A macroscopic modeling method based on the pore structure of porous media
Through macro-angle modeling and discretized parameter functions, the problem of pore-particle boundary processing complexity in pore scale flow simulation of pores and media is solved, and efficient and accurate flow simulation is achieved, reducing the calculation amount and time, and improving the simulation. Accuracy and speed.
Patent Information
- Application Number
- CN202210235422.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-11
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2042-03-11
AI Technical Summary
The existing pore-scale flow simulation methods of porous media need to deal with complex pore-particle boundary problems, resulting in huge calculations, long simulation time and inaccurate results.
The macro-angle modeling method is adopted to construct a discrete parameter function to give different porosity and permeability values to different points in the region, ignore the existence of pore structure, and distinguish pores from particles only through parameter values, thereby avoiding boundary processing problems. Assisted meshing based on CT scan images.
It effectively reduces the calculation amount and simulation time, avoids the complexity at the pore-particle boundary, realizes accurate pore-scale flow simulation, and simplifies the model dissection process, improving the speed and accuracy of numerical calculations.
Smart Images

Figure CN114818408B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a macroscopic angle modeling method based on the pore structure of a porous medium, and belongs to the technical field of pore-scale flow simulation of porous media. Background Art
[0002] Oil and natural gas are important components of my country's energy system. It is of vital importance to develop and utilize oil and gas resources efficiently and rationally. The flow of oil and gas in underground reservoirs can be equivalent to the flow of porous media. In addition, the hydrate mining and geothermal mining that have emerged in recent years can also be considered as the flow process in porous media. Therefore, more and more scholars have conducted a lot of research on porous media. Not only the flow characteristics of porous media are observed through physical experiments, but also the flow mechanism is studied through numerical simulation.
[0003] With the widespread application of computer science, numerical simulation technology has been widely used in the study of flow simulation in porous media. Using numerical simulation technology, the flow law of fluid in porous media at the pore scale can be effectively simulated. And with the development of electronic computer tomography, namely CT (Computed Tomography) technology, it is easy to obtain high-resolution pore structure by scanning reservoir cores. Many scholars simulate the flow of fluid phases during the mining process based on the pore structure scanned by CT. At present, the main flow simulation methods for porous media at the pore scale include pore network model, unstructured grid finite element simulation and lattice-Boltzmann method. However, most of these simulation methods are based on the real pore structure of reservoir cores, and inevitably deal with complex boundary problems, which requires huge calculations.
[0004] The finite element simulation method needs to obtain the grayscale image of the pore structure based on the core CT scan to deal with the complex particle-pore boundary problem. This requires local encryption or even boundary layer processing of the boundaries between pores and particles during the meshing process, which not only increases the number of grids in the solution domain, but also a large number of pore-particle boundaries increase the complexity of solving partial differential equations. On the one hand, it increases the amount of calculation and places too high demands on computer performance; on the other hand, it is very likely that the model will be difficult to converge and the solution results will be inaccurate. The pore network model first obtains the electron microscope image of the core through CT scanning to extract the pore space information of the porous medium. Subsequently, the digital core is reconstructed through corresponding mathematical methods to obtain the geometric structure data of the digital core, including rock pores and rock skeletons. Then the key information of the pore structure of the digital core is extracted, and the pores and channels of the core are replaced by a simple pore network model. However, this method does not take into account the characteristics of fluid flow in microscale channels, and is an idealized model. Relying on a simplified pore network model cannot accurately describe the flow of fluid in porous media at the pore scale. In contrast, the lattice-Boltzmann method is more suitable for pore-scale flow simulation. The lattice-Boltzmann method can be simulated directly on the digital pore structure without solving any partial differential equations. This method assumes that the fluid exists in the form of particles, represents the fluid with the distribution function of the particles, and the fluid flow process includes two processes: particle collision and flow. The principle is relatively simple. However, the boundary treatment of the pores and rock particles of the lattice-Boltzmann method is relatively cumbersome. It is assumed that the fluid particles collide with the rock skeleton, and collision boundary conditions need to be applied to each fluid particle, which requires a large amount of calculation. For example, the Chinese patent document (application number CN201510173922.7) discloses a method and device for simulating the microscopic flow of carbonate rocks, which mainly uses the lattice Boltzmann method to simulate the microscopic flow of carbonate rocks.
