A method, device, equipment and medium for determining parameters of a block discrete element method

Abstract

The application discloses a block discrete element method parameter determination method and device, equipment and medium, and relates to the technical field of rock mechanical behavior simulation. The method comprises the following steps: determining a plurality of parameter value combinations based on the value range of each block and contact parameter; for each parameter value combination, taking the parameter value combination as input, determining the predicted value of the macroscopic mechanical parameter by using a trained macroscopic mechanical parameter prediction model, calculating the deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter, obtaining the loss value corresponding to the parameter value combination, and selecting the parameter value combination with the minimum loss value as the final parameter value combination. The application can solve the problems of poor authenticity, low efficiency and low fault tolerance of the traditional block discrete element method parameter determination method, and is beneficial to the simulation of rock mechanical behavior.

Description

Technical Field

[0001] This application relates to the field of rock mechanical behavior simulation technology, and in particular to a method, apparatus, equipment and medium for determining parameters using the block discrete element method. Background Technology

[0002] Due to the invisibility of internal fractures in rocks, numerical simulations are often used to reproduce their mechanical behavior, thereby providing a basis for the prevention and control of geological disasters in rock engineering.

[0003] Over long geological processes, rocks typically contain numerous discontinuities, which significantly impact their physical and mechanical properties, making the simulation of rock mechanical behavior extremely difficult. To address this complex characteristic, the block discrete element method treats rocks as discontinuous media, discretizing them into a collection of rigid or deformable polygonal blocks. These blocks are bonded together through contact, and the kinematic formulas of the contacts and blocks are calculated to simulate the rock's mechanical behavior. Because the block discrete element method can realistically simulate rock fracture and nonlinear large deformation, it has been widely applied in rock engineering simulation.

[0004] In the block discrete element method, the simulation of rock mechanical behavior essentially depends on the block and contact parameters, i.e., the properties of the simulated minerals and mineral boundaries. However, the block and contact parameters are microscopic parameters, and current techniques cannot directly measure them experimentally, because experiments can only measure macroscopic parameters, nor can they be indirectly derived from other measurable material physical parameters. Therefore, accurate block and contact parameters are a prerequisite for reproducing rock mechanical behavior when applying the block discrete element method.

[0005] When simulating rock mechanical behavior using the block discrete element method, the block and contact parameters cannot be directly obtained. Therefore, their determination typically relies on a trial-and-error approach. This method involves repeatedly adjusting the block and contact parameters until the simulated macroscopic mechanical parameters match those obtained in laboratory experiments. Furthermore, a fitting method has been developed based on this trial-and-error approach. This involves establishing a fitting curve characterizing the relationship between the block and contact parameters and the simulated macroscopic mechanical parameters based on extensive trial and error. The fitting curve is then used to determine the block and contact parameters that meet the macroscopic mechanical parameter requirements of the laboratory experiments.

[0006] However, trial and error and fitting methods have many problems in practical applications, as detailed below.

[0007] (1) Poor realism. The trial-and-error method relies on the process of gradually adjusting the block and contact parameters and observing the results, while the fitting method relies on the relationship between the block and contact parameters and macroscopic mechanical parameters under a locally preset gradient (i.e., the interval between two adjacent values). Therefore, it is difficult for either method to guarantee that the found parameter combination is globally optimal, and often only achieves local optimality. This limitation often leads to a certain deviation between the simulation results and the actual situation, or only meets the corresponding requirements of some block and contact parameters, which greatly reduces the realism of the simulation. In addition, since the trial-and-error method relies on human adjustment and observation, and the fitting method relies on human gradient setting, it is highly subjective.

[0008] (2) Low efficiency. Due to the large number of block and contact parameters involved, the trial-and-error method and the fitting method usually require a large number of discrete element numerical simulations. Each time the parameters are adjusted, they need to be run again, resulting in high computational costs and time consumption.

[0009] (3) Low tolerance for error. When dealing with complex parameter spaces with high-dimensional and nonlinear complex relationships, trial and error and fitting methods face huge challenges. Multiple blocks and contact parameters in the complex parameter space may interact with each other, making it difficult to find a suitable combination of parameter values ​​through simple adjustments. As the complex parameter space increases, trial and error and fitting methods gradually lose their effectiveness, especially when facing complex physical phenomena. Summary of the Invention

[0010] The purpose of this application is to provide a method, apparatus, equipment, and medium for determining parameters using the block discrete element method, which can solve the problems of poor realism, low efficiency, and low fault tolerance in traditional block discrete element method parameter determination methods.

[0011] To achieve the above objectives, this application provides the following solution.

[0012] In a first aspect, this application provides a method for determining parameters using a block discrete element method, which includes the following steps.

[0013] Based on the value range of each block and contact parameter, multiple parameter value combinations are determined; the block and contact parameters are modeling parameters in the block discrete element method, and the parameter value combinations include the values ​​of each block and contact parameter.

