CN103500259A - Modeling method of rock-soil model - Google Patents

Abstract

The invention discloses a modeling method of a rock-soil model, which is characterized in that, according to the natural formation process of the granular rock mass, the method makes the particles naturally fall and accumulate, compact and fully contact in a specified area, and then The configuration is performed by deleting particles outside the specified shape, and the initial in-situ stress field is obtained through calculation until equilibrium; compared with the classical procedure, this method does not need to calculate mul, build a boundary wall between the slope wall and the soil layer, and eliminate suspended particles; but The particle drop calculation and configuration process are added; it includes the following steps: generate the boundary wall around the model, and set the stiffness of the wall; generate particles in the vertical projection area of the stratum rock and soil; set the acceleration of gravity, particle density, stiffness, friction coefficient, and calculate to the required stacking height; set the interaction between particles; delete unnecessary particles, and calculate to balance; the invention can be used for establishment of geotechnical engineering model.

Description

A Modeling Method of Geotechnical Model

technical field

The invention relates to geotechnical engineering, in particular to the modeling of loose granular rock and soil during geotechnical engineering simulation.

Background technique

PFC3D is a high-end product released by Itasca in 2008, which is especially suitable for the research of complex mechanism problems. It is a micro/micro-mechanics program developed by using explicit difference algorithm and discrete element theory. It considers the basic mechanical properties of the medium from the perspective of the basic particle structure of the medium, and considers the basic properties of a given medium under different stress conditions It mainly depends on the change of the contact state between particles, and is suitable for studying the cracking and cracking development of granular aggregates, as well as large displacement problems such as the flow of particles. In geotechnical engineering, it can be used to study a series of problems that are difficult to solve by traditional numerical methods, such as structural cracking, rockfill material properties and stability, mine caving mining, slope disintegration, and blasting impact.

The application of PFC3D is more difficult and has higher requirements for users. There are not many applications and researches on PFC3D in China. Zhang Long et al. studied the PFC3D simulation of Jiweishan high-speed long-distance landslide movement process; Chen Yikai conducted a stability analysis on the tailings dam body based on particle flow discrete elements. But at present, the shapes and sizes of the models constructed by PFC3D are relatively simple, and it is difficult to meet the needs of practical engineering.

The author has been engaged in simulation research and engineering simulation application for a long time. It is believed that according to the modeling steps provided by the PFC3D user manual, there will be some problems in modeling, such as the difficulty in determining the radius magnification factor (mul), the difficulty in ensuring the contact degree of grain boundaries with different properties, and the removal of the boundary wall between the slope wall and the soil layer. The problem of the ball flying out, the influence of the model shape is not considered when determining the mul after specifying the porosity, etc. These problems make the model construction inaccurate, and the deformation of the model after calculation is large, resulting in rework and other problems.

The author considers that the formation process of granular rock and soil mass is due to weathering, deposition, etc., which make the particles accumulate layer by layer from bottom to top in the vertical direction, and undergo a natural compaction process. According to this idea, the "falling method" (Fall Particles Method, FPM)" to construct the initial stress field. The basic steps, advantages and disadvantages of FPM are discussed, and applied to the excavation examples of tailings ponds and coal piles.

According to the introduction in the PFC3D user manual PROBLEM SOLVING WITH PFC ^3D , the general steps of numerical analysis of geotechnical problems are shown in Figure 1.

