Shale oil and gas reservoir fracture network evaluation method and device

Abstract

The present invention relates to a method and device for evaluating a shale oil and gas reservoir fracture network. The shale oil and gas reservoir fracture network evaluation method extracts rough edges through edge detection, and further refines the shape characteristics of the fracture edges and the brightness difference of pixels to obtain accurate fracture edges, thereby ensuring accurate identification of fracture pixels, making the fracture boundaries clearer and enhancing the detection accuracy of fractures; and by connecting adjacent fractures, a complete fracture structure is obtained, which can provide basic data for a comprehensive evaluation of the fracture network; in addition, by identifying and classifying fracture nodes, the number of nodes of different types is clarified, providing a data basis for fracture network connectivity analysis. At the same time, the fracture network development intensity is combined with the fracture network connectivity results to obtain a comprehensive evaluation result of the rock fracture network. The present invention realizes accurate evaluation of shale oil and gas reservoir fractures.

Description

Technical Field

[0001] The present invention relates to the field of rock fractures, and in particular to a method and device for evaluating a shale oil and gas reservoir fracture network. Background Art

[0002] Shale oil and gas reservoir fractures are not only important channels connecting pores of different scales, but also have storage and seepage capabilities. Therefore, fracture network evaluation is crucial for shale oil and gas reserve assessment and exploitation.

[0003] However, the existing shale oil and gas reservoir fracture evaluation methods often only count the length, number, area and other basic physical parameters of the fractures for shale fracture network evaluation, but ignore the connectivity of the fracture network and its impact on the performance of the fracture network system. The evaluation results cannot truly reflect the actual situation, which is not conducive to shale oil and gas exploration and development. For example, there is a fracture random simulation method based on the fracture development intensity trend in the prior art to characterize the fracture development intensity. When establishing the fracture development intensity distribution map, only the number of fractures is considered. It is believed that the greater the number of fractures, the greater the fracture development intensity, ignoring other influencing factors such as fracture width and area. Moreover, due to the large number of shale reservoir fractures and their complex distribution, the existing method of manually counting the basic physical parameters of fractures is time-consuming, labor-intensive, and inefficient, which brings great difficulties to the statistics of the basic parameters of fractures; in addition, due to the limited number of fractures counted manually and the influence of human factors, the statistical results often cannot truly reflect the real fracture physical parameters of shale oil and gas reservoirs.

[0004] In summary, the statistical results of the existing shale oil and gas reservoir fracture network development intensity and connectivity evaluation methods deviate from the actual reservoir fracture development situation, and the traditional method of statistically analyzing the basic physical parameters of fractures and fracture nodes is not only inaccurate, but also time-consuming and labor-intensive. Therefore, there is an urgent need to provide a shale oil and gas reservoir fracture network development intensity and connectivity evaluation method with high accuracy and efficiency. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide a shale oil and gas reservoir fracture network evaluation method, device and computer equipment, which is a shale oil and gas reservoir fracture network evaluation method that combines machine vision and topological structure, and is used to solve one or more of the above-mentioned problems existing in the prior art.

[0006] In order to achieve the above objectives, on the one hand, the embodiment of the present application provides a shale oil and gas reservoir fracture network evaluation method, comprising the steps of:

[0007] Acquire a crack image of a rock sample, perform edge detection processing on the crack image to obtain a crack edge, and determine a preset threshold corresponding to pixels on the crack edge based on shape characteristics of the crack edge;

[0008] Expand or shrink the crack edge according to the brightness of the pixels on the crack edge in the crack image and a preset threshold, and determine the target crack according to the crack edge;

[0009] Connect adjacent target cracks to obtain a complete crack structure, and obtain the physical parameters of each complete crack structure, and obtain the number of cracks in the crack image, the average width of the cracks, the crack surface density and the crack surface ratio according to the physical parameters;

[0010] Based on the number of cracks, average crack width, crack surface density and crack surface ratio, the crack network development intensity of the rock sample is obtained;

[0011] Acquire the crack nodes of each complete crack structure, and perform node identification on each crack node to obtain a first number of first type nodes, a second number of second type nodes, and a third number of third type nodes, and extract the node area and the area of ​​the filling in the node for each crack node to obtain the node area and the area of ​​the filling in the node for each crack node;

[0012] Obtaining fracture network connectivity results according to the first quantity, the second quantity, the third quantity, the node area, and the area of ​​the filling material in the node;

[0013] Based on the fracture network development intensity and fracture network connectivity results, the comprehensive evaluation results of the rock fracture network are determined.

