Method for generating large-scale heterogeneous pore-throat network jointly controlled by macro and micro parameters

Abstract

The present invention discloses a method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters, which includes partitioning a set simulation area, randomly generating a plurality of mineral particle cores in each partition, simulating the growth of the mineral particle cores by using a four-parameter random growth method, regarding the boundaries of the minerals of each mineral particle core as potential pores and throats, and inversely obtaining the position distribution of each pore and throat; using a genetic algorithm to extract and optimize the position distribution of the pores and throats, matching a coordination number for each pore and throat, and assigning a target growth radius value to the pores and throats with each matching coordination number by setting the micro pore-throat structure parameters of each partition; using the four-parameter random growth method to simulate the corrosion of the fluid in the pores and throats in each partition towards the mineral particle cores until the radii of the pores and throats in each partition reach the target growth radius values one by one, thereby generating a pore-throat network image of the set simulation area.

Description

Technical Field

[0001] The invention relates to the technical field of oil development, and in particular to a method for generating a large-scale heterogeneous pore throat network by jointly controlling macroscopic and microscopic parameters. Background Art

[0002] With the updating of instruments and equipment and the development of observation methods, microfluidics technology has gradually become an important experimental means in the field of oil and gas geology to deepen the understanding of seepage laws and expand the explanation of mechanisms due to its outstanding advantages of microscopic visualization and convenience.

[0003] At present, physical simulation methods at the micro-nano scale are mainly explored from two perspectives: micro-local and macro-overall. The former focuses on phase changes and fluid storage types in pore throat structures, while the latter focuses on the macroscopic distribution characteristics of fluids in complex porous media structures. The comprehensive application of the two can more systematically clarify the seepage mechanism of reservoir fluids, more effectively evaluate the remaining oil utilization capacity and recovery potential, and provide theoretical support and experimental basis for the adjustment of measures in the later stage of oilfield development.

[0004] Among these, designing and generating models that can truly reflect the microscopic pore structure characteristics of rocks is crucial for microscopic physical simulation experiments and cross-scale numerical simulations, and is the basis for conducting related research.

[0005] Previously, the generation of microscopic pore-throat network model patterns was often based on cast thin sections or CT scan images, and on the basis of extracting the pore-throat structure, it was completed through image stitching and denoising techniques. However, due to the limitation of image accuracy, the size of the images obtained by cast thin sections and CT scans are small, and the range of pore-throat structural features that can be displayed is limited. The microscopic model formed by the stitching design can obtain larger-scale images, but it is still far from the research scale of the core and the macroscopic level of the mine, and its repetitive structure lacks theoretical support, and the reliability of the upscaling application of the results is questionable.

[0006] At present, there are two generation methods for large-scale pore-throat network modeling: equivalent modeling and random modeling. For equivalent modeling, even if the pore and throat size distributions are similar and the porosity is similar, if the spatial topological structure is quite different, the physical parameters such as permeability of the pore-throat network model will still show large differences. The random modeling method can only consider two constraints, namely porosity and correlation function, resulting in poor connectivity of the generated model. The traditional four-parameter random growth QSGS method is further developed on the basis of random modeling, simulating the growth process of rock minerals and can effectively control the anisotropy of the model. However, the constraint condition only considers the single factor of porosity, and the simulation degree of the pore throat morphology and connectivity of the model is relatively poor. Summary of the invention

[0007] The object of the present invention is to provide a method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters, so as to solve the problem in the prior art that the existing modeling methods cannot meet the requirements of fast speed, high simulation degree and generation of large-scale pore-throat network images.