[0005] The several pore-scale flow simulation methods discussed above all need to deal with real digital pore structures and cannot avoid dealing with complex pore-particle boundary conditions. Moreover, pore-scale simulation requires extremely fine mesh subdivision, and each mesh node needs to solve a large number of equations, so the amount of calculation is large, the simulation time is long, and the computer performance requirements are high.
[0006] However, if we consider modeling from a macroscopic perspective, we can effectively avoid the problem of boundary processing of porous particles. We only need to assign the required parameter values to the entire area, and different points in the area require different parameter values. For example, flow simulation requires assigning parameter values such as porosity and permeability to the entire area, but the porosity and permeability are generally different at different points in the area. Specifically, the permeability of the part where the particles are located should be assigned an extremely small value, or even 0. Through this discretization assignment method, we only need to assign different parameter values to the entire solution domain to effectively deal with the relationship between the pores and the skeleton.
[0007] Therefore, the present invention proposes a method for assigning discretized functions to simulate pore-scale flow in porous media, which can assign different parameter values to different points in the region from a macroscopic perspective. This can achieve the same results as pore-scale flow simulation, while avoiding the boundary processing problem and huge amount of calculation in pore-scale simulation. Summary of the invention
[0008] In view of the deficiencies of the prior art, the present invention provides a reasonable and effective method that can accurately simulate the pore-scale flow process of porous media and model it from a macroscopic perspective.
[0009] The technical solution of the present invention is as follows:
[0010] A macroscopic modeling method for pore-scale flow simulation in porous media comprises the following steps:
[0011] S1: core samples are obtained by core sampling of the reservoir; core slices are scanned by CT to create a two-dimensional rock grayscale image; black is represented by a value of 0, and white is represented by a value of 1; 0 represents pores and 1 represents rock skeleton particles;
[0012] S2: Based on the rock grayscale image in S1, obtain the spatial information of the pore structure. Construct the image numerical function f(x,y), Where (x, y) is the geometric position coordinate;
[0013] S3: According to the size of the two-dimensional rock grayscale image scanned by CT in S1, a two-dimensional simulation area with the same size as the CT image is established using the finite element method to perform flow simulation;
[0014] S4: Set the global parameter values required for pore-scale flow simulation. Taking porosity and permeability as an example, assume that the global porosity is φ0 and the global permeability is k0;
[0015] S5: In order to avoid considering the complex boundary problems of pores and particles, a discretized parameter function is constructed from a macroscopic perspective. This porous medium region is assumed to be a complete region, and the existence of the pore structure is ignored. It is assumed that there is no complex positional relationship between pores and particles in this region, but that different positions in the region have different parameter values, and the pores and particles are distinguished by different parameter values. Taking porosity and permeability as examples, the porosity and permeability are different at different positions in the region. Therefore, the porosity discretization function is constructed as φ(x,y)=φ0(1-f(x,y)·a), that is, the porosity at the location of the particles is a minimum value close to 0, and the porosity at the location of the pores is φ0; the permeability discretization function is That is, the permeability of the location of the particles is a small value, while the permeability of the location of the pores is extremely large, and the permeability is used to distinguish between pores and particles;
[0016] Among them, a and c are constants. In order to avoid φ=0 or k=∞, constants a and c are introduced; this makes the porosity close to zero and non-zero, and the permeability close to infinity and a calculable value. These two constants can ensure that numerical errors are avoided in numerical calculations. b is a constant. In addition, constant b is to ensure that the permeability of the particle is It is a small value, and the permeability value at this location can be determined by adjusting the size of b.
[0017] Preferably, a is to avoid φ=0, so a is taken as 99% of φ0, c is to avoid k=∞, so c is taken as one ten-thousandth of k0, and b is to ensure that the permeability at the particle is a smaller value, so b is taken as 10% of k0.
