CN114117879A

CN114117879A - Microscopic parameter calibration method for sandy gravel soil three-axis shearing discrete element model

Info

Publication number: CN114117879A
Application number: CN202111462082.8A
Authority: CN
Inventors: 魏英杰; 李鹏飞; 李刚; 夏俊卫; 王帆; 陶琦; 罗振平
Original assignee: Beijing University of Technology
Current assignee: Beijing University of Technology; China Railway 19th Bureau Group Co Ltd; Rail Transit Engineering Co Ltd of China Railway 19th Bureau Group Co Ltd
Priority date: 2021-12-02
Filing date: 2021-12-02
Publication date: 2022-03-01

Abstract

The invention belongs to the technical field of microscopic parameter calibration of discrete element models, and discloses a microscopic parameter calibration method of a sandy cobble soil triaxial shearing discrete element model, which comprises the steps of establishing a sandy cobble triaxial shearing numerical model through PFC2D/3D software, selecting a contact model, and determining calibrated macroscopic and microscopic parameters and ranges thereof; designing a numerical test scheme by utilizing an orthogonal test; performing numerical tests according to a test scheme by using PFC2D/3D software; analyzing the significance of each microscopic parameter on the influence of macroscopic mechanical parameters by adopting multi-factor variance, and screening out main microscopic parameters influencing macroscopic mechanical indexes; establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis; and converting the calibration problem into a multi-objective function mathematical programming problem to solve, and comparing and verifying the numerical test result with the physical test result of typical cobble soil. The invention solves the problems of blindness, time consumption and labor consumption in the calibration process of a trial and error method.

Description

Microscopic parameter calibration method for sandy gravel soil three-axis shearing discrete element model

Technical Field

The invention belongs to the technical field of microscopic parameter calibration of discrete element models, and particularly relates to a method for calibrating microscopic parameters of a PFC3D model by three-axis shearing of sandy gravel soil.

Background

At present, sandy gravel strata are widely distributed in Chinese Chengdu, Beijing, Shenyang, Lanzhou and the like. Sand-gravel soil is a general term for discrete debris deposits which are mainly pebbles and gravels and contain a small amount of sand and cohesive soil. The sand-gravel soil has the characteristics of high gravel content, large particle size of single gravel, high compressive strength, low soil body cohesive force, large porosity, strong water permeability and the like, and belongs to a typical mechanical unstable layer. Therefore, many engineering problems occur in the process of constructing underground engineering in the sandy gravel stratum, such as surface subsidence, stratum cavities, sand gushing and water bursting on the tunnel face and the like. Therefore, the research on the physical and mechanical properties of the sand-gravel soil has great engineering significance.

The numerical test has the advantages of low cost, easy repetition and the like and is usually used as a good supplement for the physical test. With the continuous development of Particle Flow (PFC) numerical simulation technology in recent years, the PFC numerical simulation technology can be widely applied to geotechnical mechanical numerical tests because the PFC numerical simulation technology can conveniently treat the mechanical problem of the discontinuous medium. When the numerical simulation research of the particle discrete elements is carried out on the mechanical properties of the sand and pebbles, the microscopic parameters of the particle flow numerical model are calibrated according to the macroscopic mechanical response. Scholars at home and abroad make a great deal of research on the specification of the mesoscopic parameters. The Potyondy et al study found that the maximum to minimum particle size ratio R was found without regard to the particle size distribution_max/R_minThe generated model is more uniform and accords with the actual physical and mechanical properties of the rock-soil material when the model is 1.66. Zhao nationality's research finds that the particle diameter ratio (R) is not considered in the case of particle size distribution_max/R_min) When the range is 1-5, the mechanical property of the model is hardly influenced. Metaphors or the like for the relationship between the model size and the particle sizeThe study found that when the minimum scale of the model RES ═ L/R_min)[1/(1+R_max/R_min)]And when the particle size is more than or equal to 10, the size and the quantity of the particles have small influence on the macroscopic mechanical parameters of the model. Abel et al found that when the porosity was in the range of 0.04 to 0.32, the initial modulus of elasticity, peak stress, cohesion and internal friction angle were not greatly affected. A triaxial numerical test model is established in Weilonghai, and the sand-gravel soil mesoscopic parameters with different compactness are calibrated. Gunn et al introduced a contact bonding model using PFC3D as a tool and analyzed the particle friction coefficient, bonding strength and the effect of particle clusters on the stress-strain curve of coarse-grained soil. The Japeng carries out numerical tests through PFC3D based on indoor test results, and reproduces stress-strain curves of sand and gravel soil with different water contents, densities and coarse-grain contents. Most of the above studies are based on trial and error methods, which are very time consuming.

The scholars develop a great deal of research aiming at the problem that the trial-and-error method is time-consuming and labor-consuming, and certain improvement is carried out. The Hookou adopts an orthogonal test design method to research a mesoscopic parameter calibration method of a parallel bond model, and establishes a relation between macroscopic mechanical parameters and mesoscopic parameters by using regression analysis. Wangjin Wei and the like adopt a method of combining an orthogonal test and an isoline method, and quickly determine the mesoscopic parameters of the rockfill material on the basis of optimizing the value range of the mesoscopic parameters and analyzing the sensitivity of macroscopic response to the mesoscopic parameters. Zhangbao jade and the like research the correlation between the microscopic parameters and the macroscopic parameters of the parallel joint model by adopting an orthogonal numerical test method and provide a rock microscopic parameter calibration process. Yoon and the like establish a linear relation between a macroscopic mechanical index and a microscopic parameter by using PB, examine the interaction between the parameters with obvious influence by using a response surface method, establish a nonlinear relation between the macroscopic mechanical index and the microscopic parameter, and finally convert the problem into a nonlinear multi-objective mathematical programming problem. The research is mostly focused on the problem of calibrating the rock microscopic parameters.

According to the research, the sand-gravel soil microscopic parameter calibration at present mostly adopts a trial-and-error method, and the system deep research is lacked. Therefore, a new method for calibrating the meso-scale parameters of the PFC3D model by three-axis shearing of sandy gravel soil is needed to be designed to overcome the defects in the prior art.

Through the above analysis, the problems and defects of the prior art are as follows: the existing sand-gravel soil microscopic parameter calibration research is mostly based on a trial and error method, the process is time-consuming, and deep system research is lacked.

