crack characteristic parameters as input samples and the macroscopic stress drop characteristic parameters as output samples; The sample dataset is randomly divided into a training set and a test set according to a preset ratio; A basic association representation model is configured using regression analysis algorithms or machine learning algorithms, wherein the regression analysis algorithm includes multiple linear regression, and the machine learning algorithm includes support vector machine, random forest, or neural network; Using the microscopic crack feature parameters in the training set as input and the macroscopic stress drop feature parameters in the training set as labels, the basic correlation representation model is iteratively trained. The basic correlation representation model is iteratively updated by minimizing the loss function between the predicted value and the true value. The accuracy of the trained association representation model is verified using the test set, and the prediction error is calculated. If the prediction error exceeds a predetermined effective range, the association representation model is optimized and trained until a target association representation model is obtained where the prediction error does not exceed the predetermined effective range.

7. A device for characterizing the correlation between stress drop and microcracks in granite joints under multiple working conditions, characterized in that, For the method according to any one of claims 1 to 6, the apparatus comprises: A multi-condition basic data acquisition unit is used to acquire multi-condition basic data, wherein the multi-condition basic data is used to reflect the multi-dimensional rock characteristics of granite under multiple conditions. The Voronoi-GBM model construction unit is used to construct a Voronoi-GBM numerical model based on the multi-condition basic data, wherein the Voronoi-GBM numerical model is used to reflect the heterogeneous grain structure and true joint morphology of the granite. A multi-objective optimization unit is used to construct a multi-objective optimization function for optimizing various micromechanical parameters, and to optimize the micromechanical parameters of the Voronoi-GBM numerical model based on the multi-objective optimization function; The working condition simulation and monitoring unit is used to simulate various granite shear working conditions based on the optimized Voronoi-GBM numerical model, and to monitor the macroscopic stress drop characteristic parameters and microscopic crack characteristic parameters of the granite during the working condition simulation process. The correlation characterization model construction unit is used to train the target correlation characterization model with the microcrack characteristic parameters as input samples and the macrostress drop characteristic parameters as output samples. The target correlation characterization model is used to reflect the correlation between the microcrack characteristic parameters and the macrostress drop characteristic parameters in the granite shearing condition. The Voronoi-GBM model construction unit is used to: generate a corresponding set of nucleus points for each mineral type, wherein the ratio of the number of nucleus points corresponding to each of the mineral types is the same as the ratio of the mineral content of each mineral type in the granite; based on the set of nucleus points for each mineral type, construct a polygonal grain structure using a Voronoi graph partitioning algorithm, wherein each grain in the polygonal grain structure represents a mineral unit; fill each grain in the polygonal grain structure with fine-grained units, and assign a set of mesoscopic parameters corresponding to the mineral type to the grain; A parallel bonding model is introduced between the fine particle units within the grains, wherein the parallel bonding model is used to simulate intragranular materials; a smooth joint model is introduced between the fine particle units at the adjacent boundaries of every two grains, wherein the smooth joint model is configured with a specified bonding strength and a specified friction coefficient, to obtain an initial Voronoi-GBM numerical model; the topographic point cloud data of the joint surface of the granite is mapped to the irregular contact surfaces of the upper and lower plates of the initial Voronoi-GBM numerical model, to obtain a Voronoi-GBM numerical model that reflects the heterogeneous grain structure and true joint morphology of the granite; The multi-objective optimization function is: ， Used to reflect micromechanical parameters Under the given conditions, the overall degree of error between the numerical simulation results and the indoor test results, The peak strength error of the macroscopic stress-displacement curve. The residual strength error is represented by the macroscopic stress-displacement curve. This represents the error in the stress drop amplitude of the macroscopic stress-displacement curve. , , They are respectively , and The weighting coefficients.

8. A computer device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, the instructions being configured to cause the processor to perform the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The device stores computer-executable instructions configured to perform the method as described in any one of claims 1 to 6.

Images