[0014] For each parameter value combination, using the parameter value combination as input, the predicted value of the macroscopic mechanical parameter is determined using the trained macroscopic mechanical parameter prediction model. The deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter is calculated to obtain the loss value corresponding to the parameter value combination. The first experimental value is the value of the macroscopic mechanical parameter obtained by conducting experiments on the mechanical behavior of the rock sample under the parameter value combination. The macroscopic mechanical parameter includes rock deformation characteristics and rock strength characteristics. The rock deformation characteristics include Young's modulus and Poisson's ratio. The rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle.

[0015] The parameter value combination that minimizes the loss value is selected as the final parameter value combination.

[0016] Secondly, this application provides a parameter determination device for the block discrete element method, which includes the following modules.

[0017] The data acquisition module is used to determine multiple parameter value combinations based on the value range of each block and contact parameter; the block and contact parameters are modeling parameters in the block discrete element method, and the parameter value combinations include the values ​​of each block and contact parameter.

[0018] The loss calculation module is used to determine the predicted value of the macroscopic mechanical parameters using a trained macroscopic mechanical parameter prediction model for each parameter value combination, taking the parameter value combination as input, and calculating the deviation between the predicted value of the macroscopic mechanical parameters and the first experimental value of the macroscopic mechanical parameters to obtain the loss value corresponding to the parameter value combination; the first experimental value is the value of the macroscopic mechanical parameters obtained by conducting experiments on the mechanical behavior of the rock sample under the parameter value combination, the macroscopic mechanical parameters include rock deformation characteristics and rock strength characteristics, the rock deformation characteristics include Young's modulus and Poisson's ratio, and the rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion and internal friction angle.

[0019] The parameter determination module is used to select the parameter value combination with the minimum loss value as the final parameter value combination.

[0020] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for determining parameters using the block discrete element method.

[0021] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for determining parameters using the block discrete element method.

[0022] According to the specific embodiments provided in this application, this application has the following technical effects.

[0023] This application provides a method, apparatus, device, and medium for determining parameters using the block discrete element method. Based on the value range of each block and contact parameter, multiple parameter value combinations are determined. For each parameter value combination, using the combination as input, a trained macroscopic mechanical parameter prediction model is used to determine the predicted value of the macroscopic mechanical parameter. The deviation between the predicted value and the first experimental value of the macroscopic mechanical parameter is calculated to obtain the loss value corresponding to the parameter value combination. The parameter value combination with the smallest loss value is selected as the final parameter value combination. This application can traverse multiple parameter value combinations of the block and contact parameters, reducing subjectivity in parameter selection and solving the problem of inaccurate prediction. Using a trained macroscopic mechanical parameter prediction model to predict macroscopic mechanical parameters avoids extensive discrete element numerical simulations, solving the problem of low efficiency. Simultaneously, it can handle complex parameter spaces with high dimensions and nonlinear complex relationships, solving the problem of low fault tolerance. Therefore, it can solve the problems of poor accuracy, low efficiency, and low fault tolerance in traditional block discrete element method parameter determination methods, accurately and efficiently determining the parameters of the block discrete element method. In summary, this application reduces subjectivity and avoids a large number of discrete element numerical simulations by traversing multiple parameter value combinations. At the same time, it can handle complex high-dimensional nonlinear parameter spaces and solve the problems of poor realism, low efficiency and low fault tolerance, thereby realizing an accurate and efficient parameter determination process for the block discrete element method. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, 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 This is an application environment diagram of a block discrete element method parameter determination method provided in Embodiment 1 of this application.

[0026] Figure 2 This is a flowchart illustrating a method for determining parameters using a block discrete element method, as provided in Embodiment 1 of this application.

[0027] Figure 3 This is a schematic diagram of the technical route for a method of determining parameters using a block discrete element method, as provided in Embodiment 1 of this application.

[0028] Figure 4This is a schematic diagram comparing the numerical simulation results and experimental data of the Young's modulus of igneous rocks provided in Example 1 of this application.

[0029] Figure 5 This is a schematic diagram comparing the numerical simulation results and experimental data of the Poisson's ratio of igneous rocks provided in Example 1 of this application.

[0030] Figure 6 This is a schematic diagram comparing the numerical simulation results and experimental data of the uniaxial compressive strength of igneous rocks provided in Example 1 of this application.

[0031] Figure 7 This is a schematic diagram comparing the numerical simulation results and experimental data of the Brazilian tensile strength of igneous rocks provided in Example 1 of this application.

[0032] Figure 8 This is a schematic diagram comparing the numerical simulation results and experimental data of the cohesion of igneous rocks provided in Example 1 of this application.

[0033] Figure 9 This is a schematic diagram comparing the numerical simulation results and experimental data of the internal friction angle of igneous rocks provided in Example 1 of this application.

[0034] Figure 10 This is a schematic diagram of the functional modules of a block discrete element method parameter determination device provided in Embodiment 2 of this application.

[0035] Figure 11 This is a schematic diagram of the structure of a computer device provided in Embodiment 3 of this application. Detailed Implementation

[0036] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] Example 1.