The difficulty in solving the above problems and defects is:

(1) at present, the value range of macroscopic mechanical parameters of typical sand and pebble soil is lacked;

(2) at present, contact models adopted for typical sandy gravel soil are not uniform, and the value range of microscopic parameters is not clear;

(3) most of the commonly adopted trial-and-error methods are 'parameter collection', which results in that the research on the influence of each microscopic parameter on the macroscopic parameter is not systematic and the understanding is not clear;

(4) calibrating a plurality of macro and meso parameters, and interacting influences exist among the parameters, so that the difficulty in clarifying the relationship between the meso parameters and the macro parameters is high;

(5) at present, numerical tests have unclear numerical test parameter values, and the test process consumes a lot of time;

(6) at present, a calibration method is not available, and the microscopic parameters corresponding to the macroscopic parameters are quickly given. The significance of solving the problems and the defects is as follows: the method of the invention provides a value range of macroscopic mechanical parameters of typical sand and pebble soil based on a large amount of literature research; based on a large number of numerical tests, a microscopic contact model suitable for sandy gravel soil and a value range of microscopic parameters are determined, and reasonable reference values are given to numerical test parameters; the influence of the microscopic parameters on the macroscopic parameters is scientifically analyzed by adopting a mathematical statistical method of orthogonal test and variance analysis; and finally, converting the calibration problem into a multi-objective function optimization problem, and providing a calibration method. The method solves the blindness existing in the calibration process of the trial and error method, solves the problems of time and labor consumption, enables the complicated calibration problem to be converted into the simple mathematical problem, and has important significance for the popularization and the use of discrete elements.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a method for calibrating mesoscopic parameters of a three-axis shearing discrete element model of sandy gravel soil.

The invention is realized in such a way that a sand and gravel soil triaxial shearing discrete element model mesoscopic parameter calibration method comprises the following steps:

establishing a sand-pebble triaxial shear numerical model through PFC2D/3D software, selecting a contact model, and determining calibrated macroscopic and microscopic parameters and ranges thereof;

designing a numerical test scheme by using an orthogonal test design method;

thirdly, performing a numerical test by using PFC2D/3D software according to a test scheme;

analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters by adopting a multi-factor variance analysis method, and screening out main microscopic parameters influencing the macroscopic mechanical indexes;

establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis;

and step six, converting the calibration problem into a multi-objective function mathematical programming problem to solve, and comparing and verifying a numerical test result with a physical test result of typical sand and pebble soil.

Further, in the microscopic parameter calibration method for the sandy gravel soil triaxial shearing discrete element model, the microscopic parameters required to be input by the linear contact bonding model comprise a linear group and a contact bonding group; wherein the linear group comprises the contact effective modulus E_cNormal tangential stiffness ratio k_n/k_sAnd coefficient of friction μ; the contact bonding group comprises bonding control gaps and bonding tensile strength sigma_cAnd bond shear strength τ_c。

Further, in step three, the designing of the numerical test scheme by using the orthogonal test design method includes using PFC3D to implement sample preparation, confining pressure application, and loading processes by writing a triaxial shear program:

(1) preparing a sample: the numerical test does not consider the influence of particle grading temporarily, and the size H multiplied by L of the model is 600mm multiplied by 300mm according to geotechnical test standard; the particles adopt two forms of spheres and particle clusters, and L/R is taken_min＝30，R_max/R_min1.66, porosity n 0.32, particle density 2700kg/m³(ii) a The whole test vessel was uniformly filled with particles by the expansion method, and the number of particles generated was 20000.

(2) Applying confining pressure: in the numerical test, three groups of confining pressures of 100kPa, 200kPa and 300kPa are respectively applied to each group of samples; the constant confining pressure is applied by monitoring the difference between the contact pressure between the sidewall and the particles and the set confining pressure in each calculation step, and adjusting the displacement of the sidewall.

(3) Vertical loading: the loading process is realized by endowing the upper loading plate and the lower loading plate with certain speed, the loading speed cannot be too high, the test sample is kept in a quasi-static state, otherwise, the accuracy of the test result is influenced, and the mechanical property of the sandy gravel soil is related to the loading speed; the loading rate was 1.2mm/s and the rate of unbalanced forces in the sample during shear was found to be less than 1X 10 by monitoring^-4And the variation amplitude of the confining pressure is not more than 0.3 percent.

(4) And (4) terminating the loading: when the peak value exists, when the axial strain of the test progress value reaches 3-5% of the peak value, the loading is stopped; if no peak exists, the loading is stopped when the axial strain reaches 15 percent.

Wherein, the orthogonal test principle comprises:

the orthogonal table is designed according to the orthogonal principle, is a normalized table, is a cost-saving tool for arranging tests and analyzing test results in the orthogonal design, and uses L_n(r^m) Represents; wherein L is the code number of the orthogonal table; n is the number of rows of the orthogonal table, namely the test times; r is a factor level number; m is the number of columns of the orthogonal table, i.e. the number of factors that can be arranged at most. In the orthogonal table, each level appears in any column with the same number of occurrences; all possible combinations of various levels between any two classes occur and occur equally often.

Further, in step four, the analysis of variance includes:

the analysis of variance is used for estimating the size of errors and the influence degree of each factor on the test result, and is used for data analysis of the orthogonal test. The analysis of variance analyzes influence factors influencing the fluctuation of the total variance by decomposing the variance, and the influence factors are divided into intra-group errors and inter-group errors according to the sources of the influence factors. Constructing F by calculating the ratio of the inter-mean square to the intra-mean square of factor A_AStatistics, for a given significance level a, find a critical value F_αIf F is_A>F_αThen factor A is considered to have a significant effect on the test results, and F_ARelative to F_αThe larger the factor, the more significantly the effect of the factor on the test results.

By using L_n(r^m) Orthogonal table design test with the test result of y_i(i ═ 1, 2, … n). Analysis of variance is as follows:

(1) calculating the sum of squares of total deviations

Wherein,

(2) calculating the sum of squared deviations due to factor j

Wherein,

wherein, K_iIs the sum of the test results with the j-th column with the horizontal number i.

(3) Calculating the sum of squared deviations of the errors

SS_e＝∑SS_{Empty column}；

(4) Degree of freedom of calculation

SS_TDegree of freedom of (2):

df_T＝n-1；

SS_jdegree of freedom of (2):

df_j＝r-1；

SS_efree of (2):

df_e＝∑df_{empty column}；

(5) Computing mean square

(6) Calculating the F value

(7) Significance test

For a given significance level α, a threshold value F is found_αIf F is_j>F_αThen the j factor is considered to have a significant effect on the test results, and F_ARelative to F_αThe larger the factor, the more significantly the effect of the factor on the test results.

Wherein, the selection of the test indexes and the factor levels comprises the following steps:

(1) selection of test indexes

The calibration of the three-axis shearing microscopic parameters of the particle flow is generally based on the stress-strain relationship and the damage envelope of the PFC model, which are consistent with the results of the indoor test. Therefore, the macro characteristic index selection follows the following principle: firstly, the selected index can be conveniently and easily obtained from the test; secondly, the selected index can better describe the form of the calibrated curve. Selecting an initialModulus of elasticity E₀Peak intensity σ_fCohesion c and internal friction angle

And taking four parameters as macro indexes. Meanwhile, the initial elastic modulus E and the peak strength of the sandy gravel soil are closely related to the magnitude of the confining pressure₀Peak intensity σ_fAll are obtained under the condition that the ambient pressure is 300 kPa.