[0014] In one embodiment, the step of expanding or contracting the crack edge according to the brightness of the pixels on the crack edge in the crack image and a preset threshold value includes:

[0015] Acquire the actual brightness of the pixels on the crack edge in the crack image and the corresponding background brightness, and obtain the brightness difference according to the actual brightness and the background brightness;

[0016] Confirming pixels whose brightness difference is greater than a corresponding preset threshold as first crack pixels, and confirming outwardly adjacent pixels of the first crack pixels as pixels at the edge of the crack, until the brightness difference of the outwardly adjacent pixels is less than the corresponding preset threshold;

[0017] Confirming pixels whose brightness difference is less than the corresponding preset threshold as second crack pixels, and confirming inwardly adjacent pixels of the second crack pixels as pixels at the crack edge, until the brightness difference of the inwardly adjacent pixels is less than the corresponding preset threshold;

[0018] The steps of determining the target crack according to the crack edge include:

[0019] The first crack pixel, the second crack pixel, and the pixels within the crack edge are determined as target cracks.

[0020] In one embodiment, the steps are also included:

[0021] Obtaining the brightness of adjacent pixels whose distance from any pixel on the crack edge is a preset value;

[0022] According to the brightness of adjacent pixels, the corresponding background brightness is determined.

[0023] In one embodiment, the step of obtaining the fracture network development strength of the rock sample includes:

[0024] Calculate a first weight of the number of cracks, a second weight of the average width of the cracks, a third weight of the crack surface density, and a fourth weight of the crack surface ratio;

[0025] The intensity of fracture network development is obtained according to the product of the number of fractures and the first weight, the product of the average fracture width and the second weight, the product of the fracture surface density and the third weight, and the product of the fracture surface ratio and the fourth weight.

[0026] In one embodiment, the step of calculating a first weight of the number of cracks, a second weight of the average width of the cracks, a third weight of the crack surface density, and a fourth weight of the crack surface ratio comprises:

[0027] The number of cracks, average crack width, crack surface density and crack surface ratio are standardized to obtain standardized data;

[0028] The entropy weight method is used to process the standardized data to obtain the first weight, the second weight, the third weight and the fourth weight.

[0029] In one embodiment, in the step of obtaining the fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area, and the filler area data in the node, the fracture network connectivity result is obtained based on the following formula:

[0030]

[0031] Among them, A, B, and C represent the degrees corresponding to the first type of nodes, the second type of nodes, and the third type of nodes, respectively, A=1, B=3, and C=4; f is the number of cracks and f≥2, g1 refers to the total number of the first type of nodes; g2 refers to the total number of the second type of nodes; g3 refers to the total number of the third type of nodes; S1 refers to the node area of ​​the first type of nodes; S1' refers to the node filling area of ​​each first type of node; S2 refers to the node area of ​​the second type of nodes; S2' refers to the node filling area of ​​each second type of node; S3 refers to the node area of ​​each third type of node; S3' refers to the node filling area of ​​each third type of node.

[0032] In one embodiment, the first type of nodes, the second type of nodes and the third type of nodes are obtained by dividing by introducing a topological concept;

[0033] The node area is obtained by using the Harris corner point algorithm;

[0034] The area information of the filling material in the node is obtained by using the spot detection algorithm.

[0035] In one embodiment, the step of identifying each crack node includes:

[0036] Label the rock images by node type to obtain a training set;

[0037] The training set is used to train the network model, and the trained network model is used to identify each crack node.

[0038] In one embodiment, in the step of determining the comprehensive evaluation result of the rock fracture network according to the fracture network development intensity and the fracture network connectivity results, the comprehensive evaluation result is obtained based on the following formula:

[0039] Q = z1·S′+z2·D′;

[0040] Among them, Q is the comprehensive evaluation result of the fracture network; z1 is the weight of the fracture network development intensity; z2 is the weight of the fracture network connectivity; S′ is the standardized data of the fracture network development intensity S; D′ is the standardized data of the fracture network connectivity D.

[0041] On the one hand, an embodiment of the present invention provides a shale oil and gas reservoir fracture network evaluation device, comprising:

[0042] The first edge detection module is used to obtain a crack image of a rock sample, and perform edge detection processing on the crack image to obtain a crack edge, and determine a preset threshold value corresponding to pixels on the crack edge according to shape characteristics of the crack edge;

[0043] A second edge detection module is used to expand or contract the crack edge according to the brightness of the pixels on the crack edge in the crack image and a preset threshold, and to determine the target crack according to the crack edge;

[0044] A connection module is used to connect adjacent target cracks to obtain a complete crack structure, and is also used to obtain physical parameters of each complete crack structure, and obtain the number of cracks in the crack image, the average width of the cracks, the crack surface density and the crack surface ratio according to the physical parameters;