[0008] To solve the above technical problems, the present invention specifically provides the following technical solutions:

[0009] A method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters, comprising:

[0010] Step 100: Divide the set simulation area into partitions, randomly generate a plurality of mineral particle cores in each partition, determine the number of particle cores according to the target porosity and the ideal pore-throat model, simulate the growth of the mineral particle cores by using the four-parameter random growth method, and regard the interfaces of the minerals of each mineral particle core as potential pores and throats, and invert the position distribution of each pore and throat;

[0011] Step 200: Use the genetic algorithm to extract and optimize the position distribution of pores and throats, match the coordination number for each pore and throat, and assign a target growth radius value to the pores and throats with each matching coordination number by setting the micro pore-throat structure parameters of each partition;

[0012] Step 300: Use the four-parameter random growth method to simulate the corrosion of the fluid in the pores and throats in each partition towards the mineral particle cores until the radii of the pores and throats in each partition reach the target growth radius value one by one, and generate a pore-throat network image of the set simulation area.

[0013] As a preferred solution of the present invention, control parameters are obtained through experiments on the target core sample, and the control parameters include phase permeability partition, porosity, pore radius distribution, throat radius distribution, grain size radius distribution and coordination number distribution;

[0014] Distinguish the constraint parameter system of the large-scale pore-throat network generation method based on the four-parameter random growth method through the phase permeability partition, and establish different micro pore constraint parameter systems in each partition;

[0015] Set the porosity parameter to control and constrain the final volume distribution of mineral particles and pore throats;

[0016] Set the pore radius distribution, throat radius distribution, grain size radius distribution and coordination number distribution to control and constrain the micro morphology and structure of mineral particles, pores and throats;

[0017] As a preferred solution of the present invention, use the ideal model of the pore-throat network of regular hexagons to determine the number of randomly generated mineral particle cores in each partition, specifically including:

[0018] Assume that the pores are spherical and the throats are rod-shaped. When the entire regional space is filled with this basic structure, calculate the number of pores and the number of mineral particle cores in the simulation area under ideal conditions;

[0019] Based on the calculated values of the ideal model, the finally determined number of pores and the number of particles will be adjusted and optimized according to the ratio of the number of particles to the number of pores in the real core.

[0020] As a preferred embodiment of the present invention, regarding each randomly generated mineral particle core as a kind of mineral and setting the corresponding fractional porosity for each kind of mineral can achieve more precise control;

[0021] And to speed up the calculation speed, the interaction between mineral particle cores can be optionally ignored during the process of simulating the growth of mineral particle cores.

[0022] As a preferred embodiment of the present invention, according to the obtained position distribution of each pore and throat, use the connected domain algorithm to identify and extract the intersection lines and the intersection points of the intersection lines between mineral particle cores in the position distribution result, define the intersection lines as throats, and define the intersection points of the intersection lines as potential pores.

[0023] As a preferred embodiment of the present invention, based on the coordination number obtained from the experiment of the target core sample, calculate the fitness of the pores and throats in each partition to the target coordination number through the genetic algorithm, and delete the pores or throats in the partition through the fitness to complete the optimization of the pores and throats in each partition;

[0024] Among them, the fitness calculation function of the genetic algorithm is:

[0025]

[0026] Among them, n is the number of partitions in the model area, m is the total number of pores extracted in each partition, c i,j is the coordination number of the j-th pore in the i-th partition calculated according to the individual chromosome, c t i,j is the target coordination number distribution of all pores in each partition, sort(c i,j ) and sort(c t i,j ) are sorting functions; sum is the summation function, and the summation function is set as a penalty term to prevent the genetic algorithm from setting pores as dead pores.

[0027] As a preferred embodiment of the present invention, by setting the microscopic pore-throat structure parameters of each partition, a distribution rule for assigning target growth radius values to the pores and throats that match the coordination number is set. The distribution rule specifically includes:

[0028] For the pores and throats within each partition, random sampling is performed according to the given pore and throat radii;

[0029] The higher the extracted pore area and the more pores are merged, the larger the pore radius is assigned;

[0030] The radius of the throat cannot exceed the minimum pore radius among the pores at both ends of the partition.

