PFC heterogeneous mechanical parameter modeling method for soluble rock corrosion simulation, electronic equipment and storage medium

Abstract

The invention relates to the technical field of tunnel engineering, in particular to a PFC heterogeneous mechanical parameter modeling method for soluble rock corrosion simulation, electronic equipment and a storage medium, and the method comprises the following steps: firstly exporting a mechanical parameter code template for controlling bonding among particles from a PFC model; secondly, on the basis of a random field theory, Python is used for generating a normal distribution parameter field conforming to geological statistical characteristics; and finally, through a written interface program, normal distribution parameter field data generated by Python is mapped and imported into a PFC model, and heterogeneous assignment of mechanical parameters of each bonding unit is realized. According to the method, the limitation of a traditional PFC homogeneous model is broken through, and a key technical support is provided for refined and heterogeneous numerical simulation of the soluble rock dissolution process.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering technology, and more specifically, to a PFC heterogeneous mechanical parameter modeling method, electronic device and storage medium for karstification simulation of soluble rocks. Background Technology

[0002] Soluble rock strata are widely distributed in karst-developed areas, and as tunnel engineering extends deeper, the number of engineering scenarios traversing complex karst geological conditions is increasing. Due to the long-term dissolution effect of groundwater, irregular cavities such as fissures and caves form within the rock mass, resulting in strong spatial variability and heterogeneity in its physical and mechanical parameters (such as elastic modulus and strength). This heterogeneous structure dominated by dissolution significantly reduces the stability of the surrounding rock, becoming a key cause of major engineering geological disasters such as tunnel water inrush and collapse. In recent years, investigations and analyses of multiple water inrush accidents during tunnel construction have shown that accurately characterizing the heterogeneous properties of karst rock masses is of paramount importance for accurate disaster prediction and prevention.

[0003] The particle flow discrete element method (PFC) simulates the macroscopic mechanical behavior of rock masses through particle-contact constitutive models and is currently one of the commonly used numerical methods for studying rock mass fracturing processes. However, when using PFC to simulate the karstification process of soluble rocks, how to reasonably characterize the heterogeneity of its mechanical parameters remains a challenge. Existing research methods mainly follow two approaches: One method is the homogeneous parameter weakening method. This method treats the entire soluble rock region as a homogeneous material and simulates the overall strength degradation caused by dissolution by globally reducing the mechanical parameters of its bonding model (such as cohesion and tensile strength). The drawback of this method is that it cannot reflect the spatial heterogeneity of dissolution, ignoring the coexistence of dissolved and undissolved zones (rock mass skeleton) in the actual rock mass. Therefore, this method struggles to accurately simulate the formation and expansion of dominant seepage channels, significantly underestimating the risk of tunnel water inrush disasters and leading to biased engineering decisions.

[0004] Another approach is the geometry-based assignment method. This method first generates the geometry of the dissolution space (such as the distribution of caves and fissures) based on geological survey data or stochastic theory. Then, in the PFC model, lower mechanical parameters are assigned to the particle contacts located inside these geometric shapes, while the external contacts retain their original parameters. Although this method introduces heterogeneity to some extent, its heterogeneous distribution depends entirely on the pre-defined geometric boundaries and cannot simulate the continuous and gradual spatial transition and correlation of parameters during dissolution. In reality, the deterioration of rock mass caused by dissolution is a continuous process from micro to macro, and parameter changes have significant spatial correlations. The "binary" (strong / weak) parameter distribution generated by this method does not conform to this physical fact and cannot truly reflect the complex mechanical response of dissoluted rock masses.