Problems in the PFC3D model building process

1. Determination of radius magnification factor (mul)

The first step of model construction is to generate particles, there are BALL and GENERATE commands. The BALL command is generally used for regular structures, and GENERATE is used for geotechnical structures, where the parameter rad r1 r2 means that the radius of the particle is randomly or regularly distributed in [r1, r2]. The radius of the particles and the size of the filled space determine the number of particles. After using PROP to set the density, shear modulus and elastic modulus of the particles, the radius magnification factor mul of the particles is initialized (see the user manual for the definition). The problem is how to determine mul. If the mul is small and the specified space is not filled enough, pfc3d will automatically expand the mul to continue the calculation. In specific engineering problems, there are many particles and the time cost is very high; if the mul is large, pfc3d will automatically shrink the mul to continue the calculation, but due to Density, shear modulus, and modulus of elasticity have been set so that the particle spheres are elastically deformed under the constraints of the boundary wall and the radius after the initial mul enlargement. At this time, the designated space can accommodate particles, but the particles have accumulated elastic energy, which cannot be eliminated even after performing solve. When the boundary wall is deleted after balancing, the particles will flow in the direction restricted by the boundary wall. This is because during the process of adjusting mul, the boundary wall has been exerting force on the particles. If the boundary wall is not used in the process of adjusting the mul, the shape of the model and the shape of the layered rock and soil mass are difficult to guarantee. If a boundary wall is used, it must be removed in the final calculation of the initial ground stress balance to ensure adequate contact between the different soil layers. This problem exists regardless of whether the mul is large or small, and it is difficult to avoid. According to Fig. 1 and Fig. 2, the models of the tailing ponds are shown in Fig. 3 and Fig. 4 respectively when the balance calculation 100 is carried out before and after the demarcation wall is deleted according to Fig. 1 and Fig. 2 .

2. The contact of particle boundaries with different properties

According to the steps described in the first section and the relevant content of the user manual, particles with different properties are produced separately. As shown in Figure 3, the geotechnical layers with different attributes have different shapes. To construct a geotechnical layer with a specified shape, a dividing wall must be used, but there are problems in using the dividing wall. The radius of the particles and the size of the filled space determine the number of particles. I found that the shape of the space formed by the boundary wall also affects the number of particles and related parameters such as porosity. The infill problem occurs when the spatial shape has small angles. As shown in Figure 5, it is an enlarged view of the back side of the initial dam in Figure 3.

As can be seen in Figure 5, there is no particle filling at the tip of the soil layer. Of course, the smaller the particles, the less obvious this problem is, but the computational cost will increase exponentially. Looking at this phenomenon from another aspect, it can be considered that before the boundary wall is removed, the degree of contact between rock and soil particles in each layer is difficult to guarantee, which is obviously wrong, as shown in Figure 6. If the boundary wall is removed, the upper particles of the unfilled space will move downward under the action of gravity, and at the same time, the contact with the boundary wall is limited. The ball will lose the constraint of the wall and move in the opposite direction, which will seriously deform the model. 4 reasons for the phenomenon.

3 Porosity problem

In practical engineering problems, an important parameter of granular body is porosity. In PFC3D, it is often necessary to construct granular bodies with specified porosity. The definition of porosity n in PFC3D is shown in the formula in Figure 21.

In the formula: Vp is the volume of particles accommodated in the construction model of the boundary wall, and V is the volume of the construction model of the boundary wall.

PFC3D provides a method for constructing a specified porosity model, and the derivation process is shown in Figure 22.

In the formula: R is the particle radius, R _old is the particle radius obtained from the previous calculation, n _old is the model porosity obtained from the previous calculation, and m is the adjustment coefficient, namely mul.

The command flow in PFC3D is:

loop while bp # null

sum = sum + (4.0/3.0) * pi * b_rad(bp)ˆ3

bp = b_next(bp)

end_loop

pmeas = 1.0 - sum / tot_vol

_mult=((1.0-poros)/(1.0-pmeas))ˆ(1.0/3.0).

However, there is a problem with the formula in Figure 22. The first step in the derivation is that all the non-void parts in the model are the volume of the particle sphere. This is ideal, without considering the elastic deformation of the sphere, and what is more serious is that the tip void part of the model is not considered, such as the tip void part of the model in Figure 5. This ideal situation results in the unfilled space that should be filled by the sphere, making Vp decrease, n increase, and m increase. The final mul is greater than the applicable mul. In turn, the sphere has a greater deformation, and the particle flow is more serious after the boundary wall is deleted. The more complex the shape, the less accurate the mul.