The rule of the influence of each microscopic parameter of the linear contact bonding model on the macroscopic parameter is as follows: the macroscopic elastic modulus is linearly related to the contact effective modulus; the poisson's ratio increases with increasing normal tangential stiffness ratio; the tangential and normal bond strengths and friction coefficients between particles are mainly related to the peak strength; the content of the particle clusters can effectively improve the shear strength. And selecting the effective contact modulus, the normal tangential stiffness ratio, the inter-particle friction coefficient, the bonding shear strength, the bonding tensile strength and the content of the particle cluster as the microscopic parameter indexes of the numerical test based on the conclusion.

(2) Determination of levels of factors in orthogonal experiments

6 microscopic parameters and 4 macroscopic parameters are selected as calibration indexes, and each factor is selected to be four levels. Through numerical tests, the value ranges of the macroscopic parameters under different combinations of the selected factor levels comprise:

initial modulus of elasticity E₀: 10MPa to 100MPa, and the confining pressure is 300 kPa; peak intensity σ_f: 500kPa to 2050kPa, and the confining pressure is 300 kPa; cohesive force c: 0 to 600 kPa; internal friction angle

17°～47°。

The factor level is selected from the following steps:

factor level 1: e_c＝20MPa，k_n/k_s＝1，μ＝0.25，σ_c＝0.25MPa，τ_c＝0.25MPa，clump＝25％；

Factor level 2: e_c＝80MPa，k_n/k_s＝2，μ＝0.50，σ_c＝0.50MPa，τ_c＝0.50MPa，clump＝50％；

Factor level 3: e_c＝140MPa，k_n/k_s＝3，μ＝0.75，σ_c＝0.75MPa，τ_c＝0.75MPa，clump＝75％；

Factor level 4: e_c＝200MPa，k_n/k_s＝4，μ＝1.00，σ_c＝1.00MPa，τ_c＝1.00MPa，clump＝100％。

Under the selected factor level, the obtained macroscopic mechanical index range covers the macroscopic force index range of most of the sand-gravel soil.

Further, in step five, the regression analysis includes:

according to the result of the variance analysis, carrying out regression analysis on the significant influence factors and the macroscopic index to obtain the corresponding relation of the macro-microscopic parameters:

E₀＝0.142E_c-5.4k_n/k_s+9.83τ_c+15.87clump+16.41(R²＝0.68)；

σ_f＝-2.47E_c+985.27μ+631.81τ_c+653.08clump-5.71(R²＝0.903)；

c＝-1.33E_c+28.08k_n/k_s+163.68σ_c+253.86τ_c-66.61(R²＝0.828)；

the initial modulus of elasticity increases with increasing contact modulus, bond shear strength and content of particle clusters, and decreases with increasing normal to tangential stiffness ratio; the peak strength is in positive correlation with the friction coefficient, the contact shear strength and the contact tensile strength, and is in negative correlation with the contact modulus; the cohesive force is increased along with the increase of the normal tangential stiffness ratio, the cohesive tensile strength and the cohesive shear strength, and is reduced along with the increase of the contact modulus; the internal friction angle is in positive correlation with the contact modulus, the friction coefficient and the content of the particle clusters.

Further, in the sixth step, the step of converting the calibration problem into a multi-objective function mathematical programming problem to solve and comparing and verifying the problem with the physical test result of typical sand and pebble soil includes optimizing each microscopic parameter of a numerical test by using a multi-objective mathematical programming method based on a regression equation, and includes:

(1) objective function

In order to make the physics obtained by numerical and laboratory tests as close as possible, the objective function is:

wherein E is₀、σ_fC and

regression equations obtained by regression respectively; e₀ ^*、σ_f ^*、c^*And

respectively, macroscopic parameters obtained by indoor tests.

(2) Constraint conditions

Since the regression equation is obtained at a certain factor level, the value range of the mesoscopic parameter needs to be constrained, and the constraint conditions are as follows:

(3) optimization of mesoscopic parameters

For the problem of optimization of the multi-objective function, solving by adopting an fgoalatain function in MATLAB; converting the target planning problem into a minimum solving problem by setting the target matrix good to 0; by giving an initial value x₀And finally optimizing the weight value and the constraint condition to obtain the microscopic parameter.

Another object of the present invention is to provide a system for calibrating mesoscopic parameters of a sandy gravel soil triaxial shearing discrete element model, which applies the method for calibrating mesoscopic parameters of a sandy gravel soil triaxial shearing discrete element model, wherein the system for calibrating mesoscopic parameters of a sandy gravel soil triaxial shearing discrete element model comprises:

the numerical model establishing module is used for establishing a sand-pebble triaxial shearing numerical model through PFC2D/3D software, selecting a contact model and determining calibrated macroscopic and microscopic parameters and ranges thereof;

the test scheme design module is used for designing a numerical test scheme by using an orthogonal test design method;

the numerical test module performs numerical test according to a test scheme by using PFC2D/3D software;

the significance analysis module is used for analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters by adopting a multi-factor variance analysis method and screening out main microscopic parameters influencing the macroscopic mechanical indexes;

the linear relation establishing module is used for establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis;

and the solving and comparison verification module is used for converting the calibration problem into a multi-target function mathematical programming problem to solve and compare and verify the problem with the physical test result of typical sand and pebble soil.

It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:

establishing a sand-pebble triaxial shear numerical model through PFC2D/3D software, selecting a contact model, and determining calibrated macroscopic and microscopic parameters and ranges thereof; designing a numerical test scheme by using an orthogonal test design method; the method is used for designing a numerical test scheme by using an orthogonal test design method; analyzing the test result by using a multi-factor variance analysis method, analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters, and screening out the main microscopic parameters influencing the macroscopic mechanical indexes; establishing a linear expression between a macroscopic mechanical index and a main microscopic parameter by using a regression analysis method; and finally, converting the parameter calibration problem into a multi-target mathematical programming problem to solve, and comparing the problem with a typical sandy gravel soil triaxial shear test to find that a numerical test result is basically consistent with a physical test result and verify the feasibility of the calibration method.

It is another object of the present invention to provide a computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:

The invention also aims to provide an information data processing terminal, which is used for realizing the microscopical parameter calibration system of the sandy gravel soil triaxial shearing discrete meta-model.