[0005] Ultimately, existing simulation methods suffer from the following shortcomings: First, at the level of physical mechanisms, both the "homogeneous parameter weakening method" and the "geometric morphology-based assignment method" are essentially static, binary, and simplified approximations of the dissolution results. They cannot characterize the continuous, gradual, and statistically correlated heterogeneous evolution of rock mass parameters in space caused by dissolution as a dynamic geological process, which is seriously inconsistent with the actual deterioration mechanism of soluble rocks. Second, at the level of technical implementation, existing methods fail to deeply couple the mature and widely validated random field theory used in geotechnical engineering to describe spatial variability with the micromechanical parameter assignment of PFC. This results in the PFC model lacking a rigorous geostatistical basis for constructing heterogeneity. The spatial distribution of parameters is artificial, arbitrary, or based on simple geometric rules, rather than driven by key statistical features controlling the dissolution process. Thirdly, in terms of simulation performance, due to the lack of mechanisms and technology, existing models cannot realistically reproduce the spontaneous formation and penetration process of dominant seepage channels in dissolution rock masses. As a result, they cannot accurately predict risk indicators of water inrush risk, such as the connectivity of water inrush channels, water inrush flux, and critical conditions for disaster occurrence. This reduces the accuracy of the prediction function of numerical simulation.

[0006] Therefore, there is an urgent need to develop a novel modeling method that can organically integrate geological random field theory with discrete element method (DEM) parameter assignment. By establishing a mapping from macroscopic geostatistical parameters to microscopic DEM parameters, the heterogeneous characteristics of karst rock masses can be accurately characterized, fundamentally improving the accuracy and reliability of tunnel water inrush disaster prediction and early warning, and providing technical support for the safe construction and operation of tunnel engineering in karst areas. Summary of the Invention

[0007] The present invention provides a PFC heterogeneous mechanical parameter modeling method, electronic device and storage medium for soluble rock karstification simulation, which can overcome some or more defects of the prior art.

[0008] A method for modeling PFC heterogeneous mechanical parameters for soluble rock karstification simulation according to the present invention includes the following steps: S1: Exporting parameter templates: After establishing an initial homogeneous numerical model in the PFC software, the built-in script commands are used to export a code template describing the mechanical properties of interparticle contact and bonding. The code template includes parameter definitions for bonding normal stiffness, tangential stiffness, cohesion, and tensile strength. S2: Generation of non-homogeneous random fields: Based on random field theory, the spatial variability of mechanical parameters is quantified using a Python programming environment outside of PFC; S2.1: Determine the statistical characteristics and spatial correlation structure of the target mechanical parameters; S2.2: Use the covariance matrix decomposition method to generate multiple sets of standard normal distribution random fields that conform to the above statistical characteristics and spatial correlation structure; S2.3: Introduce the cocorrelation between different mechanical parameters; S2.4: The standard normal distribution random field is mapped to a physically meaningful normal distribution parameter field through log-normal transformation to characterize the heterogeneous distribution of rock mass parameters under dissolution. S3: Parameter Mapping and Integration Read the normal distribution parameter field data file generated in step S2.4; based on the center spatial coordinates of each contact bonding unit in the PFC model, accurately assign the parameter values ​​at the corresponding positions in the normal distribution parameter field to the unit, replacing the homogeneous parameter values ​​in the code template exported in step S1, thereby generating a new PFC command flow script containing spatially variable mechanical parameters; execute the script to complete the assignment of heterogeneous mechanical parameters to the soluble rock PFC model.

[0009] As a preferred embodiment, in S2.2, the specific method for generating a standard normally distributed random field is as follows: The two-dimensional random field r is generated using the covariance matrix decomposition method, and its expression is:

[0010] in, m x The mean of the parameters. s x Let L be the standard deviation, L be the lower triangular matrix obtained by Cholesky decomposition of the spatial correlation matrix R, and N = [N1, N2, …, N]. n Let ] be a set of independent random variables that satisfy a standard normal distribution; Correlation between any two points i and j in space r ij Calculated by the following formula:

[0011] Where (x, y) are the coordinates of the point. l x and l y These are the relevant lengths in the x and y directions, respectively.

[0012] As a preferred option, in S2.3, specifically: Construct the cocorrelation coefficient matrix and perform Cholesky decomposition to obtain the lower triangular matrix. L corr ;

[0013] The above equation transforms the multiple independent sets of standard normal distributed random fields Z generated in S2.2 into a set of variables with a cocorrelation structure. Z corr .