[0031] As a preferred embodiment of the present invention, the four-parameter random growth method is used to simulate the corrosion of the fluid in the pores and throats in each partition towards the mineral particle core. During the introduction of the fluid corrosion process:

[0032] The total pore area is set as the primary pores between mineral particles and the intergranular epigenetic corrosion region;

[0033] The percentage of the area of the epigenetic corrosion region between adjacent mineral particle cores in the total pore area is set as the corrosion threshold, and the corrosion stops when any pore reaches the corrosion threshold.

[0034] As a preferred embodiment of the present invention, after the pore and throat radius distribution of each partition is completed, an initial pore-throat network image is formed. The specific method for thickening the throat using the image dilation algorithm includes:

[0035] Obtain the size of the initial image and create a completely black image of the same size;

[0036] Traverse each pixel and determine if there is an intersection of non-zero common elements between the two images;

[0037] After the traversal is completed, output the target image.

[0038] The present invention has the following beneficial effects compared with the prior art:

[0039] The present invention combines the advantages of the equivalent modeling and random modeling algorithms, inversely simulates the geological formation of pores and throats, simulates the mineral growth and pore corrosion processes, and controls the generation process with the distribution of four parameters: porosity, coordination number, pore radius, and throat radius at each stage, realizing high-fidelity, large-scale, controllable, and heterogeneous pore-throat network modeling. Among them, the improved genetic algorithm is used for the fitting of the coordination number distribution, which has the advantages of shorter convergence time and faster calculation speed. Brief Description of the Drawings

[0040] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only exemplary, and for those of ordinary skill in the art, without creative efforts, other implementation drawings can also be obtained based on the provided drawings.

[0041] Figure 1 This is the core cast thin section image and pore-throat network image of X reservoir for this embodiment;

[0042] Figure 2 This is the process flow of the pore-throat inversion method for this embodiment;

[0043] Figure 3 This is the flow chart of rock mineral growth for this embodiment;

[0044] Figure 4 This is the result of simulating mineral growth and interface extraction for this embodiment;

[0045] Figure 5 This is the comparison chart of the implementation of the dilation algorithm for this embodiment;

[0046] Figure 6 This is the principle of the genetic algorithm for this embodiment;

[0047] Figure 7 This is the flow chart of the implementation of the dilation algorithm for this embodiment;

[0048] Figure 8 This is the comparison chart before and after the growth of pore growth and dissolution for this embodiment;

[0049] Figure 9 This is the schematic diagram of distance transformation in this embodiment;

[0050] Figure 10 This is the algorithm flow chart of pore growth and dissolution for this embodiment;

[0051] Figure 11 This is the result of pore-throat network generation under the gradual change of porosity for this embodiment;

[0052] Figure 12 This is the result chart of pore-throat network generation under the gradual change of porosity for this embodiment;

[0053] Figure 13 This is the result of pore-throat network generation with various image sizes under the custom partition in this embodiment;

[0054] Figure 14 These are the microscopic pore-throat structure parameters of the X reservoir core obtained in this embodiment;

[0055] Figure 15Schematic diagram of pore ranges under different coordination numbers in this embodiment. Specific implementation mode

[0056] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0057] As Figures 1 to 13 shown, this embodiment provides a method for generating a large-scale heterogeneous pore-throat network with co-control of macro and micro parameters, which includes two growth processes in total:

[0058] The first growth process: The formation of pores and throats is inversely deduced through mineral growth to determine the positions of pores and throats. During this process, relative permeability zoning and porosity are introduced to control the number and positions of pores and throats.

[0059] After determining the distribution positions of pores and throats, an improved genetic algorithm is used to optimize the distribution of throats, further match the distribution of the target coordination number, and assign radius values to each pore and throat according to the micro pore-throat structure parameters of each zone according to the target micro pore-throat structure parameter distribution.

[0060] The second growth process: The formation process of pores is inversely deduced, that is, while minerals grow into the intergranular space, they are continuously washed, shed, and dissolved by the fluid. By finely controlling the formation process of each pore, the pores have specified sizes and complex and diverse shapes.