By combining all the technical schemes, the invention has the advantages and positive effects that: according to the mesoscopic parameter calibration method for the three-axis shearing discrete element model of the sandy cobble soil, provided by the invention, the problem of mesoscopic parameter calibration of the three-axis shearing particle flow model of the sandy cobble soil is researched, and firstly, the value range of macroscopic mechanical parameters of typical sandy cobble soil is given based on a large amount of literature research; based on a large number of numerical tests, a microscopic contact model suitable for sandy gravel soil and a value range of microscopic parameters are defined; reasonable reference values are given to the numerical test parameters; then, designing a numerical test scheme by using an orthogonal test design method; then, analyzing the test result by using a multi-factor variance analysis method, analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters, and screening out the main microscopic parameters influencing the macroscopic mechanical indexes; then, establishing a linear expression between the macroscopic mechanical index and the main microscopic parameter by using a regression analysis method; and finally, converting the parameter calibration problem into a multi-target mathematical programming problem to solve. The invention is compared with a typical sandy gravel soil triaxial shear test to find that the numerical test result and the physical test result are basically consistent, and the feasibility of the calibration method is verified. The method solves the blindness existing in the calibration process of the trial-and-error method, solves the problems of time and labor consumption, and enables the complicated calibration problem to be converted into the simple mathematical problem, provides an efficient calibration method, and has important significance for the popularization and the use of discrete elements.

Firstly, designing a numerical test scheme by utilizing orthogonal test design, and then carrying out significance analysis on the influence of each microscopic parameter on a macroscopic parameter by adopting a multi-factor variance analysis method; then establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis; and finally, converting the calibration problem into a multi-objective function mathematical programming problem to solve, and comparing and verifying the problem with the physical test result of typical sand and pebble soil. According to the sand-gravel-soil discrete element model microscopic-parameter calibration method, a sand-gravel-soil triaxial shear numerical model is established through PFC3D, a test numerical test scheme is designed by using an orthogonal test method, and the influence of 6 microscopic parameters of a contact bonding model on the macroscopic mechanical characteristics of the sand-gravel soil is analyzed by using multi-factor variance analysis, so that the sand-gravel-soil discrete element model microscopic-parameter calibration method is provided.

According to the invention, a sand-gravel-soil triaxial shear numerical model is established through PFC3D, a test numerical test scheme design is carried out by using an orthogonal test method, the influence of 6 mesoscopic parameters of the contact bonding model on the macroscopic mechanical characteristics of the sand-gravel soil is analyzed by using multi-factor variance analysis, and a method for calibrating the mesoscopic parameters of the contact bonding model is provided. The main conclusions are:

(1) the contact bonding model can increase the bonding among the particles to simulate the occlusion effect among the pebble particles, better reappear the result of the indoor test, and can be used for simulating the mechanical properties of the sandy gravel soil; meanwhile, by adding the particle clusters and combining the contact bonding model, the microscopic parameter calibration of the sand-gravel soil can be realized more efficiently.

(2) Through orthogonal test design and multi-factor variance analysis, the sensitivity of the macro-mechanical index of the sandy gravel soil to microscopic parameters is analyzed. The research finds that the initial elastic modulus increases along with the increase of the contact modulus, the bonding shear strength and the content of the particle cluster, and decreases along with the increase of the normal tangential stiffness ratio; the peak strength is in a positive correlation with the coefficient of friction, contact shear strength and contact tensile strength, and in a negative correlation with the contact modulus. The cohesive force is increased along with the increase of the normal tangential stiffness ratio, the cohesive tensile strength and the cohesive shear strength, and is reduced along with the increase of the contact modulus; the internal friction angle is in positive correlation with the contact modulus, the friction coefficient and the content of the particle clusters.

(3) Based on the result of the multi-factor variance analysis, a linear expression between the macro-mechanical index and the main microscopic parameters of the sandy gravel soil is further established through simple linear regression analysis.

(4) And converting the parameter calibration problem into a multi-objective function optimization problem according to the obtained expression among the macro and micro parameters, and finally solving to obtain the corresponding micro parameters. And finally obtaining the microscopic parameters which are consistent with the indoor test by micro-adjusting the microscopic parameters. The feasibility of the microscopic parameter calibration method is verified.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a flow chart of a mesoscopic parameter calibration method for a three-axis shearing discrete element model of sandy gravel soil according to an embodiment of the present invention.

FIG. 2 is a structural block diagram of a mesoscopic parameter calibration system of a three-axis shearing discrete element model of sandy gravel soil provided by the embodiment of the invention;

in the figure: 1. a test scheme design module; 2. a significance analysis module; 3. a linear relationship establishing module; 4. and comparing the verification module.

Fig. 3 is a schematic view illustrating a constitutive relation of a contact force and a relative displacement relationship according to an embodiment of the present invention.

FIG. 3(a) is a schematic diagram of the normal component of the contact force provided by an embodiment of the present invention.

FIG. 3(b) is a schematic diagram of the tangential component of the contact force provided by an embodiment of the present invention.

Fig. 4 is a schematic diagram of a numerical triaxial test model according to an embodiment of the present invention.

Fig. 4(a) is a schematic diagram of a three-dimensional particle flow numerical model according to an embodiment of the present invention.

Fig. 4(b) is a schematic diagram of a particle cluster used in a sampling process according to an embodiment of the present invention.

Fig. 5 is a stress-strain curve diagram of a conventional triaxial test of sandy gravel soil provided by an embodiment of the invention.

FIG. 6 is a Moire coulomb intensity envelope plot provided by an embodiment of the present invention.

FIG. 7 is a drawing of E provided by an embodiment of the present invention₀Is generated from the multi-factor analysis of variance histogram.

FIG. 8 is a graph of σ according to an embodiment of the present invention_fIs generated from the multi-factor analysis of variance histogram.

FIG. 9 is a multi-factor ANOVA histogram of c provided by an embodiment of the present invention.

FIG. 10 is a schematic diagram of an embodiment of the present invention

Is generated from the multi-factor analysis of variance histogram.

Fig. 11 is a triaxial shear test stress-strain curve diagram of sandy gravel soil provided by the embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Aiming at the problems in the prior art, the invention provides a sand-gravel soil triaxial shearing discrete element model microscopic parameter calibration method, which is described in detail below with reference to the accompanying drawings.

As shown in fig. 1, the method for calibrating microscopic parameters of a three-axis shearing discrete element model of sandy gravel soil provided by the embodiment of the present invention includes the following steps:

s101, establishing a sand-pebble triaxial shear numerical model through PFC2D/3D software, selecting a contact model, and determining calibrated macroscopic and microscopic parameters and ranges thereof;

s102, designing a numerical test scheme by using an orthogonal test design method;

s103, performing a numerical test according to a test scheme by using PFC2D/3D software;

s104, analyzing the significance of each microscopic parameter on the influence of macroscopic mechanical parameters by adopting a multi-factor variance analysis method, and screening out main microscopic parameters influencing macroscopic mechanical indexes;

s105, establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis;

and S106, converting the calibration problem into a multi-objective function mathematical programming problem to solve, and comparing and verifying the multi-objective function mathematical programming problem with the physical test result of typical cobble soil.