[0038] The parameter determination method for the block discrete element method provided in this application can be applied to, for example... Figure 1The application environment shown depicts a terminal communicating with a server via a network. A data storage system stores the data the server needs to process. This system can be set up independently, integrated into the server, or located in the cloud or on another server. The terminal can send multiple combinations of parameter values ​​for the block to be processed and contact parameters to the server. Upon receiving these combinations, the server, for each parameter value combination, uses the combination as input and a trained macroscopic mechanical parameter prediction model to determine the predicted value of the macroscopic mechanical parameter. It then calculates the deviation between the predicted value and the first experimental value of the macroscopic mechanical parameter to obtain the loss value corresponding to the parameter value combination. The parameter value combination with the smallest loss value is selected as the final parameter value combination. The server can then feed back the final parameter value combination to the terminal.

[0039] Furthermore, in some embodiments, the parameter determination method of the block discrete element method can also be implemented by the server or the terminal alone. For example, the terminal can directly process multiple parameter value combinations of the block and contact parameters to be processed, or the server can obtain multiple parameter value combinations of the block and contact parameters to be processed from the data storage system and process them.

[0040] The terminals can be, but are not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices, while portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. Servers can be implemented using independent servers, server clusters composed of multiple servers, or cloud servers.

[0041] In one exemplary embodiment, such as Figure 2 As shown, a method for determining parameters using a block discrete element method is provided. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is applied to... Figure 1 The following steps are used as an example of a server in the example.

[0042] Step S1: Based on the value range of each block and contact parameter, determine multiple parameter value combinations; the block and contact parameters are modeling parameters in the block discrete element method, and the parameter value combinations include the values ​​of each block and contact parameter.

[0043] Step S2: For each parameter value combination, using the parameter value combination as input, the predicted value of the macroscopic mechanical parameter is determined using the trained macroscopic mechanical parameter prediction model. The deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter is calculated to obtain the loss value corresponding to the parameter value combination. The first experimental value is the value of the macroscopic mechanical parameter obtained by conducting experiments on the mechanical behavior of the rock sample under the parameter value combination. The macroscopic mechanical parameter includes rock deformation characteristics and rock strength characteristics. The rock deformation characteristics include Young's modulus and Poisson's ratio. The rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle.

[0044] Step S3: Select the parameter value combination with the smallest loss value as the final parameter value combination.

[0045] To determine block and contact parameters, traditional techniques such as trial and error and fitting methods are currently widely used. However, these traditional techniques suffer from poor realism, low efficiency, and low error tolerance, posing a significant challenge to the rapid and accurate determination of block and contact parameters. In recent years, artificial intelligence and machine learning technologies have demonstrated strong potential in parameter optimization and uncertainty analysis, especially in handling complex, nonlinear relationships and high-dimensional parameter spaces. Therefore, this embodiment applies machine learning to quickly and accurately determine block and contact parameters in the block discrete element method. Specifically, addressing the problems of poor realism, low efficiency, and low error tolerance in traditional block discrete element method parameter determination methods, a machine learning-based block discrete element method parameter determination method is proposed. Implementing steps S1 to S3 above can solve the problems of poor realism, low efficiency, and low error tolerance in traditional block discrete element method parameter determination methods, accurately and quickly determining the parameters of the block discrete element method, thereby serving rock mechanics experiments and simulation of geological disaster prevention and control in rock engineering.

[0046] This embodiment uses a numerical simulation tool based on the block discrete element method to obtain input and output parameters, constructs a dataset, and builds a multi-scale pre-trained model framework including a random forest model (RF), a Gaussian process regression model (GPR), a support vector regression model (SVR), and a deep neural network model (DNN). This multi-scale pre-trained model framework is trained using the dataset to obtain a trained macroscopic mechanical parameter prediction model. Based on this trained macroscopic mechanical parameter prediction model, parameter inversion is performed using a grid search method, specifically defining the value range and creating a parameter network (i.e., determining multiple parameter value combinations). The optimal parameter value combination for the block discrete element method is determined by minimizing the loss value. Finally, simulation results and experimental results are used for verification. Figure 3 As shown, the specific steps include the following.

[0047] (a) Collection and preparation of datasets.

[0048] Based on the properties of rock materials and previous research, the variation range and gradient (i.e., the interval between two adjacent values) of microscopic block and contact parameters are reasonably set to generate multiple parameter value combinations. Each parameter value combination includes the value of each block and contact parameter. The values ​​of each block and contact parameter in different parameter value combinations are not entirely the same. The block and contact parameters can include deformation-related parameters and strength-related parameters. Deformation-related parameters can include contact normal stiffness. k n Contact shear stiffness k s The ratio of contact normal stiffness to contact shear stiffness k n / k s Young's modulus of rock E b Poisson's ratio of rocks ν b Strength-related parameters may include contact cohesion. c cont Contact internal friction angle cont and contact tensile strength σ t cont A numerical simulation model of the rock sample is established in discrete element method (DEM) software. The geometry and boundary conditions of the rock sample are defined, and the initial size of the rock particles is set. The initial size can be 2 mm, 3 mm, or 4 mm, etc., which can be set according to user needs. Size division is to enhance the universality of the results and construct the corresponding numerical simulation model. According to the experimental design, different combinations of parameter values ​​are input into the numerical simulation model. The numerical simulation model is then run, and the simulation results of the rock sample under each parameter combination are recorded to obtain the simulated values ​​of macroscopic mechanical parameters. These macroscopic mechanical parameters may include rock deformation characteristics and rock strength characteristics. Rock deformation characteristics may include Young's modulus. E Compared to Poisson μ Rock strength properties can include uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), and cohesion. c and internal friction angle The simulation results are checked for rationality, and abnormal or non-physical results are eliminated. Finally, the dataset is obtained by using the preset parameter combinations and the simulated values ​​of macroscopic mechanical parameters obtained in the numerical simulation as the input and output parameters.