[0014] As a preferred option, in S2.4, the variable set... Z corr Each parameter field in the model is further subjected to a log-normal distribution to ensure that the parameters are positive, and is then transformed into a log-normal distributed parameter field using the following formula. X :

[0015] in, and Variables X Mean and standard deviation after taking the natural logarithm; Z x represent Z corr The corresponding mechanical parameters are a single standard normal random field.

[0016] As a preferred embodiment, in S3, the normal distribution parameter field data is imported into the PFC model via an interface program, specifically as follows: Read the normal distribution parameter field data file generated by Python, which includes the mechanical parameter values ​​for each spatial location point; based on the position coordinates of particle contact in the PFC model, assign the parameter values ​​at the corresponding positions in the random field to the bonding model of particle contact.

[0017] Preferably, the mechanical parameters include elastic modulus, cohesion, and tensile strength.

[0018] As an option, the model validation step is also included: constructing a benchmark PFC model with homogeneous parameters and comparing it with a model established using a heterogeneous mechanical parameter modeling method; performing dissolution simulation under the same boundary conditions, and verifying the effectiveness of the heterogeneous model in more realistically reflecting the disaster risk of dissolution rock mass by comparing the differences between the two in terms of seepage channel development, water inrush flux and macroscopic mechanical response.

[0019] Preferably, in the model validation step, if the simulation results do not meet expectations, the process returns to S2, and the correlation length of the random field is adjusted. l x and l y， mean m x Standard deviation s x Alternatively, the cocorrelation coefficient matrix can be used to regenerate the random field of mechanical parameters, and iterative optimization can be performed until satisfactory simulation results are obtained.

[0020] The present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described method for modeling PFC heterogeneous mechanical parameters for soluble rock karstification simulation.

[0021] The present invention provides a non-transitory computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor, implements the above-described method for modeling PFC heterogeneous mechanical parameters for soluble rock karstification simulation.

[0022] The beneficial effects of this invention are as follows: 1) This invention is the first to systematically introduce geological random field theory into the PFC micro-parameter assignment process. It generates a parameter field with spatial correlation structure through covariance matrix decomposition and introduces cocorrelation between parameters, overcoming the limitations of traditional homogeneous models and simple geometric assignment methods. This method can realistically reflect the spatial variability and statistical distribution characteristics of parameters caused by dissolution, making the model more consistent with the physical nature of soluble rock dissolution and deterioration.

[0023] 2) By endowing the contact bonding parameters with heterogeneous characteristics that conform to geostatistical laws, this invention can realistically simulate the spontaneous formation and penetration process of weakened zones within the rock mass under dissolution, accurately reproducing the development mechanism of seepage channels. Compared with traditional modeling methods, this invention significantly improves the prediction accuracy of key disaster indicators such as the connectivity of water inrush channels and the flow rate of water inrush.

[0024] 3) This invention constructs a complete technical chain from "PFC template export, Python random field generation, and data mapping integration," realizing the automation and standardization of non-homogeneous parameter assignment. By adjusting the random field parameters (correlation length, statistical characteristics, etc.), the strength and pattern of non-homogeneity can be flexibly controlled, providing convenience for parameter sensitivity analysis and inverse analysis. Attached Figure Description

[0025] Figure 1 This is a flowchart of a PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation provided in the embodiments; Figure 2 This is a schematic diagram illustrating the principle of generating a two-dimensional random field using the covariance matrix decomposition method in the embodiment. Figure 3 This is a schematic diagram illustrating the spatial variation of the surrounding rock parameters in the embodiment; Figure 4 This is a schematic diagram illustrating the spatial variation of the contact bond strength of soluble rock in the PFC software of the embodiment. Detailed Implementation

[0026] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention.

[0027] Example like Figure 1 As shown in the figure, this embodiment provides a PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation, which includes the following steps: S1: Exporting parameter templates: After establishing an initial homogeneous numerical model in the PFC software, the built-in script commands are used to export a code template describing the mechanical properties of interparticle contact and bonding. The code template includes parameter definitions for bonding normal stiffness, tangential stiffness, cohesion, and tensile strength.