As shown in fig. 2, the system for calibrating mesoscopic parameters of a three-axis shearing discrete element model of sandy gravel soil provided by the embodiment of the present invention includes:

the test scheme design module 1 is used for designing a numerical test scheme by using an orthogonal test design method;

the significance analysis module 2 is used for analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters by adopting a multi-factor variance analysis method and screening out main microscopic parameters influencing the macroscopic mechanical indexes;

the linear relation establishing module 3 is used for establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis;

and the comparison and verification module 4 is used for converting the calibration problem into a multi-objective function mathematical programming problem to solve and comparing and verifying the multi-objective function mathematical programming problem with the physical test result of typical sand and pebble soil.

The technical solution of the present invention is further described below with reference to specific examples.

1. The invention researches the microscopic parameter calibration problem of the sandy gravel soil triaxial shear particle flow model. Firstly, a numerical test scheme is designed by using an orthogonal test design method. And then, analyzing the test result by using a multi-factor variance analysis method, analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters, and screening out the main microscopic parameters influencing the macroscopic mechanical indexes. And then establishing a linear expression between the macroscopic mechanical index and the main microscopic parameter by using a regression analysis method. And finally, converting the parameter calibration problem into a multi-target mathematical programming problem to solve. The three-axis shear test of the typical sandy gravel soil is compared to find that the numerical test result and the physical test result are basically consistent, and the feasibility of the calibration method is verified.

The method firstly utilizes orthogonal test design to design a numerical test scheme, and then adopts a multi-factor variance analysis method to perform significance analysis on the influence of each microscopic parameter on the macroscopic parameter. And then establishing a linear relation between the macroscopic parameters and the main microscopic parameters through regression analysis. And finally, converting the calibration problem into a multi-objective function mathematical programming problem to solve, and comparing and verifying the problem with the physical test result of typical sand and pebble soil.

2. Contact bonding model

The physical and mechanical properties of the rock-soil material are greatly dependent on the contact property between particles, and the contact model is the visual embodiment of the contact property. PFC5.0 provides 10 kinds of built-in contact models to simulate contact properties of rock-soil mass, such as linear contact model (linear), linear contact bond model (linear bond), linear parallel bond model (linear bond), and the like. The linear contact bonding model treats contact bonding as a pair of springs with constant normal and tangential stiffnesses, which also have both normal and tangential tensile strengths. As shown in FIG. 3, the strength parameters of the contact-bonding model include the normal strength F_c ⁿAnd tangential strength F_c ^s. When the overlapping amount of the adjacent particles is less than zero, the particles generate a pulling force at the contact position due to relative separation, and if the pulling force is greater than F_c ⁿThe contact bonding is subject to tensile failure, at which point both the normal contact force and the tangential contact force become zero. If the tangential force is greater than F_c ^sShear failure of the contact bond occurs. Compared with a linear model, the parallel bonding model has an inhibiting effect on the sliding behavior among particles, and compared with the parallel bonding model, the contact part can only transmit force but not torque. Many scholars introduce the linear contact model into the parameter calibration of the sandy cobble soil according to the characteristic of the linear contact model, and simulate the embedding and biting action among irregular cobble particles by increasing the bonding strength, thereby obtaining better effect.

The mesoscopic parameters required to be input by the linear contact bonding model include a linear group and a contact bonding group, wherein the linear group comprises: contact effective modulus (E)_c) Normal tangential stiffness ratio (k)_n/k_s) And coefficient of friction (μ). The contact adhesive group includes: bonding control gap, bonding tensile strength (sigma)_c) And bond shear strength (τ)_c)。

3. Numerical model

3.1 procedure for triaxial test

The PFC3D is adopted in the numerical test, and three processes of sample preparation, confining pressure application and loading are realized by writing a triaxial shearing program:

(1) preparing a sample: the numerical test does not consider the influence of the grain composition temporarily, and the model size H multiplied by L is 600mm multiplied by 300mm according to geotechnical test standard. The particles take the form of both spheres (ball) and particle clusters (clusters) where a series of spheres are passed through a rigid bedLinked together to effectively increase the interlocking effect between particles), according to the research of the related literature, the L/R is taken_min＝30，R_max/R_min1.66, porosity n 0.32, particle density 2700kg/m³. The whole test vessel was uniformly filled with the particles by the expansion method, and the number of particles generated was about 20000, as shown in FIG. 4.

(2) Applying confining pressure: in the numerical test of the present invention, three sets of confining pressures of 100kPa, 200kPa and 300kPa were applied to each set of samples, respectively. The constant confining pressure is applied by monitoring the difference between the contact pressure between the sidewall and the particles and the set confining pressure in each calculation step, and adjusting the displacement of the sidewall.

(3) Vertical loading: the loading process is realized by endowing the upper loading plate and the lower loading plate with certain speed, the loading speed cannot be too high, the test sample is kept in a quasi-static state, otherwise, the accuracy of the test result is influenced, and the mechanical property of the sandy gravel soil is related to the loading speed. The loading rate of the invention is 1.2mm/s, and the unbalanced force rate in the sample is found to be less than 1 multiplied by 10 by monitoring in the shearing process^-4And the variation amplitude of the confining pressure is not more than 0.3 percent.

(4) And (4) terminating the loading: when the peak value exists, the axial strain of the test should reach 3% -5% of the peak value, and the loading can be stopped. If no peak exists, the loading is stopped when the axial strain reaches 15 percent.

4. Orthogonal test and analysis of variance

4.1 principles of orthogonal experiments

In scientific experiments, factors needing to be investigated and researched are more often, and the level number of the factors is more than two. If a full test is performed with each level of each factor matched to each other, the number of tests will increase exponentially as the number of factors increases. Although the number of tests can be reduced by adopting a simple comparison method, the matching among levels of all factors is not uniform in the test process, the uniformity of data point distribution is not guaranteed, and an erroneous conclusion can be often obtained under the condition. The orthogonal test design can obtain a more reliable test result on the premise of obviously reducing the test times.

The orthogonal table is designed according to the orthogonal principle, is a normalized table, is a cost-saving tool for arranging tests and analyzing test results in the orthogonal design, and is generally L_n(r^m) And (4) showing. Wherein L is the code number of the orthogonal table; n is the number of horizontal rows (test times) of the orthogonal table; r is a factor level number; m is the number of columns (the number of factors that can be arranged at most) of the orthogonal table. E.g. orthogonal table L₈(2⁷) It shows that there are 8 rows and 7 columns in total, and if it is used to arrange orthogonal tests, a maximum of 7 2-level factors can be arranged, and the number of tests is 8. Several important properties of the orthogonal table: each level appears in any column and occurs the same number of times; all possible combinations of various levels between any two classes occur and occur equally often. The two properties determine the balanced dispersion and comprehensive comparability of the orthogonal test.