[0061] Its specific implementation steps include:

[0062] Step 100: Divide the set simulation area into zones, randomly generate multiple mineral particle cores in each zone, determine the number of particle cores in each zone according to the porosity and the ideal pore-throat model, simulate the growth of mineral particle cores using the four-parameter random growth method, and regard the boundary of each mineral particle core mineral as potential pores and throats, and inversely obtain the position distribution of each pore and throat;

[0063] Step 200: Use the genetic algorithm to extract and optimize the position distribution of pores and throats, match the coordination number for each pore and throat, and assign the target growth radius value to each pore and throat with a matched coordination number by setting the micro pore-throat structure parameters of each zone;

[0064] Step 300: Use the four-parameter random growth method to simulate the dissolution of the fluid in the pores and throats in each partition towards the mineral particle core until the radii of the pores and throats in each partition reach the target growth radius values one by one, and generate the pore-throat network image of the set simulation area.

[0065] For the experiment on the target core sample, obtain the microscopic pore-throat structure parameters and the phase permeability partition. The microscopic pore-throat structure parameters include porosity, pore radius distribution, throat radius distribution, grain size radius distribution, and coordination number distribution.

[0066] Distinguish the partitions in the set simulation area through the phase permeability partition, and correspondingly establish the microscopic pore-throat structure parameters in each partition according to the distinction results.

[0067] Among them, match the coordination numbers for the pores and throats in each partition through the coordination number distribution.

[0068] Control the target growth radii of the pores and throats in the partition through the pore radius distribution and throat radius distribution; control the grain size radius of the mineral particle core in the partition through the grain size radius distribution.

[0069] Use the above to obtain the representative microscopic pore-throat characteristics of the target reservoir rock:

[0070] As Figure 1 shown, the core samples in this article are taken from the X reservoir, which are grayish-brown fine-grained sandstone and grayish-brown medium-fine-grained sandstone respectively, Figure 1 (a) and Figure 1 (c) The thin section images of the cast specimens.

[0071] Figure 1 (a) 34.45% of the pore diameters are less than 100 μm, and the rest of the pore diameters are greater than 100 μm. Moreover, the pores in the size ranges of 75 μm - 100 μm and 100 μm - 125 μm are in the majority. The radii of the main flow throats are mostly between 1 - 10 μm, belonging to a homogeneous reservoir with large pores and fine throats.

[0072] Figure 1 (c) 61.30% of the pore diameters are less than 100 μm, and the pores in the size ranges of 50 μm - 75 μm and 75 μm - 100 μm are in the majority. The rest of the pore diameters are greater than 100 μm, and about 89.03% of the radii of the flow throats are between 1 - 10 μm, belonging to a homogeneous reservoir with medium pores and fine throats.

[0073] For Figure 1 (a) and Figure 1 (c), perform binarization processing through the color threshold segmentation method, extract the blue pore part, and along the edge of the mineral particles, complete the extraction of the throats with light blue color, and obtain as Figure 1(b) and Figure 1 the pore-throat network images shown in (d).

[0074] Based on the pore-throat network images of the cast thin sections or CT digital cores, the microscopic pore-throat structure parameter characteristics including porosity, coordination number distribution, pore radius distribution, and throat radius distribution can be statistically analyzed. For the specific data rules, see Figure 13 .

[0075] The number of pores and the number of mineral grains in each partition are the main parameters that need to be obtained through calculation.

[0076] By using the phase permeability partition to constrain the growth density (quantity) of the mineral grain cores simulated by the four-parameter random growth method, the mineral grain cores with different diameter distributions in each partition can be obtained;

[0077] The porosity parameter is the percentage of the final pore-throat volume (area) in the total volume (area) of the region in the model.

[0078] Using the ideal model of the pore-throat network of regular hexagons to calculate the number of randomly generated mineral grain cores in each partition of the simulation region, specifically including:

[0079] It is agreed that the pores are circular (spherical) and the throats are rectangular (rod-shaped), and the number of pores and the number of mineral grain cores in the simulation region under ideal conditions are calculated;

[0080] Furthermore, based on the calculated values of the ideal model, the finally determined number of pores and the number of grains will be adjusted and optimized according to the ratio of the number of grains and pores in the real core.