[0028] S2: Generation of non-homogeneous random fields: Based on random field theory, the spatial variability of mechanical parameters is quantified using a Python programming environment outside of PFC.

[0029] S2.1: Determine the target mechanical parameters (elastic modulus E, cohesion c, and tensile strength). s t Statistical characteristics (mean) m x Standard deviation s x The spatial correlation structure consists of the correlation function ρ and the correlation lengths λ in the x and y directions. x , λ y definition; In this embodiment, the spatial correlation function adopts an exponential correlation function, and the correlation between any two points i and j in space is... r ij The expression is:

[0030] in, and Let i and j represent the coordinates of two points i and j in space along the x and y directions, respectively. l x and l y The correlation length controls the spatial fluctuation characteristics of the parameters. A smaller correlation length indicates a more fragmented parameter field; a larger correlation length indicates better spatial continuity. The correlation length is adjusted within the range of 1-10 meters based on the degree of dissolution, with smaller values ​​used in strongly dissoluted areas and larger values ​​used in weakly dissoluted areas.

[0031] S2.2: Multiple sets of standard normal distribution random fields conforming to the above statistical characteristics and spatial correlation structure are generated using the covariance matrix decomposition method; a two-dimensional random field r is generated using the covariance matrix decomposition method, with the expression:

[0032] in, m x The mean of the parameters. s x Let L be the standard deviation, L be the lower triangular matrix obtained by Cholesky decomposition of the spatial correlation matrix R, and N = [N1, N2, …, N]. n Let be a set of independent random variables that follow a standard normal distribution.

[0033] Specifically, the implementation process of Cholesky decomposition is as follows: First, construct an n×n spatial correlation matrix R, where each element R ij The matrix R is calculated using the correlation function in step S2.1. Then, the lower triangular matrix L is obtained by Cholesky decomposition of the matrix R. Finally, a standard normal random field with spatial correlation is generated by matrix operation L·N. Figure 2 This paper demonstrates the computational principle of generating random fields using the covariance matrix factorization method. Random field generation and parameter mapping are key steps in achieving heterogeneous characterization.

[0034] S2.3: Introduce the cocorrelation between different mechanical parameters; To reflect the inherent statistical correlation among parameters such as elastic modulus, cohesion, and tensile strength, a cocorrelation coefficient matrix among the parameters is constructed, and the lower triangular matrix L is obtained by Cholesky decomposition of this matrix. corr By transforming Z corr =Z·L corr T transforms the multiple independent sets of standard normal distributed random fields Z generated in S2.2 into a variable set Z with a cocorrelation structure. corr .

[0035] In this embodiment, the cocorrelation coefficient matrix is ​​constructed based on field experimental data combined with literature statistics to ensure that different parameter fields maintain consistent statistical correlation characteristics in spatial distribution. The typical correlation coefficient between elastic modulus and cohesion is 0.6-0.8, the correlation coefficient between elastic modulus and tensile strength is 0.5-0.7, and the correlation coefficient between cohesion and tensile strength is 0.7-0.9.

[0036] S2.4: The standard normal distribution random field is mapped to a physically meaningful normal distribution parameter field through log-normal transformation to characterize the heterogeneous distribution of rock mass parameters under dissolution. The standard normal random field is further adopted using a log-normal distribution to ensure that the parameters are positive, and the standard normal random field Z after cocorrelation processing is expressed by the following formula. x Convert to a log-normal distribution parameter field X:

[0037] in, Z x represent Z corr The corresponding mechanical parameters are a single standard normal random field; and The mean and standard deviation of variable X (including target mechanical parameters: elastic modulus, cohesion, or tensile strength) after taking the natural logarithm, respectively, are compared with the mean of the original space. m X and standard deviation s X There is a definite transformation relationship:

[0038]

[0039] Specifically, Figure 3 An example of the elastic modulus random field (i.e., log-normally distributed parameter field X) generated in this embodiment is shown. This random field clearly shows the spatial variation characteristics of the parameters, which conforms to the actual distribution law of the dissolved rock mass. Through the above processing procedure, three complete two-dimensional parameter random fields are finally obtained, corresponding to the distribution of elastic stiffness, cohesion and tensile strength of limestone under dissolution.