4.2 analysis of variance

The analysis of variance can estimate the magnitude of errors, can accurately estimate the importance degree of influence of each factor on the test result, and is commonly used in data analysis of orthogonal tests. The analysis of variance analyzes influence factors influencing the fluctuation of the total variance by decomposing the variance, and the influence factors are divided into intra-group errors and inter-group errors according to the sources of the influence factors. Constructing F by calculating the ratio of the inter-mean square to the intra-mean square of factor A_AStatistics, for a given significance level a, find a critical value F_αIf F is_A>F_αThen factor A is considered to have a significant effect on the test results, and F_ARelative to F_αThe larger the factor, the more significantly the effect of the factor on the test results.

If L is adopted_n(r^m) Orthogonal table design test with the test result of y_i(i ═ 1, 2, … n). The basic steps of analysis of variance are as follows:

(1) calculating the sum of squares of total deviations

Wherein,

(2) calculating the sum of squared deviations due to factor j

Wherein,

K_iis the sum of the test results with the j-th column with the horizontal number i.

(3) Calculating the sum of squared deviations of the errors

SS_e＝∑SS_{Empty column} (5)

(4) Degree of freedom of calculation

SS_TDegree of freedom of (2):

df_T＝n-1 (6)

SS_jdegree of freedom of (2):

df_j＝r-1 (7)

SS_efree of (2):

df_e＝∑df_{empty column} (8)

(5) Computing mean square

(6) Calculating the F value

(7) Significance test

4.3 selection of test index and factor level

(1) Selection of test indexes

The calibration of the three-axis shearing microscopic parameters of the particle flow is generally based on the stress-strain relationship and the damage envelope of the PFC model, which are consistent with the results of the indoor test. Therefore, the macro characteristic index selection follows the following principle: firstly, the selected index can be conveniently and easily obtained from the test; secondly, the selected index can better describe the form of the calibrated curve. The stress-strain curve of the sand-gravel soil routine triaxial test has an initial elastic modulus E as shown in FIGS. 5 and 6₀Peak intensity σ_fResidual intensity σ_r(ii) a The indexes for describing the Mohr-coulomb strength envelope include cohesive force c and internal friction angle

And the like. The selection of specific indexes needs to be determined according to the form of the calibration curve and the focus of attention. The invention focuses on the pre-peak behavior of the stress-strain curve of the triaxial shear test, so that the initial elastic modulus E is selected₀Peak intensity σ_fCohesion c and internal friction angle

Scholars at home and abroad have qualitatively obtained the influence rule of each microscopic parameter of the linear contact bonding model on the macroscopic parameter through research: the macroscopic elastic modulus is linearly related to the contact effective modulus; the poisson's ratio increases with increasing normal tangential stiffness ratio; the tangential and normal bond strengths and friction coefficients between particles are mainly related to the peak strength; the content of the particle clusters can effectively improve the shear strength. According to the invention, based on the conclusion, the effective contact modulus, the normal tangential stiffness ratio, the inter-particle friction coefficient, the bonding shear strength, the bonding tensile strength and the content of particle clusters are selected as the mesoscopic parameter indexes of the numerical test.

(2) Determination of levels of factors in orthogonal experiments

6 microscopic parameters and 4 macroscopic parameters in the table 1 are selected as calibration indexes, and each factor is selected to have four levels. The factor levels were selected as shown in Table 1. Numerical tests show that the values of the macroscopic parameters in different combinations of the selected factor levels are as follows: initial modulus of elasticity E₀: 10MPa to 100MPa (the confining pressure is 300 kPa); peak intensity σ_f: 500kPa to 2050kPa (the confining pressure is 300 kPa); cohesive force c: 0 to 600 kPa; internal friction angle

17 to 47 degrees. Therefore, under the selected factor level, the obtained macroscopic mechanical index range covers the macroscopic force index range of most of sand and gravel soil.

Table 1 levels of factors

4. Analysis of results

4.1 numerical test results

According to the number of factors and the level of the factors, the invention adopts L₃₂(4⁹) Orthogonal tables numerical tests were performed, and the test protocol and test results are shown in table 2.

Table 2 numerical calculation scheme and results based on orthogonal design

4.2 multifactor analysis of variance

The results of the numerical tests were analyzed for variance, and the analysis results are shown in Table 3. The total sum of squares of deviations, the sum of squares of deviations of the respective factors, and the sum of squares of errors can be calculated according to the equations (1) to (5). The total degree of freedom method f can be calculated according to the formulas (6) to (8)_T31, degree of freedom f of each factor _j3 and degree of freedom of error f _e13. Selecting significance level alpha-0.05 and alpha-0.01, and looking up table to obtain F_0.05(3,13)＝3.41，F_0.01(3,13) ═ 5.74. When F is present_0.05<F_j<F_0.01Then, factor j is considered to have a significant effect on the test results, denoted as "+; when F is present_j>F_0.01When considered, factor j was considered to have a very significant effect on the test results, noted as "×". In order to more intuitively analyze the degree of influence of the test factors on the test index, the analysis of variance results are plotted into a histogram, as shown in fig. 7.

TABLE 3 ANOVA TABLE

From FIG. 7, the effective contact modulus E between particles can be seen_cNormal tangential stiffness ratio k_n/k_sAnd the content of the particle clusters to the initial modulus of elasticity E₀There is a very significant effect. Bond tensile strength sigma_cFor initial modulus of elasticity E₀There is a significant impact. Other factors though to the initial modulus of elasticity E₀Also has a certain influence, but with respect to E_c、k_n/k_sParticle cluster content and sigma_cIts effect is completely negligible. Meanwhile, the sensitivity sequence of each microscopic parameter to the initial elastic modulus is obtained as follows: e_c>k_n/k_s>clump>τ_c。

FIG. 8 showsShowing the coefficient of friction mu, the content of column, the bond shear strength tau_cAnd contact modulus E_cFor peak intensity σ_fHaving a very significant effect on the bond tensile strength σ_cAnd normal tangential stiffness ratio k_n/k_sFor peak intensity σ_fThe effect of (c) is negligible. At the same time, the peak intensity σ is obtained_fMost sensitive to changes in the coefficient of friction mu, followed by the content of particle clusters, the bond shear strength tau_cAnd contact modulus E_cHowever, the significance of the latter three effects on peak intensity was approximately the same.

Figure 9 gives a multi-factor analysis of variance histogram of the adhesion force. The contact modulus E can be seen_cAdhesive shear strength tau_cAnd bond tensile strength σ_cHas a very significant effect on cohesion c; tangential normal stiffness ratio k_n/k_sThe effect of changes in the amount of column and coefficient of friction μ on cohesion c is not significant. The sensitivity of the main microscopic parameters to cohesion is ranked: e_c>τ_c>σ_c。

FIG. 10 shows the coefficient of friction μ and the contact modulus E_cAnd the amount of column versus internal angle of friction

Having a very significant effect on the bond tensile strength σ_cAdhesive shear strength tau_cAnd normal tangential stiffness ratio k_n/k_sFor internal friction angle

The effect of (c) is negligible. Of these, the coefficient of friction has the most significant effect on the internal friction angle, followed by the contact modulus and the content of particle clusters.