[0018] S6: Through S5, the positional relationship between pores and particles has been successfully distinguished, and the geometric position information of the pore structure has been determined by means of a discretized assignment function; then, the entire two-dimensional rock grayscale image area where the model is located is meshed, and the mesh is assisted by the numerical function based on the grayscale image. The area where f(x,y) is 0 needs to be refined, while the area where f(x,y) is 1 is the particle area and does not participate in the calculation, so the mesh can be coarsened. This can greatly save the number of meshes in the area and the time of simulation calculation, and effectively avoid the complex problem of meshing on the pore-particle boundary.
[0019] Coarsening the model's grid and refining the grid are basic concepts in the field of fluid mechanics computer simulation, which refers to the number of grids used to characterize the constructed model involved in the calculation. Considering that different models have different sizes, the coarsening and refining in this application are relative concepts proposed in the same model and the same scenario. In this step of the present invention, it means that the number of grids divided in the area where f(x,y) is 0 is greater than the number of grids divided in the position where f(x,y) is 1.
[0020] S7: Apply the Brinkman equation of flow simulation to the simulation area established in S3, set the initial boundary conditions, and solve the velocity field and pressure field of the flow process in the grid divided in S6; through S1-S7, that is, modeling from a macroscopic perspective to realize the pore-scale flow simulation equation. Note: The Brinkman flow equation mentioned in S7 is a mature theoretical equation, which does not belong to the content of the present invention and is not elaborated here.
[0021] Preferably, the initial boundary conditions in S7 are: inlet boundary P=P0, outlet boundary P=0, and the other two boundaries are set u·n=0, where u is the velocity vector, n is the normal vector, and u·n=0 means that there is no flow velocity perpendicular to the boundary, that is, there is no flow boundary.
[0022] Further preferably, the P0 value is 0.715Pa.
[0023] When simulating microscopic flow, the present invention does not process the complex boundary relationship between particles and pores, but uses a discretized numerical function to assign different values in different regions, thereby realizing microscopic boundary processing from a macroscopic modeling perspective.
[0024] The beneficial effects of the present invention are:
[0025] 1. The macro-angle modeling method for pore-scale flow simulation of porous media proposed in the present invention includes constructing an image numerical function of a rock slice CT scan image, a discretization parameter assignment function for the entire region, and an image numerical function-assisted regional grid division method. The invention can truly restore the distribution of rock particle pore channels based on CT scan images, obtain a real porous medium physical model, and discretize and assign simulation parameters in the real porous medium region. Compared with the complexity of pore-scale simulation, the macro-modeling method proposed in the present invention avoids the problem of pore-scale boundary processing of particles and pores, and greatly eliminates the uncertainty caused by this boundary problem in numerical calculation. Moreover, macro-angle modeling reduces the number of grids for model division, and avoids the problem of local encrypted and refined grid division at the pore-particle boundary. That is, the macro-angle modeling method for porous media involved in the present invention can greatly reduce the amount of calculation and shorten the time of numerical simulation. It can achieve the results of pore-scale flow simulation and avoid the shortcomings of pore-scale simulation.
[0026] 2. The macroscopic modeling method proposed in the present invention can be used to study the flow simulation of porous media at the microscopic scale, to study the influence of the distribution of pore channels at the pore scale on the velocity field of the flow process, and to further study the distribution of microscopic streamlines during the flow process, and to calculate parameter values such as tortuosity and permeability.
[0027] 3. The macroscopic modeling method involved in the present invention greatly reduces the amount of calculation for numerical simulation of porous flow, and has the advantages of fast numerical calculation speed and high precision. The proposed parameter discretization function can effectively assign corresponding parameters to the model according to the actual pore distribution of porous media. The principle is simple and easy to operate. Therefore, this method can be effectively applied to the study of porous media flow simulation, and provide technical support for porous media flow research from the perspective of numerical simulation. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 It is a technical flow chart of the present invention.
[0029] Figure 2 2D grayscale image of a core sample scanned by CT.
[0030] Figure 3a , Figure 3bAssignment diagram of discretized parameter functions for porosity and permeability. Figure 3a is the discretized value of the porosity in the model area; Figure 3b is the discretized value of the permeability in the model area.
[0031] Figure 4 Initial boundary conditions set for the present invention.