[0081] The pore-throat area of the ideal model satisfies the formula:

[0082] πNR p 2 +2cNR t (L - 2R p ) = φS;

[0083] N is the number of pores, Rp is the average pore radius, Rt is the average throat radius, L is the average pore spacing, c is the average coordination number, φ is the porosity, and S is the area of the partition.

[0084] Considering randomly distributing N pores in a region with an area of S, when ignoring the throat area, there is:

[0085] NL 2 ≈ S ≈ πNR p 2 / φ;

[0086] However, in the algorithm, the pores that are relatively close to each other will be merged into one, so the number of grains is often less than 2 / 3 of the number of pores.

[0087] In this embodiment, based on a large number of models, this parameter is set to 0.78, that is, the number of pores is 1.3 times the number of mineral particle cores. At this time, the mineral particle cores and pores are more natural and realistic with a loose and orderly arrangement.

[0088] This embodiment only assumes that the pore structure of the rock can be approximated as an ideal hexagonal pore-throat network, but it can also be replaced by other ideal unit forms.

[0089] Using Figure 13 the microscopic pore-throat structure data of the medium-fine grained sandstone core samples, the obtained mineral distribution map is as shown in Figure 4 (a) and Figure 4 (b).

[0090] Among them, each randomly generated mineral particle core in the partition is regarded as a kind of mineral, and the interaction between mineral particle cores is ignored during the process of simulating the growth of mineral particle cores using the four-parameter random growth method.

[0091] According to the obtained position distribution of each pore and throat, the boundary line and the intersection points of the boundary line between mineral particle cores in the position distribution result are identified and extracted using the connected domain algorithm, and the boundary line is defined as the throat, and the intersection points of the boundary line are defined as potential pores.

[0092] When the length of the throat is less than the value l, it is changed into a pore. Let the equivalent radius of the pore be Rp and the pore area be Sp, where:

[0093] S p = πR p 2 ;

[0094] When the coordination number is 3, as shown in Figure 15 (a), then: l / Rp = 2.

[0095] When the coordination number is 4, as shown in Figure 15 (b), assuming that the distance between pore 1 and pore 2 is greater than l, and a pair of pores has been merged in pore 1 and pore 2, and its final pore shape is a square and two semi-circles. At this time, it satisfies: l / Rp = 1.33 (equivalent radius).

[0096] Therefore, it can be considered that the more pores are merged, the smaller the value of l / Rp, and at the same time, in the pore radius distribution, this pore tends to be assigned a larger value.

[0097] In a real core, although there are obvious boundaries between some mineral particles, no channels for fluid to pass through are formed. Therefore, these mineral particles and the surrounding mineral particles can be regarded as one particle.

[0098] In other words, some pores or throats can be deleted. By selecting appropriate pores or throats for deletion, the distribution of coordination number and grain size can be better fitted.

[0099] In this paper, the genetic algorithm is adaptively improved. The numerical value originally input into the genetic algorithm in decimal is changed to binary, where 0 represents deletion and 1 represents retention. The length of the binary is the number of throats in the model.

[0100] The flow chart of the genetic algorithm is as Figure 6 shown. Among them, roulette wheel selection is used for natural selection, and single-point crossover operator is used for the gene recombination part.

[0101] In the improved genetic algorithm, each individual has only one chromosome, which is composed of 0 and 1. The numerical value at each position represents whether the throat corresponding to the number is deleted.