4.3 regression analysis

According to the result of 4.2 variance analysis, regression analysis is carried out on the significant influence factors and the macroscopic index, and the corresponding relation of macro-microscopic parameters can be obtained:

E₀＝0.142E_c-5.4k_n/k_s+9.83τ_c+15.87clump+16.41(R²＝0.68) (12)

σ_f＝-2.47E_c+985.27μ+631.81τ_c+653.08clump-5.71(R²＝0.903) (13)

c＝-1.33E_c+28.08k_n/k_s+163.68σ_c+253.86τ_c-66.61(R²＝0.828) (14)

as can be seen from equation (12), the initial modulus of elasticity increases with increasing contact modulus, bond shear strength and content of particle clusters, and decreases with increasing normal-to-tangential stiffness ratio; from the same equation (13), the relationship in which the peak strength is positively correlated with the friction coefficient, the contact shear strength and the contact tensile strength is negatively correlated with the contact modulus. The cohesive force is increased along with the increase of the normal tangential stiffness ratio, the bond tensile strength and the bond shear strength and is reduced along with the increase of the contact modulus, which is obtained by the formula (14); from the formula (15), the internal friction angle is in a positive correlation with the contact modulus, the friction coefficient and the content of the particle clusters.

5. Calibration process and verification of mesoscopic parameters

5.1 optimization of the mesoscopic parameters

In order to make the macroscopic mechanical indexes obtained by numerical tests and the macroscopic mechanical indexes obtained by indoor tests match as much as possible, the section optimizes each microscopic parameter of the numerical tests by using a multi-objective mathematical programming method based on the regression equation obtained in section 4.3.

(1) Objective function

In order to make the physics obtained by numerical and laboratory tests as close as possible, the objective function can be written as:

wherein E is₀、σ_fC and

the regression equations obtained by the regression are shown in the formulas (12) to (15);

E₀ ^*、σ_f ^*、c^*and

respectively, macroscopic parameters obtained by indoor tests.

(2) Constraint conditions

(3) optimization of mesoscopic parameters

For the problem of optimization of the multi-objective function, the fgoalattain function in MATLAB can be adopted for solving. The goal planning problem may be converted to a minimum solving problem by letting the goal matrix goul be 0. By giving an initial value x₀And finally optimizing the weight value and the constraint condition to obtain the microscopic parameter.

5.2 example validation

In Baiyong and the like, slightly dense sandy pebbles buried 8-10 m deep along No. 2 lines of Chengdu subways are used as materials to carry out indoor large-scale triaxial consolidation drainage tests, and the test results are shown in Table 4. The invention carries out the microscopic parameter calibration of the sandy gravel soil based on the characteristic.

TABLE 4 test values and simulation values of sandy gravel soil macro parameters

The macro parameters corresponding to the indoor test in table 4 are substituted into the formula (16), the corresponding microscopic parameters obtained by optimization are shown in table 5, the microscopic parameters are input into the PFC3D numerical test model, and the obtained macro parameters are shown in table 4. Comparing with the indoor test result, it is found that there is a large error in the peak intensity and the cohesive force, and based on the analysis results of the sections 4.2 and 4.3 of the present invention, the microscopic parameters are adjusted, as shown in table 5, and are substituted into the numerical model, and the obtained macroscopic parameters are shown in fig. 11.

TABLE 5 Fine parameters of sandy gravel soil calibration results

Fig. 11 shows stress-strain curves of the sand-gravel soil in the laboratory test and the numerical test. It can be seen from the figure that compared with the indoor test, the stress-strain curve obtained by the numerical test has a larger elastic range and a small peak value change, and a certain softening phenomenon can occur along with the increasing of the strain, but the initial elastic modulus, the peak value strength and the curve form of the numerical test result and the physical test result are basically consistent, and the characteristics of the sandy gravel soil can be better simulated.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.

Claims

1. A microscopical parameter calibration method for a sandy gravel soil triaxial shearing discrete element model is characterized by comprising the following steps of:

designing a numerical test scheme by using an orthogonal test design method;

2. The sand gravel soil triaxial shear discrete element model meso-scale parameter calibration method as claimed in claim 1, wherein in the sand gravel soil triaxial shear PFC2D/3D model establishment:

(1) selection of contact model

The microscopic contact model adopted by the numerical model is a linear contact bonding model, and the linear contact bonding model simulates the embedding and occluding actions among pebble particles by increasing the bonding strength among the particles so as to reflect the characteristics of the sandy pebble soil;

(2) selection of macroscopic parameters

The calibration of the particle flow triaxial shearing microscopic parameters is generally based on the standard that the stress-strain relationship and the damage envelope of a numerical model conform to the indoor test result; the macro characteristic index selection follows the following principle: firstly, obtaining selected indexes from tests; secondly, the selected index can better describe the form of the calibrated curve; selecting an initial modulus of elasticity E₀Peak intensity σ_fCohesion c and internal friction angle

Four parameters are used as macroscopic indicators.

(3) Selection of mesoscopic parameters

The rule of the influence of each microscopic parameter of the linear contact bonding model on the macroscopic parameter is as follows: the macroscopic elastic modulus is linearly related to the contact effective modulus; the poisson's ratio increases with increasing normal tangential stiffness ratio; the tangential and normal bond strengths and friction coefficients between particles are mainly related to the peak strength; the content of the particle clusters improves the shear strength; selection of contact effective modulus E based on conclusion_cNormal tangential stiffness ratio k_n/k_sInter-particle friction coefficient mu and adhesive shear strength sigma_cBonding tensile strength tau_cAnd the content of the particle cluster is used as a microscopic parameter index of the numerical test;

(4) value range of macro and micro parameters

Selecting 6 microscopic parameters and 4 macroscopic parameters as calibration indexes, wherein each factor is selected to be four levels; through numerical tests, the values of the mesoscopic parameters are found to be in the following ranges under different combinations of selected factor levels:

E_c:20MPa～200MPa；k_n/k_s:1～4；σ_c:0.25～1.0MPa；τ_c:0.25～1.0MPa；

obtained macroscopic parameter ranges:

17°～47°；

3. The sand-gravel soil triaxial shear discrete element model microscopic parameter calibration method as claimed in claim 1, wherein in the second step, the design of the numerical test scheme by using the orthogonal test design method comprises the selection of factor levels and an orthogonal table:

the factor level is selected from the following steps:

Factor level 4: e_c＝200MPa，k_n/k_s＝4，μ＝1.00，σ_c＝1.00MPa，τ_c＝1.00MPa，clump＝100％；

The design method of experiment scheme by using orthogonal test method is characterized in that the orthogonal table is designed according to orthogonal principle, is normalized, is a cost-saving tool for arranging test and analyzing test result in orthogonal design, and uses L_n(r^m) Represents; wherein L is the code number of the orthogonal table; n is positiveThe number of horizontal rows of the cross table is the test times; r is a factor level number; m is the number of columns of the orthogonal table and the number of factors which can be arranged at most; in the orthogonal table, each level appears in any column with the same number of occurrences; all possible combinations of various levels between any two classes occur and occur equally often.