[0032] Figure 5 The flow results simulated for the invented porous media macroscopic modeling method. DETAILED DESCRIPTION
[0033] In order to enable those skilled in the art to better understand the present invention, the present invention is further described below through embodiments in conjunction with the accompanying drawings, but the present invention is not limited thereto.
[0034] The present invention first obtains reservoir core samples by core sampling, and obtains grayscale images of the core samples by CT scanning technology. After obtaining the geometric position confidence of the pore structure, an image numerical function is constructed. In order to avoid considering the complex boundary problem of pores and particles, a discretized parameter function is constructed from a macroscopic perspective. It is assumed that there is no distinction between pores and rock skeletons in the entire area, only the corresponding parameter values are different. Finally, the regional grid is solved by auxiliary segmentation based on the numerical function of the grayscale image, and the flow equation is calculated. The pore-scale flow simulation of porous media can be realized by modeling from a macroscopic perspective, and the next step of research can also be carried out on the basis of this method.
[0035] Example 1
[0036] In order to further illustrate the effectiveness of the technical method, the macroscopic modeling method of porous media proposed in the present invention is further explained by taking the CT scanning image of a real core sample as an example. Figure 1 It can be seen from the technical route flow chart that the specific steps of the present invention are as follows:
[0037] S01: Perform CT scanning on real reservoir core samples to obtain two-dimensional rock scanning grayscale images (see Figure 2 ).
[0038] S02: Based on the grayscale image of the core scan in S01, obtain the spatial information of the pore structure. Construct the image numerical function f(x,y), Where (x, y) is the geometric position coordinates.
[0039] S03: According to the size of the CT image (360 μm×360 μm), a two-dimensional simulation area (360 μm×360 μm) of the same size as the CT image is established using a conventional finite element method to perform flow simulation.
[0040] S04: Initialize the global parameter values required for pore-scale flow simulation. For the sake of convenience, the two basic parameters of porosity and permeability are taken as examples. Assume that the initial porosity φ0 = 1 and the permeability k0 = 1000 mD;
[0041] S05: In order to avoid the complex boundary problem between pores and particles, a discretized parameter function is constructed from a macroscopic perspective. It is assumed that there is no complex positional relationship between pores and particles in the entire area of the model, but that different positions in the area have different parameter values, and the pores and particles are distinguished by the different parameter values. Taking porosity and permeability as examples, only porosity and permeability are used as examples here. In actual operation, the discretized parameter function can be selectively used for the parameters involved;
[0042] Assuming that the porosity and permeability in the model solution area vary with location, the porosity discretization function constructed is φ(x, y) = 1-0.99×f(x, y), that is, the porosity at the location of the particle is 0.01, and the porosity at the location of the pore is 1 (see Figure 3a ); The constructed permeability discretization function is That is, the permeability of the particles is about 10 millidarcy, and the permeability of the pores is 10,000 millidarcy (see Figure 3b ); particles and pores are distinguished by the size of the porosity and permeability parameter values in the entire area of the model. Note: The specific constant values set for a and b in this step are not fixed. They are only used as a reference here. More appropriate and realistic values can be assigned according to the actual physical background.
[0043] S06: Through S05, the positional relationship between pores and particles has been successfully distinguished, and the geometric position information of the pore structure has been determined by means of a discretized parameter assignment function. Then, the mesh is assisted by the grayscale image numerical function f(x,y), and the mesh is refined at the position where f(x,y)=0. The non-parametric model calculation of the part where f(x,y)=1 can coarsen the mesh. This can greatly reduce the number of meshes in the model area and the time for numerical calculation, and also avoid the processing of the particle-pore boundary.
[0044] S07: Apply the Brinkman equation for flow simulation to the simulation area established in S03 and set initial boundary conditions (such as Figure 4 As shown in the figure, the inlet boundary P = P0, the outlet boundary P = 0, and the other two boundaries are set to u·n = 0, where u is the velocity vector and n is the normal vector. u·n = 0 means that there is no flow velocity perpendicular to the boundary, that is, there is no flow boundary. The value of P0 is 0.715Pa. The velocity field and pressure field of the flow process are solved in the grid divided by S06. The velocity field distribution solved by S01-S07 is shown in Figure 5 .Depend on Figure 5It can be seen that the flow process occurs in the pore roar, and the flow channel can be clearly seen in the figure.