[0040] In this embodiment, the correlation function may also be a Gaussian function or other correlation function forms that conform to geological statistics. The value of the correlation length is adjusted within the range of 0.1-10 meters according to the actual degree of dissolution to meet the simulation needs of different engineering scenarios.

[0041] The random field generated in this embodiment has the following technical features: 1) Spatial variability: The parameters exhibit a continuous but uneven distribution in space, which can truly reflect the deterioration of the rock mass caused by dissolution. 2) Statistical controllability: By adjusting parameters such as mean, standard deviation, and correlation length, the statistical characteristics of random fields can be precisely controlled; 3) Parameter cocorrelation: There is a reasonable statistical correlation between different mechanical parameters, which conforms to the inherent laws of rock mass mechanical behavior; 4) Physical rationality: All parameter values ​​are positive and follow a reasonable probability distribution.

[0042] After applying the generated random field to the PFC model, the following can be obtained: Figure 4 The spatial distribution of contact bonding strength of soluble rock shown effectively simulates the heterogeneous characteristics of rock mass caused by dissolution, providing a reliable technical means for accurately predicting water inrush disasters.

[0043] S3: Parameter Mapping and Integration Read the normal distribution parameter field (data file) generated in step S2.4; based on the center spatial coordinates of each contact bonding unit in the PFC model, accurately assign the parameter values ​​at the corresponding positions in the normal distribution parameter field to the unit, replacing the homogeneous parameter values ​​in the code template exported in step S1, thereby generating a new PFC command flow script containing spatially variable mechanical parameters; execute the script to complete the assignment of heterogeneous mechanical parameters to the soluble rock PFC model.

[0044] In S3, the normal distribution parameter field data is imported into the PFC model through an interface program, specifically as follows: Read the normal distribution parameter field data file generated by Python, which includes the mechanical parameter values ​​for each spatial location point; based on the position coordinates of particle contact in the PFC model, assign the corresponding parameter values ​​in the random field to the bonding model of particle contact; the mechanical parameters include elastic modulus, cohesion, and tensile strength.

[0045] Specifically, S3 combined Figure 4 The spatial variation of the contact bond strength of soluble rock was implemented in the PFC software shown: S3.1: PFC Contact Information Extraction and Coordinate Matching. First, the geometric information and mechanical parameters of contact bonding in all soluble rock regions are extracted from the PFC model, establishing the correspondence between contact elements and their spatial locations. The specific code is as follows (PFC code for exporting bonding mechanical parameters): ; In this embodiment, the spatial coordinates (x, y) of the center point of each contact bonding unit are obtained through the contact query command built into PFC. The center point coordinates are calculated by connecting the center coordinates of the circles connecting the two contact particles. Simultaneously, key information such as the ID number, contact type, and current mechanical parameter values ​​of each contact unit are recorded to establish a basic database for subsequent parameter mapping.

[0046] S3.2: Random Field Data Interface Program Design. Develop a dedicated data interface program to achieve seamless integration between random field data generated by Python and the PFC model. The interface program has the following functions: The interface program is written in Python and its main functions include: reading random field data files, parsing PFC contact information, establishing coordinate mapping relationships, and performing parameter assignment operations. The program structure is divided into three modules: the data input module is responsible for reading random field text files; the data processing module performs coordinate interpolation calculations; and the data output module generates executable command stream scripts for PFC.

[0047] Specifically, the coordinate mapping algorithm uses the nearest neighbor interpolation method. For the center coordinates (x, y, z) of each PFC contact unit, ... i ,y i The interpolation process involves finding the nearest grid point in the random field and assigning the parameter values ​​of that grid point to the corresponding contact element. The interpolation formula is as follows:

[0048] Where (xj,yj) are the coordinates of the grid point closest to (xi,yi) in the random field, P random P represents the parameter values ​​of a random field. PFC The parameter values ​​to assign to the PFC contact.