[0049] (ii) Establish four different machine learning models.

[0050] (1) Divide the dataset into a training set and a test set to ensure that the data in each subset is evenly distributed. The main purpose of the training set is to allow the model to learn the patterns and features of the data, i.e., to train the model. The test set is used to evaluate the model performance.

[0051] (2) Determine whether the four machine learning models—random forest model, Gaussian process regression model, support vector regression model, and deep neural network model—have been established.

[0052] (3) If not, return to step (2) and build each machine learning model in turn for training.

[0053] (4) If so, proceed to the next step.

[0054] (5) Perform hyperparameter optimization. Specifically, use grid search to determine multiple hyperparameter combinations. For each hyperparameter combination, split the training set into a training subset and a validation subset. Use the training subset for training and test the training effect on the validation subset (accuracy can be used to characterize it). Finally, select the optimal hyperparameter combination based on the training effect. The main hyperparameter optimization methods are GridSearchCV and RandomizedSearchCV. This method will automatically perform cross-validation. The default is to use 5-fold cross-validation. Depending on the size of the dataset, this embodiment can select 5-fold cross-validation or 10-fold cross-validation.

[0055] (6) Determine whether each machine learning model using the optimal hyperparameter combination has achieved the training objective. The training objective can be to reach the maximum number of iterations or the loss value is less than the preset value.

[0056] (7) If not, return to step (6).

[0057] (8) If so, proceed to the next stage, save the trained machine learning model, and obtain the trained macroscopic mechanical parameter prediction model. At the same time, the importance of features can be analyzed based on the trained macroscopic mechanical parameter prediction model, and how the block and contact parameters affect the macroscopic mechanical parameters can be explored. For example, in the random forest model, the importance ranking of features can be viewed directly.

[0058] In this embodiment, before determining the predicted values ​​of macroscopic mechanical parameters using the trained macroscopic mechanical parameter prediction model with parameter value combinations as input, the block discrete element method parameter determination method of this embodiment further includes: acquiring a dataset, which includes multiple samples and labels corresponding to each sample; the samples are parameter value combinations of block and contact parameters; the labels are the first simulated values ​​of macroscopic mechanical parameters, which are the values ​​of macroscopic mechanical parameters obtained by simulating the mechanical behavior of rock samples using the block discrete element method under the samples; establishing multiple initial prediction models, which are machine learning models; for each initial prediction model, training the initial prediction model using the dataset to obtain the trained model and the prediction performance of the trained model; and selecting the trained model with the best prediction performance as the trained macroscopic mechanical parameter prediction model.

[0059] The initial prediction models include random forest, Gaussian process regression, support vector regression, and deep neural network models.

[0060] The process of training an initial prediction model using a dataset to obtain a trained model and its prediction performance includes: dividing the dataset into a training set and a test set; training the initial prediction model using the training set to obtain a trained model; using the test set to determine the prediction performance of the trained model; if the prediction performance does not meet the requirements, using the trained model as the initial prediction model for the next iteration and returning to the step of "training the initial prediction model using the training set"; if the prediction performance meets the requirements, using the trained model as the trained model and using the test set to determine the prediction performance of the trained model.

[0061] Among them, prediction performance can be characterized by accuracy, and the prediction performance meeting the requirements means that the prediction performance is greater than the preset value.

[0062] (iii) Parameter inversion is performed using the grid search method.

[0063] (1) Based on the range of values ​​of blocks and contact parameters in the dataset, determine the reasonable range of values ​​for each block and contact parameter, and set the step size, i.e. the size of the parameter increment each time in the grid search, which is beneficial for generating parameter value combinations in the future.

[0064] (2) Based on the defined range of values ​​and step size, determine whether all possible combinations of parameter values ​​have been generated.

[0065] (3) If not, return to step (2).

[0066] (4) If so, proceed to the next step.

[0067] (5) Initialize the minimum loss value to a larger value and proceed to the next step.

[0068] (6) Input the current parameter value combination into the trained macroscopic mechanical parameter prediction model to predict the macroscopic mechanical parameters of the rock sample. Based on the deviation between the prediction results of the trained macroscopic mechanical parameter prediction model and the actual experimental data, calculate the loss value of the current parameter value combination. The loss value is obtained by summing the square of the difference between each predicted value and the experimental value.

[0069] (7) Determine whether the loss value of the current parameter combination is less than the recorded minimum loss value.

[0070] (8) If not, skip the update and continue to the next parameter value combination.

[0071] (9) If so, update the minimum loss value and set the current parameter value combination as the optimal parameter value combination.

[0072] (10) Determine whether all parameter value combinations have been traversed.