[0102] Based on the coordination number obtained from the experiment of the target core sample, the fitness of the pores and throats in each partition to the target coordination number is calculated by the genetic algorithm. The pores or throats in the partition are deleted through the fitness, and the optimization of the pores and throats in each partition is completed;

[0103] Among them, the fitness calculation function of the genetic algorithm is:

[0104]

[0105] Among them, n is the number of partitions in the model area, m is the total number of pores extracted in each partition, c i,j is the coordination number of the j-th pore in the i-th partition calculated according to the individual chromosome, c t i,j is the target coordination number distribution of all pores in each partition, sort(c i,j ) and sort(c t i,j ) are sorting functions; sum is the summation function, and the summation function is set as a penalty term to prevent the genetic algorithm from setting pores as dead pores.

[0106] By setting the microscopic pore-throat structure parameters of each partition to assign target growth radius values to the pores and throats that match each coordination number, the assignment rules are as follows:

[0107] For the pores and throats in each partition, random sampling is performed according to the given pore-throat radius;

[0108] The higher the area of the extracted pores and the more pores merged, the larger the pore radius is assigned;

[0109] The radius of the throat cannot exceed the minimum pore radius among the pores at both ends of the partition.

[0110] The fluid in the pores and throats in each partition is simulated to corrode towards the mineral particle cores by using the four-parameter random growth method. During the introduction of the fluid corrosion process:

[0111] The area of the pores is set to include the voids between mineral particles and the mineral particle corrosion regions;

[0112] Among them, the percentage of the area between adjacent mineral particles in the void area is set as the corrosion threshold;

[0113] Among them, if the area of a certain mineral particle core reaches the corrosion threshold, the space left by the adjacent mineral particle cores during growth is not included in the intergranular range of the current mineral particle core.

[0114] After the pore and throat radii in each partition are allocated, an initial pore-throat network image is formed. The specific method of thickening the throats by using the image dilation algorithm includes:

[0115] Obtain the size of the initial image and create a completely black image of the same size;

[0116] Traverse each pixel and determine that there are non-zero common element intersections between the two images;

[0117] After the traversal is completed, output the target image.

[0118] According to the allocated pore and throat radii, the QSGS algorithm is used to simulate the erosion effect of the fluid on the rock mineral particles.

[0119] In this process, the distance transformation is used to determine the distance between each pixel point in the image and the pores or throats, and the possible erosion range is determined.

[0120] In order to find the position distribution of the pores in the image, it is necessary to use the distance transformation to determine the distance between each pixel point in the image and the pores or throats;

[0121] Define the zero point set on the image as G = {(x1, y1), (x2, y2), …, (x n , y n )}, then the distance transformation function is:

[0122]

[0123] In the formula, i and j are the positions of points in the image, and x and y are the positions of any point in the zero points, that is, (x, y) ∈ G.

[0124] The distance transformation results Dp and Dt of the pores and throats are calculated respectively, and the final comprehensive transformation result Dc is:

[0125] Dc = 0.5D p + 0.5D t ;

[0126] The image obtained through the above formula is as shown in Figure 9 (a) and Figure 9 (b). The brighter the color in the image, the farther away from the pores and throats.

[0127] When the truncation distance is larger, the area of the pores is larger, and the area of adjacent mineral grains is smaller, indicating more severe erosion. When the truncation distance is larger, the area of the pores is larger, and the area of adjacent mineral grains is smaller, indicating more severe erosion.

[0128] For the simulation of the pore-throat structure generated by fluid erosion, the four-parameter stochastic growth method (QSGS) is used to simulate the relative growth of the liquid phase to the rock phase. The fluid in the pores starts to dissolve the particles continuously until the radius of each pore and throat reaches the assigned value and the porosity reaches the predetermined value, and the erosion expansion stops.

[0129] During the mineral growth inversion process, the pores with different numbers continuously fill the blank space through the four-parameter stochastic growth method, and the growth probability is 1.0. During the fluid dissolution process, the fluid in the pores still "grows" into the minerals through the four-parameter stochastic growth method.

[0130] Generate multiple microscopic pore-throat network images, count their particle sizes, and select the result with the particle size radius distribution range closest to the target particle size radius distribution.