[0049] The Python code for generating random fields is as follows: ; S3.3: Parameter Assignment and Model Update. This section precisely assigns random field parameter values ​​to the corresponding contact units in the PFC model, completing the assignment of heterogeneous mechanical parameters. The specific Python code for importing random fields into PFC is as follows: ; In this embodiment, the parameter assignment process maintains the integrity of the contact bonding model, ensuring mechanical consistency among the elastic modulus, strength parameters, and deformation parameters. For the parallel bonding model, multiple parameters such as normal stiffness, tangential stiffness, tensile strength, and cohesion need to be updated simultaneously.

[0050] The accuracy of parameter assignments is verified by comparing the statistical distribution of the original random field data with that of the actual parameter values ​​in the PFC model; visualization tools are used to check whether the spatial distribution of the parameters matches the expectations, such as... Figure 4 The distribution of contact bond strength shown should exhibit a reasonable spatial variation pattern.

[0051] Specifically, if errors are found in parameter assignment or model instability during the verification process, these can be corrected by adjusting the random field grid resolution, optimizing the interpolation algorithm, or modifying the parameter limits. This ensures that the final heterogeneous PFC model conforms to both geostatistical laws and numerical computation stability requirements.

[0052] In this embodiment, the export of the PFC bonding parameter template in S1 specifically includes: establishing a discrete element numerical model of the soluble rock region, which consists of particle units and a parallel bonding model that defines their contact mechanical behavior; and systematically exporting the code templates of all contact bonding in the model using the built-in script commands of PFC. The mechanical parameters include key microscopic parameters such as elastic modulus, cohesion, and tensile strength, providing a structured interface for subsequent replacement of heterogeneous parameters.

[0053] In this embodiment, the Python random field generation in S2 is implemented using the covariance matrix decomposition method. Its core principle is to generate a spatially correlated random field with specified statistical characteristics by decomposing the spatial correlation matrix using Cholesky method. During the random field generation process, the spatial autocorrelation characteristics of the parameters and the cocorrelation structure between different mechanical parameters are considered simultaneously to ensure that the generated heterogeneous parameter field conforms to the geological statistical laws of actual karst rock masses.

[0054] In this embodiment, the parameter mapping of S3 is implemented by writing a dedicated data interface program. This program reads the random field data file generated by Python, which contains all contact bonding units in the PFC model. It performs interpolation matching in the random field according to the spatial coordinates of their centers, accurately assigns the obtained parameter values ​​to the corresponding contact units, and finally generates and executes a new PFC command stream script containing spatial variation parameters to complete the entire heterogeneous assignment process.

[0055] This embodiment includes a model validation step: constructing a baseline PFC model with homogeneous parameters and comparing it with a model established using a heterogeneous mechanical parameter modeling method; performing dissolution simulation under the same boundary conditions (such as seepage-stress coupling conditions), and verifying the effectiveness of the heterogeneous model in more realistically reflecting the disaster risk of dissolution rock masses by comparing the differences between the two in terms of seepage channel development, water inrush flux, and macroscopic mechanical response. If the simulation results do not meet expectations, return to S2 and adjust the correlation length of the random field. l x and l y， mean m x Standard deviation s x Alternatively, the cocorrelation coefficient matrix can be used to regenerate the random field of mechanical parameters, and iterative optimization can be performed until satisfactory simulation results are obtained.

[0056] The heterogeneous mechanical parameter model constructed by the above three-step method in this embodiment can effectively simulate the spatial variability of rock mass parameters caused by dissolution, overcome the limitations of traditional homogeneous models, and provide technical support for the accurate prediction of soluble rock erosion disasters.

[0057] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation.

[0058] This embodiment provides a non-transitory computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for modeling PFC heterogeneous mechanical parameters for soluble rock karstification simulation.