[0073] (11) If not, select an untraversed parameter value combination as the current parameter value combination and return to step (6).

[0074] (12) If so, the grid search ends and the next stage begins. Finally, the parameter combination with the minimum loss value is recorded and output. This parameter combination represents the values ​​of the block and contact parameters that make the macroscopic mechanical parameters predicted by the model closest to the experimental data.

[0075] In this embodiment, based on the value range of each block and contact parameter, multiple parameter value combinations are determined. The block and contact parameters are modeling parameters in the block discrete element method, and the parameter value combinations include the values ​​of each block and contact parameter. For each parameter value combination, the parameter value combination is used as input, and the predicted value of the macroscopic mechanical parameter is determined using the trained macroscopic mechanical parameter prediction model. The deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter is calculated to obtain the loss value corresponding to the parameter value combination. The first experimental value is the value of the macroscopic mechanical parameter obtained by conducting experiments on the mechanical behavior of the rock sample under the parameter value combination. The parameter value combination with the smallest loss value is selected as the final parameter value combination.

[0076] The block and contact parameters include deformation-related parameters and strength-related parameters. Deformation-related parameters include contact normal stiffness, contact shear stiffness, the ratio of contact normal stiffness to contact shear stiffness, Young's modulus, and Poisson's ratio. Strength-related parameters include contact cohesion, contact internal friction angle, and contact tensile strength. Macroscopic mechanical parameters include rock deformation characteristics and rock strength characteristics. Rock deformation characteristics include Young's modulus and Poisson's ratio. Rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle.

[0077] Specifically, based on the value range of each block and contact parameter, multiple parameter value combinations are determined. This includes: obtaining the value range and step size of each block and contact parameter; for each block and contact parameter, taking multiple values ​​within the value range based on the step size to obtain multiple values ​​for the block and contact parameter, that is, generating multiple values ​​within the value range based on the value range and step size, specifically starting from the lower limit of the value range and gradually increasing the value based on the step size until the upper limit of the value range is reached; and randomly combining the multiple values ​​of all blocks and contact parameters to obtain multiple parameter value combinations for the block and contact parameters.

[0078] Specifically, the calculation of the deviation between the predicted value and the first experimental value of the macroscopic mechanical parameters yields the loss value corresponding to the parameter value combination. This includes: calculating the difference between the predicted value and the first experimental value of each macroscopic mechanical parameter (including Young's modulus, Poisson's ratio, uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle); and calculating the sum of the squares of all differences to obtain the loss value corresponding to the parameter value combination.

[0079] (iv) Simulation and comparison verification using the block discrete element method.

[0080] (1) Input the parameter combination with the minimum loss value predicted by machine learning into the discrete element method software to perform uniaxial compression numerical simulation, triaxial compression numerical simulation and Brazilian splitting numerical simulation. The numerical simulation model size and boundary conditions are the same as those of the specimens and loading conditions used for uniaxial compression, triaxial compression and Brazilian splitting, and the simulated values ​​of macroscopic mechanical parameters are obtained.

[0081] (2) Determine whether the macroscopic mechanical parameters obtained by numerical simulation are equal to those obtained by experiment.

[0082] (3) If not, return to the previous process. You can return to the training model step or the parameter back-inference step. If the loss value is large due to poor model performance, you can choose to return to the training model step. If the loss value is large due to improper parameter settings, you can choose to return to the parameter back-inference step, that is, adjust the machine learning model or adjust the value range, perform parameter back-inference again, and then input it into the discrete unit method software for verification.

[0083] (4) If so, the parameter combination with the minimum loss value predicted by machine learning is directly output as the modeling parameter of the block discrete unit method.

[0084] In this embodiment, the parameter value combination with the minimum loss value is selected as the final parameter value combination. Specifically, this includes: selecting the parameter value combination with the minimum loss value as a candidate parameter value combination; obtaining the second simulated value and the second experimental value of the macroscopic mechanical parameters corresponding to the candidate parameter value combination. The second simulated value is the value of the macroscopic mechanical parameter obtained by simulating the mechanical behavior of the rock sample using the block discrete element method under the candidate parameter value combination; the second experimental value is the macroscopic mechanical parameter obtained by experimentally testing the mechanical behavior of the rock sample under the candidate parameter value combination. The algorithm determines whether the second simulated value and the second experimental value of the macroscopic mechanical parameter are equal. If they are equal, the candidate parameter combination is used as the final parameter combination. If not, the algorithm adjusts the value range of each block and contact parameter and returns to the step of "determining multiple parameter combinations based on the value range of each block and contact parameter", or updates the trained macroscopic mechanical parameter prediction model and returns to the step of "for each parameter combination, using the parameter combination as input, using the trained macroscopic mechanical parameter prediction model to determine the predicted value of the macroscopic mechanical parameter".

[0085] The method in this embodiment has wide applicability and is applicable to multiple block discrete element method software such as UDEC (Universal Distinct Element Code) and 3DEC (3-Dimensional Distinct Element Code). It is suitable for any size within a reasonable range and can be applied to different lithological materials, such as igneous rocks, sedimentary rocks, and metamorphic rocks. In the following embodiment, we will specifically use UDEC software, select 2 mm, 3 mm, and 4 mm, and igneous rocks as examples for analysis and explanation. The results analysis is based on igneous rocks, but the results of other lithologies are also fully considered to ensure the universality and accuracy of the method.