[0045] The first evaluation module is used to obtain the fracture network development intensity of the rock sample based on the number of fractures, the average width of fractures, the fracture surface density and the fracture surface ratio;

[0046] A node identification module is used to obtain the crack nodes of each complete crack structure, and perform node identification on each crack node to obtain a first number of first type nodes, a second number of second type nodes, and a third number of third type nodes, and extract the node area and the area of ​​the filling inside the node from each crack node to obtain the node area and the area of ​​the filling inside the node of each crack node;

[0047] A second evaluation module is used to obtain the fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area and the filling area in the node;

[0048] The third evaluation module is used to determine the comprehensive evaluation results of the rock fracture network based on the fracture network development intensity and fracture network connectivity results.

[0049] On the other hand, an embodiment of the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0050] One of the above technical solutions has the following advantages and beneficial effects:

[0051] The above-mentioned shale oil and gas reservoir fracture network evaluation method extracts rough edges through edge detection, and further refines the accurate fracture edges through the shape characteristics of the fracture edges, etc., ensuring the accurate identification of fracture pixels, making the fracture boundaries clearer and enhancing the detection accuracy of fractures. And by connecting adjacent fractures, a complete fracture structure is obtained. In this way, the number, average width, surface density and porosity of fractures can be analyzed in detail, providing basic data for the comprehensive evaluation of fracture networks. In addition, by identifying and classifying fracture nodes and counting the fracture node area and the filling area in the nodes, the number and node area of ​​different types of nodes and the filling area data in the nodes are clarified, providing a data basis for the analysis of fracture network connectivity, helping to deeply understand the topological structure and connectivity characteristics of the fracture network. At the same time, the fracture network development intensity is combined with the fracture network connectivity results to obtain a comprehensive evaluation result of the shale fracture network. This comprehensive evaluation not only improves the accuracy of the evaluation, but also enhances the comprehensive understanding of the structural characteristics of shale fractures. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0054] Figure 1 is a first schematic flow chart of a method for evaluating a shale oil and gas reservoir fracture network in one embodiment;

[0055] Figure 2 is a schematic flow chart of the steps of obtaining the fracture network development strength of a rock sample in one embodiment;

[0056] Figure 3 A schematic flow chart of the steps of calculating a first weight of the number of cracks, a second weight of the average width of the cracks, a third weight of the crack surface density, and a fourth weight of the crack surface porosity in one embodiment;

[0057] Figure 4 is a schematic flow chart of the steps of performing node identification on each crack node in an embodiment;

[0058] Figure 5 A visualization diagram of a single image effect after annotation in an embodiment;

[0059] Figure 6 A visualization training result diagram in one embodiment;

[0060] Figure 7 This is a diagram showing the effect of node detection in an embodiment;

[0061] Figure 8 An example diagram of crack node types and node filling conditions in an embodiment;

[0062] Fig. 9 A crack node map located by performing Harris corner point detection algorithm processing on the cropped node image in one embodiment;

[0063] Fig.10 A result diagram of the filling material in the crack node extracted by the spot detection algorithm in one embodiment;

[0064] Fig.11 4 is a second schematic flow chart of a method for evaluating a shale oil and gas reservoir fracture network in one embodiment. DETAILED DESCRIPTION

[0065] In order to facilitate understanding of the present application, the present application will be described more fully below with reference to the relevant drawings. Embodiments of the present application are provided in the drawings. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0066] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0067] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application and have no specific meaning. Therefore, "module" and "component" can be used interchangeably.

[0068] It can be understood that the “connection” in the following embodiments should be understood as “electrical connection”, “communication connection”, etc. if the connected circuits, modules, units, etc. have electrical signals or data transmission between each other.

[0069] When used herein, the singular forms "a", "an", and "said / the" may also include plural forms, unless the context clearly indicates otherwise. It should also be understood that the terms "include / comprise" or "have" etc. specify the presence of stated features, wholes, steps, operations, components, parts or combinations thereof, but do not exclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts or combinations thereof.

[0070] In one embodiment, Figure 1 As shown, a method for evaluating a shale oil and gas reservoir fracture network is provided, comprising the steps of:

[0071] S110, obtaining a fracture image of a rock sample, and performing edge detection processing on the fracture image to obtain a fracture edge, and determining a preset threshold corresponding to pixels on the fracture edge according to shape characteristics of the fracture edge;

[0072] The rock sample may be any type of rock, such as conventional and unconventional oil and gas reservoirs such as shale oil and gas reservoirs, coalbed methane reservoirs, and tight sandstone gas reservoirs. The following takes a shale sample as an example, and an image of a shale sample can be collected by a scanning electron microscope. A fracture image is an image of a shale containing fractures. Shape features may include curvature.