[0059] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for modeling PFC heterogeneous mechanical parameters for soluble rock karstification simulation, characterized in that: Includes the following steps: S1: Exporting parameter templates: After establishing an initial homogeneous numerical model in the PFC software, the built-in script commands are used to export a code template describing the mechanical properties of interparticle contact and bonding. The code template includes parameter definitions for bonding normal stiffness, tangential stiffness, cohesion, and tensile strength. S2: Generation of non-homogeneous random fields: Based on random field theory, the spatial variability of mechanical parameters is quantified using a Python programming environment outside of PFC; S2.1: Determine the statistical characteristics and spatial correlation structure of the target mechanical parameters; S2.2: Use the covariance matrix decomposition method to generate multiple sets of standard normal distribution random fields that conform to the above statistical characteristics and spatial correlation structure; S2.3: Introduce the cocorrelation between different mechanical parameters; S2.4: The standard normal distribution random field is mapped to a physically meaningful normal distribution parameter field through log-normal transformation to characterize the heterogeneous distribution of rock mass parameters under dissolution. S3: Parameter Mapping and Integration Read the normal distribution parameter field data file generated in step S2.4; based on the center spatial coordinates of each contact bonding unit in the PFC model, accurately assign the parameter values ​​at the corresponding positions in the normal distribution parameter field to the unit, replacing the homogeneous parameter values ​​in the code template exported in step S1, thereby generating a new PFC command flow script containing spatially variable mechanical parameters; execute the script to complete the assignment of heterogeneous mechanical parameters to the soluble rock PFC model.

2. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 1, characterized in that: In S2.2, the specific method for generating a standard normally distributed random field is as follows: The two-dimensional random field r is generated using the covariance matrix decomposition method, and its expression is: ； in, μ x The mean of the parameters. σ x Let L be the standard deviation, L be the lower triangular matrix obtained by Cholesky decomposition of the spatial correlation matrix R, and N = [N1, N2, …, N]. n Let ] be a set of independent random variables that satisfy a standard normal distribution; Correlation between any two points i and j in space ρ ij Calculated by the following formula: ； Where (x, y) are the coordinates of the point. λ x and λ y These are the relevant lengths in the x and y directions, respectively.

3. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 2, characterized in that: In S2.3, specifically: Construct the cocorrelation coefficient matrix and perform Cholesky decomposition to obtain the lower triangular matrix. L corr ; ； The above equation transforms the multiple independent sets of standard normal distributed random fields Z generated in S2.2 into a set of variables with a cocorrelation structure. Z corr .

4. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 3, characterized in that: In S2.4, the variable set Z corr Each parameter field in the model is further subjected to a log-normal distribution to ensure that the parameters are positive, and is then transformed into a log-normal distributed parameter field using the following formula. X : ； in, and Variables X Mean and standard deviation after taking the natural logarithm; Z x represent Z corr The corresponding mechanical parameters are a single standard normal random field.

5. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 4, characterized in that: In S3, the normal distribution parameter field data is imported into the PFC model through an interface program, specifically as follows: Read the normal distribution parameter field data file generated by Python, which includes the mechanical parameter values ​​for each spatial location point; based on the position coordinates of particle contact in the PFC model, assign the parameter values ​​at the corresponding positions in the random field to the bonding model of particle contact.

6. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 5, characterized in that: Mechanical parameters include elastic modulus, cohesion, and tensile strength.

7. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 6, characterized in that: It also includes model validation steps: constructing a benchmark PFC model with homogeneous parameters and comparing it with a model established using a heterogeneous mechanical parameter modeling method; conducting dissolution simulations under the same boundary conditions, and verifying the effectiveness of the heterogeneous model in more realistically reflecting the disaster risk of dissolution rock masses by comparing the differences between the two in terms of seepage channel development, water inrush flux and macroscopic mechanical response.

8. The PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation according to claim 7, characterized in that: In the model validation step, if the simulation results do not meet expectations, return to S2 and adjust the correlation length of the random field. λ x and λ y， mean μ x Standard deviation σ x Alternatively, the cocorrelation coefficient matrix can be used to regenerate the random field of mechanical parameters, and iterative optimization can be performed until satisfactory simulation results are obtained.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements a PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation as described in any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements a PFC heterogeneous mechanical parameter modeling method for soluble rock karstification simulation as described in any one of claims 1 to 8.

Images