[0086] The parameter space includes deformation-related parameters and strength-related parameters. Reasonable gradients are set, and corresponding parameter value combinations are generated. Numerical simulations of uniaxial compression, triaxial compression, and Brazilian splitting are performed using UDEC software. The compression simulation model has a width of 50 mm and a height of 100 mm, while the splitting simulation model has a diameter of 50 mm, conforming to ASTM (American Society of Testing Materials) standards. All simulation models are composed of deformable polygons bonded together with contact points forming a Voronoi mesh. The contact points follow the classical Coulomb friction law. Each simulation model is sandwiched between upper and lower simulation steel plates, each 50 mm wide and 10 mm high. The simulation steel plates have a width of 50 mm, a height of 10 mm, a Young's modulus of 140 GPa, and a Poisson's ratio of 0.28. The contact normal stiffness of the interface between the simulation model and the simulation steel plates is also considered. k n Set to 10 6 GPa / m, and no contact shear stiffness k s Contact cohesion c cont Contact internal friction angle cont and contact tensile strength σ t cont To eliminate boundary effects, during the simulation process, the lower numerical simulation steel plate is fixed, while the upper numerical simulation steel plate compresses the numerical simulation model at a constant speed of 0.015 m / s.

[0087] Based on the pre-defined combinations of interparticle bulk and contact parameters and their corresponding macroscopic mechanical parameters from numerical simulations, machine learning models were trained and their performance evaluated. Specifically, a multi-scale pre-trained model framework was constructed, including a random forest model, a Gaussian process regression model, a support vector regression model, and a deep neural network model. Results show that the random forest model exhibits the best predictive performance, with a prediction accuracy consistently above 0.99. The deep neural network model demonstrates significant advantages in handling complex nonlinear relationships, with prediction accuracies ranging from 0.97 to 0.99, making it particularly suitable for simulating mechanical behaviors with highly nonlinear characteristics. In contrast, the support vector regression model and the Gaussian process regression model show relatively poor predictive performance, especially for a 3 mm particle size, where the prediction accuracy for uniaxial compressive strength and cohesion significantly decreases.

[0088] Taking a deep neural network model as an example, the grid search method is applied to perform parameter inversion, define the range of parameter values ​​and create a parameter network, and input all possible combinations of parameter values ​​into the trained deep neural network model in sequence to predict the macroscopic mechanical parameters of the rock sample. By calculating the deviation between the model prediction results and the experimental data, the optimal combination of parameter values ​​that makes the model prediction accuracy is finally determined.

[0089] At this point, the optimal parameter combinations predicted by the machine learning model, including contact normal stiffness, contact cohesion, contact internal friction angle, and contact tensile strength, have been determined. These optimal parameter combinations are then input into the UDEC software for uniaxial compression numerical simulation and Brazilian splitting numerical simulation, and the simulation results are compared with experimental results. Taking igneous rocks as an example... Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 and Figure 9 This paper presents a comparative analysis between numerical simulation results and experimental data, covering key macroscopic mechanical parameters such as Young's modulus, Poisson's ratio, uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle. The predictive performance of the model is quantitatively evaluated by calculating statistical indicators (mean square error, mean absolute error, coefficient of determination, and relative error) for each macroscopic mechanical parameter. Figures 4-9 Predicted values ​​1, 2, and 3 correspond to predicted dimensions of 2 mm, 3 mm, and 4 mm, respectively. l The results show that as the preset size increases from 2 mm to 4 mm, the prediction accuracy of most macroscopic mechanical parameters significantly improves, with Young's modulus and uniaxial compressive strength being particularly accurate. From a lithological perspective, igneous rocks show the best prediction performance, especially at the 4 mm size. Sedimentary rocks show slightly lower prediction performance than igneous rocks, while metamorphic rocks show relatively poor prediction performance, particularly with significant deviations in the prediction of internal friction angle and Brazilian tensile strength. Young's modulus and uniaxial compressive strength have high prediction accuracy across all lithologies, with coefficients of determination close to 1. However, the prediction errors for internal friction angle and Brazilian tensile strength are relatively large, especially in sedimentary and metamorphic rocks. Overall, the model's prediction results have high accuracy and can effectively reflect the mechanical properties of different lithological materials, successfully achieving rapid and accurate determination of parameters using the block discrete element method.

[0090] This embodiment uses machine learning to establish the relationship between simulated block and contact parameters and macroscopic mechanical parameters, and uses the grid search method for parameter inversion, which has the following advantages.

[0091] (1) It can accurately obtain real block and contact parameters. The subjectivity and uncertainty of trial and error and fitting methods may miss some optimal solutions, especially when the parameter space is large. Trial and error methods may not be able to fully explore all possible combinations of parameter values, thus missing the opportunity to find the optimal solution. Machine learning, on the other hand, is data-driven and can learn potential patterns and rules by using a large amount of numerical simulation data. Machine learning can also dynamically adjust hyperparameters during model training to optimize performance, thereby improving realism, and can traverse all possible combinations of parameter values. This method reduces the subjectivity in parameter selection and makes the parameter optimization process more objective and systematic.