Aiming at the problems encountered in the above-mentioned practical engineering, the author proposes FPM to construct the PFC3D model.

Contents of the invention

According to the modeling steps provided by the PFC3D user manual, there will be some problems in modeling, such as the difficulty in determining the radius magnification factor (mul), the difficulty in ensuring the contact degree of grain boundaries with different properties, and the removal of the boundary wall between the slope wall and the soil layer. The problem of the ball flying out, the influence of the model shape is not considered when determining the mul after specifying the porosity, etc. These problems make the model construction inaccurate, and the deformation of the model after calculation is large, resulting in rework and other problems. Considering that the formation process of rock and soil with granular properties is due to weathering, deposition, etc., the particles accumulate layer by layer in the vertical direction from bottom to top, and undergo natural compaction. According to this idea, the "falling method" (Fall Particles Method, FPM)" to construct the geotechnical model.

1. The process of building a model by the falling method

FPM constructs the model through the process of accumulating and compacting particles layer by layer from bottom to top in the vertical direction. The drop method is divided into the overall drop method (Overall Particles Fall Method, OPFM) and Hierarchical Fall Method (Hierarchical Particles Fall Method, HPFM). The processes are shown in Figures 7 and 8 respectively. Figure 9 shows the process of HPFM constructing the third rock and soil layer of the tailings pond.

Figure 9 shows the process of constructing a layer of geotechnical mass by HPFM. In essence, the above-mentioned OPFM and HPFM are consistent, the difference is that HPFM’s geotechnical property setting and balance calculation are step-by-step, which is closer to the actual situation; while OPFM uses FISH statements to determine the soil layer where each ball is located and then assign values of. The former consumes more time for balance calculation, and the latter consumes more time for attribute setting.

2. Method of removing unwanted particles

What needs to be explained here is that in the two methods, the method for deleting unnecessary particles is a FISH statement. However, the specific implementation is divided into two methods. These two methods can be simply expressed as deleting the particles in the specified area and judging that the particles are deleted after the specified deletion area. The former uses the command flow range to locate, and the latter uses the FISH statement find_ball( id) positioning. The former is more efficient, but imprecise, and the latter is the opposite. The more complicated the shape of the rock and soil layer, the closer the efficiency of the two is. The model in Figure 9 uses the latter to delete particles, and its command flow is shown below, and the former command flow is shown in Section 4. The final calculation to equilibrium model of the tailings pond constructed using HPFM is shown in Figure 10. Model CForce Chains are shown in Figure 11.

def delz3

loop nn(20001,23000)

bp=find_ball(nn)

_brad = b_rad(bp)

_bx = b_x(bp)

_by = b_y(bp)

_bz = b_z(bp)

sx=-140

lx=100

sy=0

ly=20

sz=0

lz1=0.125*_bx+17.5 ; The determination of the curve is based on the highest point and the lowest point in the vertical direction of the third rock and soil layer in Figure 9.

if _bx>sx then

if_bx<100 then

if _bz>lz1 then

command

del ball range id = nn

end_command

end_if

end_if

end_if

end_loop

end

Compared with Fig. 3, the outer envelope line and the boundary line of different rock-soil layers in Fig. 10 are not smooth, but rough, which conforms to the actual situation. Comparing Fig. 10 with Fig. 4, Fig. 4 only calculates 100 steps and there is a serious phenomenon of particle drifting, and the particles of different rock-soil layers have entered other rock-soil layers, which is wrong and causes serious deformation of the whole model. Figure 10 has been calculated to the equilibrium state, and the error phenomenon in Figure 4 does not appear. The only obvious deformation is that the left end of the bottom bedrock is squeezed and uplifted by the gravity of the tailings pond.