[0045] This implementation case further verifies that the discretized parameter assignment function proposed in the present invention realizes the flow simulation process of porous media pore scale from a macroscopic perspective. This method is suitable for pore scale flow simulation of porous media, which can not only cleverly avoid dealing with complex pore-particle boundary problems, but also reduce the number of grids for model segmentation and reduce calculation time.
[0046] Based on the macroscopic modeling method provided by the present invention, further research can be carried out. After realizing the pore-scale flow simulation, the streamlines of the model can be drawn and the parameters such as tortuosity and permeability can be calculated. Therefore, the method provided by the present invention has excellent universality.
[0047] It should be noted that the above examples are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that, in view of the simple operation of the discretized parameter function and the macro-angle modeling method proposed in the present invention, relevant personnel in the field can still modify the technical solutions recorded in the aforementioned embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A macroscopic modeling method based on the pore structure of porous media, characterized by: The method comprises the following steps: S1: Create a two-dimensional rock grayscale image according to the preset method; S2: construct an image numerical function f(x,y) based on the two-dimensional rock grayscale image obtained in S1 according to a preset method; S3: Based on the size of the two-dimensional rock grayscale image scanned by CT in S1, a two-dimensional simulation area with the same size as the CT image is established using the finite element method; S4: Set the global parameter values required for pore-scale flow simulation, setting the global porosity to φ0 and the global permeability to k0; S5: According to the preset method, a discretized parameter function is constructed from a macroscopic perspective; specifically, the following steps are included: The porosity discretization function is constructed as φ(x,y)=φ0(1-f(x,y)·a), and the permeability discretization function is Among them, a, b, c are constants; S6: S5 can be used to successfully distinguish the geometric position relationship between pores and particles by discretizing the assignment function, and then the entire two-dimensional rock grayscale image area is assisted in meshing based on the grayscale image numerical function f(x, y) according to the preset method; S7: Apply the Brinkman equation for flow simulation to the simulation area established in S3, set the initial boundary conditions, and solve the velocity field and pressure field of the flow process within the grid divided by S6; through S1-S7, that is, modeling from a macroscopic perspective, the pore-scale flow simulation equation is realized.
2. The macroscopic modeling method based on the pore structure of porous media according to claim 1, characterized in that: The step S1 of creating a two-dimensional rock grayscale image according to a preset method specifically includes the following steps: Core sampling is carried out on the reservoir to obtain reservoir core samples; rock thin sections are scanned using CT to create a two-dimensional rock grayscale image; in which black is represented by a value of 0 and white is represented by a value of 1; 0 represents the pore throat and 1 represents the rock skeleton particles.
3. The macroscopic modeling method based on the pore structure of porous media according to claim 1 is characterized in that: The image numerical function f(x, y) constructed in S2 is specifically: Based on the core scanning grayscale image in S1, construct the image numerical function f(x,y), Where (x, y) is the geometric position coordinates.
4. The macroscopic modeling method based on the pore structure of porous media according to claim 1, characterized in that: a is 99% of φ0, c is 1 / 10,000 of k0, and b is 10% of k0.
5. The macroscopic angle modeling method based on the pore structure of porous media according to claim 1 is characterized in that: The S6 specifically includes the following steps: The numerical function-assisted meshing based on the two-dimensional rock grayscale image is used. The area where f(x, y) is 0 needs to be refined, while the area where f(x, y) is 1 is the granular area and the mesh needs to be coarsened.
6. The macroscopic angle modeling method based on the pore structure of porous media according to claim 5 is characterized in that: The number of grids in the area where f(x,y) is 0 is greater than the number of grids in the area where f(x,y) is 1.
7. The macroscopic angle modeling method based on the pore structure of porous media according to claim 1 is characterized in that: The initial boundary conditions in S7 are: inlet boundary P=P0, outlet boundary P=0, and the other two boundaries are set u·n=0, where u is the velocity vector and n is the normal vector. u·n=0 means that there is no flow velocity perpendicular to the boundary, that is, there is no flow boundary.
8. The macroscopic angle modeling method based on the pore structure of porous media according to claim 7 is characterized in that: The P0 value is 0.715Pa.