[0092] (2) High computational efficiency. Machine learning can quickly find the optimal combination of parameter values ​​through efficient optimization algorithms (such as grid search). Through automated data processing and model training, it can significantly reduce the time and workload of manually adjusting parameters. Compared with traditional trial and error and fitting methods, especially when dealing with large amounts of simulated data, it can make more effective use of computing resources. Machine learning reduces computing requirements in an intelligent way, which greatly saves computing costs.

[0093] (3) High fault tolerance. Machine learning models can handle complex nonlinear relationships between input parameters and output results, which is especially important for complex physical phenomena in the block discrete element method. Machine learning can handle high-dimensional and complex parameter spaces, and is particularly suitable for situations with many combinations of micro parameters and complex interactions in the discrete element method.

[0094] This application also provides an application scenario in which the above-described block discrete element method for parameter determination is applied. Specifically, the block discrete element method for parameter determination provided in this embodiment can be applied to a rock mechanical behavior simulation scenario. The rock mechanical behavior simulation scenario includes a parameter determination stage and a simulation stage. The parameter determination stage is used to determine the final combination of parameter values, and the simulation stage is used to simulate the rock mechanical behavior based on the final combination of parameter values. The block discrete element method for parameter determination provided in this embodiment belongs to the parameter determination stage.

[0095] Example 2.

[0096] Based on the same inventive concept, this application also provides a device for determining the parameters of the block discrete element method for implementing the aforementioned method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the device for determining the parameters of the block discrete element method provided below can be found in the limitations of the method described above, and will not be repeated here.

[0097] In one exemplary embodiment, such as Figure 10 As shown, a parameter determination device for the block discrete element method is provided, which includes the following modules.

[0098] The data acquisition module M1 is used to determine multiple parameter value combinations based on the value range of each block and contact parameter; the block and contact parameters are modeling parameters in the block discrete element method, and the parameter value combinations include the values ​​of each block and contact parameter.

[0099] The loss calculation module M2 is used to determine the predicted value of the macroscopic mechanical parameters using a trained macroscopic mechanical parameter prediction model for each parameter value combination, taking the parameter value combination as input, and calculating the deviation between the predicted value of the macroscopic mechanical parameters and the first experimental value of the macroscopic mechanical parameters to obtain the loss value corresponding to the parameter value combination; the first experimental value is the value of the macroscopic mechanical parameters obtained by conducting experiments on the mechanical behavior of the rock sample under the parameter value combination, the macroscopic mechanical parameters include rock deformation characteristics and rock strength characteristics, the rock deformation characteristics include Young's modulus and Poisson's ratio, and the rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion and internal friction angle.

[0100] The parameter determination module M3 is used to select the parameter value combination with the minimum loss value as the final parameter value combination.

[0101] Example 3.

[0102] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 11 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a block discrete element method for parameter determination.

[0103] Those skilled in the art will understand that Figure 11The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0104] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the block discrete element method parameter determination method of embodiment 1.

[0105] Example 4.

[0106] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the block discrete element method parameter determination method of Embodiment 1.

[0107] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0108] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0109] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for determining parameters of a block discrete element method, characterized in that, The block discrete element method parameter determination method comprises: determining a plurality of parameter value combinations based on the value range of each block and contact parameter; the block and contact parameter is a modeling parameter in the block discrete element method, the block and contact parameter comprises a deformation-related parameter and a strength-related parameter, the deformation-related parameter comprises a contact normal stiffness, a contact shear stiffness, a ratio of the contact normal stiffness and the contact shear stiffness, a rock Young's modulus, and a rock Poisson's ratio, the strength-related parameter comprises a contact cohesion, a contact internal friction angle, and a contact tensile strength, and the parameter value combination comprises the value of each block and contact parameter; for each parameter value combination, determining a predicted value of a macroscopic mechanical parameter by using the trained macroscopic mechanical parameter prediction model with the parameter value combination as input, calculating the deviation between the predicted value of the macroscopic mechanical parameter and a first experimental value of the macroscopic mechanical parameter, and obtaining a loss value corresponding to the parameter value combination; the first experimental value is a value of the macroscopic mechanical parameter obtained by performing an experiment on the mechanical behavior of a rock sample under the parameter value combination, the macroscopic mechanical parameter comprises a rock deformation characteristic and a rock strength characteristic, the rock deformation characteristic comprises a Young's modulus and a Poisson's ratio, and the rock strength characteristic comprises a uniaxial compressive strength, a Brazilian tensile strength, a cohesion, and an internal friction angle; selecting the parameter value combination with the minimum loss value as the final parameter value combination; determining a plurality of parameter value combinations based on the value range of each block and contact parameter, specifically comprising: obtaining the value range and step size of each block and contact parameter; for each block and contact parameter, based on the step size, the value is taken multiple times in the value range, starting from the lower limit of the value range, gradually increasing the value based on the step size, until reaching the upper limit of the value range, obtaining a plurality of values of the block and contact parameter; randomly combining the plurality of values of all the block and contact parameters to obtain a plurality of parameter value combinations; calculating the deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter to obtain the loss value corresponding to the parameter value combination, specifically comprising: calculating the difference between the predicted value and the first experimental value of each macroscopic mechanical parameter, and calculating the sum of the squares of all the differences to obtain the loss value corresponding to the parameter value combination.