[0131] Furthermore, evaluate the method provided in this embodiment:

[0132] In terms of parameter similarity comparison:

[0133] According to the data representation method and storage method, the pore structure characteristic parameters can be divided into two categories: numerical type and vector type.

[0134] The numerical type parameters include porosity, and the vector type parameters include pore radius distribution, throat radius distribution, particle size distribution, and coordination number distribution.

[0135] These parameters can characterize the characteristics of the microscopic pore-throat network, are easy to count from the images, and do not rely on any empirical formulas.

[0136] For the numerical type of microscopic pore-throat structure parameters, linear normalization is used to calculate the similarity.

[0137] For the vector type of microscopic pore-throat structure parameters, the distribution frequency difference between the target generation and the result generation in each partition is used as the calculation method for the similarity of each partition.

[0138] The comprehensive similarity of the model is the weighted average of the similarities of each partition, which is used as the evaluation of the authenticity of the pore structure characteristic parameters of the model.

[0139] According to the calculation results, the comprehensive similarities between the generated model and the real model can both reach over 80%, among which the comprehensive similarities of the numerical parameters are all over 90%. And the similarity in the throat radius distribution is as high as 99.6%.

[0140] Partition control analysis:

[0141] The pore-throat network random model realizes the control of microscopic pore-throat structure characteristic parameters such as porosity, pore-throat radius, and coordination number during the model generation process. Figure 11 It is the generation result of the pore-throat network random model, and the result reflects the gradual change effect from top to bottom in partitions.

[0142] The pores at the upper part of the image are smaller, and the rock particles are smaller.

[0143] The pores at the lower part of the image are larger, and the rock particles are larger.

[0144] From top to bottom of the image, it reflects a relatively uniform gradual change, reflecting the macroscopic heterogeneity of the pore-throat network model.

[0145] Such as Figure 12 As shown, according to the pore-throat network model generation algorithm designed in this paper, the parameters such as porosity, pore coordination number, pore radius, throat radius, and grain size of the pore-throat network are respectively constrained, and the pore-throat network models under the gradual change of these parameters can be obtained. Here, the gradual change of porosity is taken as an example.

[0146] According to the actual needs of researchers for the model, a custom partition map is provided to generate pore-throat network models of various image sizes, such as Figure 13 as shown.

[0147] The model generated by the method provided in this embodiment is highly consistent with the target core in terms of microscopic pore-throat structure and macroscopic physical properties parameters, and has better simulation performance for the macroscopic-microscopic physical structure of the real core.

[0148] In addition, based on the partition function, a pore-throat network model corresponding to the isoline map established according to the geological model can be generated.

[0149] The above embodiments are only exemplary embodiments of the present application and are not used to limit the present application. The protection scope of the present application is defined by the claims. Those skilled in the art can make various modifications or equivalent replacements within the essence and protection scope of the present application, and such modifications or equivalent replacements should also be regarded as falling within the protection scope of the present application.

Claims

1. A method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters, characterized in that, Including: Step 100: Partition the set simulation area, randomly generate multiple mineral particle cores in each partition, simulate the growth of the mineral particle cores using the four-parameter random growth method, regard the boundaries between the minerals of each mineral particle core as potential pores and throats, and inversely obtain the position distribution of each pore and throat; Step 200: Use the genetic algorithm to extract and optimize the position distribution of the pores and throats, match the coordination number for each pore and throat, and assign the target growth radius value to the pores and throats with each matched coordination number by setting the microscopic pore-throat structure parameters of each partition; Step 300: Use the four-parameter random growth method to simulate the corrosion of the fluid in the pores and throats in each partition towards the mineral particle cores until the radii of the pores and throats in each partition reach the target growth radius value one by one, and generate the pore-throat network image of the set simulation area; Obtain the microscopic pore-throat structure parameters and phase permeability partitions through experiments on the target core sample. The microscopic pore-throat structure parameters include porosity, pore radius distribution, throat radius distribution, grain size radius distribution, and coordination number distribution; Distinguish the partitions in the set simulation area through the phase permeability partition, and correspondingly establish the microscopic pore-throat structure parameters in each partition according to the distinction result; Among them, match the coordination number for the pores and throats in each partition through the coordination number distribution; Control the target growth radius of the pores and throats in the partition through the pore radius distribution and throat radius distribution; Control the grain size radius distribution of the mineral particle cores in the partition through the grain size radius distribution.