4. The method for calibrating the microscopic parameters of the sandy gravel soil triaxial shear discrete element model as claimed in claim 1, wherein in the third step, the numerical test mainly comprises the following steps:

(1) preparing a sample: the numerical test does not consider the influence of grain composition temporarily, and the model size H multiplied by L is 600mm multiplied by 300 mm; the particles adopt two forms of spheres and particle clusters, and L/R is taken_min＝30，R_max/R_min1.66, porosity n 0.32, particle density 2700kg/m³Inputting a contact model and parameters according to a test scheme; the whole test container is uniformly filled with particles by an expansion method, and the number of the generated particles is 20000;

(2) applying confining pressure: in the numerical test, three groups of confining pressures of 100kPa, 200kPa and 300kPa are respectively applied to each group of samples; the constant confining pressure is applied by monitoring the difference between the contact pressure between the side wall and the particles in each calculation step and the set confining pressure so as to adjust the displacement of the side wall;

(3) vertical loading: the loading process is realized by endowing the upper loading plate and the lower loading plate with certain speed, the loading speed cannot be too high, the test sample is kept in a quasi-static state, otherwise, the accuracy of the test result is influenced, and the mechanical property of the sandy gravel soil is related to the loading speed; the loading rate was 1.2mm/s and the rate of unbalanced forces in the sample during shear was found to be less than 1X 10 by monitoring^-4The variation amplitude of the confining pressure is not more than 0.3%;

5. As claimed in claimThe sand-gravel soil triaxial shearing discrete element model microscopic parameter calibration method is characterized in that in the fourth step, the variance analysis comprises the following steps: the analysis of variance is used for estimating the size of an error and the influence degree of each factor on a test result, and is used for data analysis of an orthogonal test; the variance analysis analyzes influence factors influencing the fluctuation of the total variance through variance decomposition, and the influence factors are divided into intra-group errors and inter-group errors according to the sources of the influence factors; constructing F by calculating the ratio of the inter-mean square to the intra-mean square of factor A_AStatistics, for a given significance level a, find a critical value F_αIf F is_A>F_αThen factor A is considered to have a significant effect on the test results, and F_ARelative to F_αThe larger the factor, the more significant the effect of the factor on the test results;

by using L_n(r^m) Orthogonal table design test with the test result of y_i(i ═ 1, 2, … n); analysis of variance is as follows:

(1) calculate the sum of the squares of the total deviations:

wherein,

(2) calculating the sum of squared deviations due to factor j

Wherein,

wherein, K_iIs at the jth column levelThe sum of the test results with number i;

(3) calculating the sum of squared deviations of the errors:

SS_e＝∑SS_{empty column}；

(4) And (3) calculating the degree of freedom:

SS_Tdegree of freedom of (2):

df_T＝n-1；

SS_jdegree of freedom of (2):

df_j＝r-1；

SS_efree of (2):

df_e＝∑df_{empty column}；

(5) Calculating the mean square:

(6) calculating the F value:

(7) and (3) significance test:

6. The method for calibrating the meso-scale parameters of the three-axis shear discrete element model of sandy gravel soil as claimed in claim 1, wherein in the fifth step, the regression analysis comprises:

E₀＝0.142E_c-5.4k_n/k_s+9.83τ_c+15.87clump+16.41(R²＝0.68)；

σ_f＝-2.47E_c+985.27μ+631.81τ_c+653.08clump-5.71(R²＝0.903)；

c＝-1.33E_c+28.08k_n/k_s+163.68σ_c+253.86τ_c-66.61(R²＝0.828)；

the initial modulus of elasticity increases with increasing contact modulus, bond shear strength and content of particle clusters, and decreases with increasing normal to tangential stiffness ratio; the peak strength is in positive correlation with the friction coefficient, the contact shear strength and the contact tensile strength, and is in negative correlation with the contact modulus; the cohesive force is increased along with the increase of the normal tangential stiffness ratio, the cohesive tensile strength and the cohesive shear strength, and is reduced along with the increase of the contact modulus; the internal friction angle is in positive correlation with the contact modulus, the friction coefficient and the content of the particle cluster;

in the sixth step, the step of converting the calibration problem into a multi-objective function mathematical programming problem to solve and comparing the problem with the physical test result of typical sand and pebble soil to verify, comprises the steps of optimizing each microscopic parameter of a numerical test by using a multi-objective mathematical programming method based on a regression equation, and comprises the following steps:

(1) the objective function is:

wherein E is₀、σ_fC and

are respectively returnObtaining a regression equation; e₀ ^*、σ_f ^*、c^*And

respectively obtaining macroscopic parameters obtained by an indoor test;

(2) and (3) constraining the value range of the mesoscopic parameters, wherein the constraint conditions are as follows:

(3) optimizing the microscopic parameters, and solving the problem of optimizing the multi-objective function by adopting an fgoalatain function in MATLAB; converting the target planning problem into a minimum solving problem by setting the target matrix good to 0; by giving an initial value x₀And finally optimizing the weight value and the constraint condition to obtain the microscopic parameter.

7. A sand gravel soil triaxial shearing discrete element model mesoscopic parameter calibration system for implementing the sand gravel soil triaxial shearing discrete element model mesoscopic parameter calibration method as claimed in any one of claims 1 to 6, wherein the sand gravel soil triaxial shearing discrete element model mesoscopic parameter calibration system comprises:

8. A computer device, characterized in that the computer device comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of:

designing a numerical test scheme by using an orthogonal test design method; analyzing the test result by using a multi-factor variance analysis method, analyzing the significance of the influence of each microscopic parameter on the macroscopic mechanical parameters, and screening out the main microscopic parameters influencing the macroscopic mechanical indexes; establishing a linear expression between a macroscopic mechanical index and a main microscopic parameter by using a regression analysis method; and finally, converting the parameter calibration problem into a multi-target mathematical programming problem to solve, and comparing the problem with a typical sandy gravel soil triaxial shear test to find that a numerical test result is basically consistent with a physical test result and verify the feasibility of the calibration method.

9. A computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:

10. An information data processing terminal, characterized in that the information data processing terminal is used for implementing the microscopical parameter calibration system of the sandy gravel soil triaxial shearing discrete meta-model as claimed in claim 7.