2. The discrete element method parameter determination method of a granular material mass according to claim 1, wherein, Before determining the predicted value of the macroscopic mechanical parameter by using the trained macroscopic mechanical parameter prediction model with the parameter value combination as input, the block discrete element method parameter determination method further comprises: obtaining a data set; the data set comprises a plurality of samples and a label corresponding to each sample, the sample is a parameter value combination, and the label is a first simulated value of a macroscopic mechanical parameter, which is a value of the macroscopic mechanical parameter obtained by simulating the mechanical behavior of a rock sample under the sample by using the block discrete element method; establishing a plurality of initial prediction models; the initial prediction model is a machine learning model; For each of the initial prediction models, the initial prediction models are trained by using the data set to obtain a trained model and a prediction performance of the trained model; The trained model with the best prediction performance is selected as the trained macroscopic mechanical parameter prediction model.

3. The method of claim 2, wherein, The initial prediction models include a random forest model, a Gaussian process regression model, a support vector regression model, and a deep neural network model.

4. The method of claim 1, wherein, The parameter value combination with the minimum loss value is selected as the final parameter value combination, specifically including: The parameter value combination with the minimum loss value is selected as the candidate parameter value combination; A second simulation value of the macroscopic mechanical parameter corresponding to the candidate parameter value combination and a second experimental value of the macroscopic mechanical parameter are obtained; the second simulation value is a value of the macroscopic mechanical parameter obtained by simulating the mechanical behavior of the rock sample by using the block discrete element method under the candidate parameter value combination, and the second experimental value is a value of the macroscopic mechanical parameter obtained by performing an experiment on the mechanical behavior of the rock sample under the candidate parameter value combination; It is judged whether the second simulation value of the macroscopic mechanical parameter and the second experimental value of the macroscopic mechanical parameter are equal; If yes, the candidate parameter value combination is taken as the final parameter value combination; If not, the value range of each block and contact parameter is adjusted, and the step of "determining a plurality of parameter value combinations based on the value range of each block and contact parameter" is returned, or the trained macroscopic mechanical parameter prediction model is updated, and the step of "for each of the parameter value combinations, using the trained macroscopic mechanical parameter prediction model to determine a predicted value of the macroscopic mechanical parameter by taking the parameter value combination as input" is returned.

5. A bulk discrete element method parameter determination apparatus characterized by comprising: The block discrete element method parameter determination device includes: A data acquisition module is configured to determine a plurality of parameter value combinations based on the value range of each block and contact parameter; the block and contact parameters are modeling parameters in the block discrete element method, the block and contact parameters include deformation-related parameters and strength-related parameters, the deformation-related parameters include contact normal stiffness, contact shear stiffness, a ratio of contact normal stiffness and contact shear stiffness, rock Young's modulus, and rock Poisson's ratio, the strength-related parameters include contact cohesion, contact internal friction angle, and contact tensile strength, and the parameter value combination includes the value of each block and contact parameter; A loss calculation module is configured to, for each of the parameter value combinations, determine a predicted value of the macroscopic mechanical parameter by taking the parameter value combination as input, calculate the deviation between the predicted value of the macroscopic mechanical parameter and a first experimental value of the macroscopic mechanical parameter, and obtain a loss value corresponding to the parameter value combination; the first experimental value is a value of the macroscopic mechanical parameter obtained by performing an experiment on the mechanical behavior of the rock sample under the parameter value combination, the macroscopic mechanical parameter includes rock deformation characteristics and rock strength characteristics, the rock deformation characteristics include Young's modulus and Poisson's ratio, and the rock strength characteristics include uniaxial compressive strength, Brazilian tensile strength, cohesion, and internal friction angle. A parameter determination module is configured to select the parameter value combination with the minimum loss value as the final parameter value combination. Based on the value range of each block and contact parameter, a plurality of parameter value combinations are determined, specifically including: obtaining the value range and step length of each block and contact parameter; for each block and contact parameter, based on the step length, a plurality of values are obtained in the value range, starting from the lower limit of the value range, gradually increasing the value based on the step length until reaching the upper limit of the value range; randomly combining the plurality of values of all the blocks and contact parameters to obtain a plurality of parameter value combinations; The deviation between the predicted value of the macroscopic mechanical parameter and the first experimental value of the macroscopic mechanical parameter is calculated to obtain the loss value corresponding to the parameter value combination, specifically including: calculating the difference between the predicted value and the first experimental value of each macroscopic mechanical parameter, and calculating the sum of the squares of all the differences to obtain the loss value corresponding to the parameter value combination.

6. A computer device comprising: A memory, a processor, and a computer program stored on the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the block discrete element method parameter determination method of any one of claims 1-4.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the block discrete element method parameter determination method of any one of claims 1-4.

Images