2. The method for generating a large-scale heterogeneous pore-throat network jointly controlled by macro and micro parameters according to claim 1, characterized in that Determine the number of randomly generated mineral particle cores in each partition using the ideal model of a hexagonal pore-throat network, specifically including: Set the pores as spherical and the throats as rod-shaped, and calculate the number of pores and mineral particle cores in the simulation area under ideal conditions; Among them, optimize the number of pores and mineral particle cores calculated under ideal conditions according to the ratio of mineral particle cores to pore numbers obtained from experiments on the target core sample.

3. The method for generating a large-scale heterogeneous pore-throat network jointly controlled by macro and micro parameters according to claim 1, characterized in that Regard each randomly generated mineral particle core in the partition as a kind of mineral, and set the final porosity corresponding to each mineral particle core; And ignore the interaction between the mineral particle cores during the process of simulating the growth of the mineral particle cores using the four-parameter random growth method.

4. The method for generating a large-scale heterogeneous pore-throat network jointly controlled by macro and micro parameters according to claim 1, characterized in that According to the obtained position distribution of each pore and throat, use the connected domain algorithm to identify and extract the boundary lines and the intersection points of the boundary lines between the mineral particle cores in the position distribution result, define the boundary lines as throats, and define the intersection points of the boundary lines as potential pores.

5. The method for generating a large-scale heterogeneous pore-throat network jointly controlled by macro and micro parameters according to claim 4, characterized in that Based on the coordination number obtained from experiments on the target core sample, the fitness of the pores and throats in each partition with respect to the coordination number is calculated by a genetic algorithm, and the pores or throats in the partition are deleted according to the fitness, thereby completing the optimization of the pores and throats in each partition; wherein, the fitness calculation function of the genetic algorithm is: where n is the number of partitions in the model area, m is the total number of pores extracted in each partition, and c i,j is the coordination number of the j-th pore in the i-th partition calculated according to the individual chromosome, and c t i,j is the target coordination number distribution of all pores in each partition. sort(c i,j ) and sort(c t i,j ) are sorting functions; sum is a summation function, and the summation function is set as a penalty term to prevent the genetic algorithm from setting pores as dead pores.

6. The method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters according to claim 5, characterized in that by setting the micro pore-throat structure parameters of each partition, a target growth radius value is assigned to the pores and throats each matching the coordination number, and the assignment rules are specifically as follows: For the pores and throats in each partition, random sampling is performed according to the given pore-throat radius; The higher the extracted pore area and the more pores merged, the larger the pore radius is assigned; The radius of the throat cannot exceed the minimum pore radius among the pores at both ends of the partition.

7. The method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters according to claim 1, characterized in that Using the four-parameter random growth method to simulate the corrosion of the fluid in the pores and throats in each partition towards the mineral particle core. During the introduction of the fluid corrosion process: The total pore area is set as the primary pores between the mineral particle cores and the corrosion area after the growth of the mineral particle cores; Among them, the percentage of the corrosion area after growth between adjacent mineral particle cores in the total pore area is set as the corrosion threshold.

8. The method for generating a large-scale heterogeneous pore-throat network with joint control of macro and micro parameters according to claim 7, characterized in that After completing the radius assignment of the pores and throats in each partition, an initial pore-throat network image is formed, and an image dilation algorithm is used to thicken the throats. The specific method for thickening the throats by the image dilation algorithm includes: Obtain the size of the initial image and create a completely black image of the same size; Traverse each pixel and determine that there are non-zero common elements intersecting in the two images; After the traversal is completed, output the target image.

Images