Description of drawings

Fig. 1 Recommended procedure for numerical analysis of geotechnical soils.

Fig. 2 Flowchart of establishing the actual model of particle flow.

Figure 3 Delete the model of the tailings pond before the dividing wall.

Figure 4 is the tailings pond model at 100 when the balance calculation is performed after the dividing wall is deleted.

Fig. 5 Enlarged view of the dorsal side of the initial dam.

Fig. 6 Contact situation of the dorsal area of the initial dam.

Figure 7 HPFM flow chart.

Figure 8 OPFM flow chart.

Figure 9 The process of HPFM constructing the third rock and soil layer of the tailings pond.

Figure 10 The final model of the tailings pond.

Figure 11 Model CForce Chains.

Fig. 12 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 13 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 14 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 15 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 16 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 17 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 18 Illustration of a coal pile constructed using OPFM and its important parameters after excavation.

Fig. 19 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 20 Illustration of important parameters of a coal pile constructed using OPFM and its excavation.

Fig. 21 The definition formula of porosity n.

Fig. 22 Formulas for constructing a model with specified porosity.

Detailed ways

In the above analysis, in order to illustrate the principle and use of the drop method, especially the HPFM, an example of a tailings pond was cited.

An example of OPFM and another method of particle deletion is discussed here. This example is a coal pile in a factory, which needs to use a large amount of coal in the production process, and the coal pile has a large volume. Because in the process of taking coal from the slope foot of the coal pile, the landslide accident of the coal pile was accidentally caused. Our research institute was entrusted by the factory to analyze the cause of the accident and its preventive measures. For this coal pile analysis, it is especially suitable for FPC3D, and this example illustrates the construction process of OPFM. The relevant parameters are: the height of the coal pile top from the ground (slope height) is 30m, and the slope length is 38.5m. Due to hardware limitations and analysis requirements, the width of the model is taken as 0.5m. The friction coefficient of the ground is 0.3, the friction coefficient of particles is 0.3, the density of coal is 1400kg/m ³ , the elastic modulus and shear modulus are 1×10 ⁸ Pa, and the particle radius range is [0.05m, 0.15m]. The excavation part is an oblique triangle with a height of 3.5m and a width of 3m, as shown in Figure 16.

Figures 12 to 20 show the process of using OPFM to construct coal piles, and excavate them, and obtain the vector diagrams of each field after excavation. Here is another method to delete particles, the code is as follows:

del ball range z 30.1 100

def delz2

num=(46.5-8)/0.1; 0.1 is the minimum particle size of the particles.

the loop nn(1,num)

x1=8+0.1*(nn-1)

x2=8+0.1*nn

z1=-0.7792*(x1+x2)/2+36.2338 ; The determination of the curve is based on the highest point and the lowest point in the vertical direction of the rock and soil layer shown in Figure 17.

z2=100

command

del ball range x x1 x2 z z1 z2

end_command

end_loop

end

The author proposes and implements the "drop method" for the construction of the pfc3d geotechnical model. According to the natural formation process of the granular rock mass, the small balls are naturally dropped and accumulated, compacted and fully contacted, and then the particles are deleted for configuration, and the calculation is carried out until the balance is obtained. Initial stress field model. The main features are as follows:

1) Point out the problems in the process of building geotechnical models using the modeling steps provided in the user manual. It mainly includes: the determination of the radius magnification factor (mul), the contact degree of grain boundaries with different properties, and the influence of porosity on the radius magnification factor.

2) The steps for constructing the model by the drop method are proposed. The drop method can be divided into OPFM and HPFM. The difference is that the rock and soil property setting and balance calculation of HPFM are step-by-step, which is closer to the actual situation; while OPFM uses the FISH statement to determine the soil layer where each ball is located and assign values. . The former consumes more time for balance calculation, and the latter consumes more time for attribute setting.

3) Two configuration methods for deleting particles are provided. The two methods can be expressed as deleting the particles in the specified area and judging that the particles are deleted after the specified deletion area. The former uses the command stream range to locate, and the latter uses the FISH statement find_ball(id) to locate. The former is more efficient, but imprecise, and the latter is the opposite. The more complicated the shape of the rock and soil layer, the closer the efficiency of the two is. And gives the code of the two methods.

Claims (9)

1. the modeling method of a geotechnical model, it is characterized in that, the method, according to the self-assembling formation process of particle rock mass, makes the whereabouts accumulation of particle nature, compacting and fully contacts in the regulation zone, then by deleting the outer particle of regulation shape, carries out configuration, is calculated to balance and obtains initially stress field; Compare the method with classical step and need not calculate mul, need not set up minute board between side slope wall and soil layer, need not eliminate suspended particle; But having increased particle falls to calculating and the configuration process; it comprises the steps: generation model periphery sides board, arrange the rigidity of wall; Generate particle in layer Rock And Soil vertical direction view field; Acceleration of gravity is set, particle density, rigidity, friction factor, and be calculated to the piling height needed; Interaction between particle is set; Delete unwanted particle, and be calculated to balance; the present invention can be used for the foundation of Geotechnical Engineering model.

2. the modeling method of a kind of geotechnical model according to claim 1, is characterized in that, the method is according to the self-assembling formation process of particle rock mass, make the whereabouts accumulation of particle nature, compacting and fully contact in the regulation zone, then carry out configuration by deleting the outer particle of regulation shape, be calculated to balance and obtain initially stress field, called after " whereabouts method (Fall Particles Method, FPM) ".

3. whereabouts according to claim 2 method, is characterized in that, fPM is by particle is successively piled up and the procedure construction model of compacting from top to bottom at vertical direction, the whereabouts method is divided into whole whereabouts method (Overall Particles Fall Method, OPFM) and layering whereabouts method (Hierarchical Particles Fall Method, HPFM).

4. whole whereabouts according to claim 3 method, is characterized in that, comprise: 1) generation model periphery sides board arranges the rigidity of wall; 2) the set Current Layer periphery sides board of Rock And Soil, arrange the rigidity of wall; 3) generate particle in current layer Rock And Soil vertical direction view field; 4) acceleration of gravity is set, particle density, rigidity, friction factor, and be calculated to the piling height needed; 5) interaction between particle, n_bond, s_bond are set; 6) delete unwanted particle, and be calculated to balance; 7) whether complete top layer structure; 8) adjust correlation parameter, and be calculated to balance.

5. layering according to claim 3 whereabouts method, is characterized in that, comprise: 1) generation model periphery sides board arranges the rigidity of wall; 2) the set Current Layer periphery sides board of Rock And Soil, arrange the rigidity of wall; 3) acceleration of gravity is set, particle density, rigidity, friction factor, and be calculated to the piling height needed; 4) interaction between particle, n_bond, s_bond are set; 5) delete unwanted particle, and be calculated to balance; 6) adjust correlation parameter, and be calculated to balance.

6. modeling method according to claim 1, is characterized in that, compare the method with classical PFC3D modeling procedure and need not calculate mul, need not set up minute board between side slope wall and soil layer, need not eliminate suspended particle.

7. modeling method according to claim 1, is characterized in that, compare the calculating of particle whereabouts and the configuration process of having increased with classical PFC3D modeling procedure.

8. the outer particle of deletion regulation shape according to claim 1, is characterized in that, in two kinds of methods, it is all the FISH statement that deletion does not need the method for particle.

9. in two kinds of methods according to claim 8, it is characterized in that, these two kinds of methods can simply be expressed as particle and the judgement particle deleted in appointed area and delete after specifying the deletion zone, and the former utility command stream range locates, and the latter uses FISH statement find_ball (id) location; The former efficiency is higher, but out of true, the latter is contrary; More complicated both efficiency of rock-soil layer shape is more approaching.

Images