[0073] Specifically, edge detection is performed on the crack image to find the edge of each crack. Common edge detection algorithms include Canny edge detection, Sobel operator, etc. The curvature of each point on the crack edge is determined, and different curvatures correspond to different preset thresholds. The corresponding curvature can be confirmed according to the position of the pixel, and the corresponding preset threshold can be further determined.

[0074] Furthermore, after obtaining the fracture image of the rock sample, the image may be preprocessed, such as graying, increasing contrast, etc. The object of edge detection processing is the preprocessed fracture image.

[0075] S120, expanding or contracting the crack edge according to the brightness of the pixels on the crack edge in the crack image and a preset threshold, and determining the target crack according to the crack edge;

[0076] Specifically, the step of expanding or shrinking the crack edge can be performed by the following means: obtaining the actual brightness of the pixels on the crack edge in the crack image and the corresponding background brightness, and obtaining the brightness difference according to the actual brightness and the background brightness;

[0077] Confirming pixels whose brightness difference is greater than a corresponding preset threshold as first crack pixels, and confirming outwardly adjacent pixels of the first crack pixels as pixels at the edge of the crack, until the brightness difference of the outwardly adjacent pixels is less than the corresponding preset threshold;

[0078] Confirming pixels whose brightness difference is less than the corresponding preset threshold as second crack pixels, and confirming inwardly adjacent pixels of the second crack pixels as pixels at the crack edge, until the brightness difference of the inwardly adjacent pixels is less than the corresponding preset threshold;

[0079] The background brightness refers to any pixel on the edge of the crack, and each pixel corresponds to a corresponding background brightness.

[0080] Specifically, when photographing rock cracks under a scanning electron microscope, the brightness of the rock crack area is usually higher than that of the non-crack area. This is because the crack area is usually filled with air or other low-density materials. When the scanning electron microscope illuminates these areas, the electrons will scatter less with the air or other low-density materials, so they appear relatively brighter. In contrast, the non-crack area of ​​the rock is usually a dense rock mineral component. When the scanning electron microscope illuminates these dense areas, there will be more electron scattering, so it appears relatively dark. The acquisition of background brightness can refer to the following steps: obtain the brightness of the adjacent pixels that are a preset value away from any pixel on the edge of the crack; determine the corresponding background brightness based on the brightness of the adjacent pixels. The preset value can be confirmed according to the actual situation. Adjacent pixels refer to pixels that are a preset value away from the target pixel.

[0081] If the pixel with brightness difference greater than the corresponding preset threshold is confirmed as the first crack pixel, the first crack pixel needs to be expanded, specifically, the outward neighboring pixels of the first crack pixel are executed once, that is, the actual brightness of the outward neighboring pixels and the corresponding background brightness are obtained, and the brightness difference is obtained according to the actual brightness and the background brightness; according to the shape characteristics of the crack edge, the preset threshold corresponding to the outward neighboring pixels is determined; the outward neighboring pixels with brightness difference greater than the corresponding preset threshold are confirmed as the first crack pixel. Repeat the execution until the brightness difference of the outward neighboring pixels is less than the corresponding preset threshold. It should be noted that the outward neighboring pixels are the neighboring pixels that are not facing the crack. For example, if the upper, lower and right pixels of the first crack pixel are all within the crack edge, the left pixel is the outward neighboring pixel.

[0082] If the pixel whose brightness difference is less than the corresponding preset threshold is confirmed as the second crack pixel, the first crack pixel needs to be shrunk, specifically, the inward neighboring pixel of the second crack pixel is executed once, that is, the actual brightness of the inward neighboring pixel and the corresponding background brightness are obtained, and the brightness difference is obtained according to the actual brightness and the background brightness; according to the shape characteristics of the crack edge, the preset threshold corresponding to the inward neighboring pixel is determined; the inward neighboring pixel whose brightness difference is less than the corresponding preset threshold is confirmed as the second crack pixel. Repeat the execution until the brightness difference of the inward neighboring pixel is greater than the corresponding preset threshold. It should be noted that the inward neighboring pixel is the neighboring pixel facing the crack. For example, if the left pixel of the first crack pixel is outside the crack edge, the pixels above, below, and right of it are inward neighboring pixels.

[0083] Furthermore, the step of determining the target crack according to the crack edge includes: determining the first crack pixel, the second crack pixel, and the pixel within the crack edge as the target crack.

[0084] S130, connecting adjacent target cracks to obtain a complete crack structure, and obtaining physical parameters of each complete crack structure, and obtaining the number of cracks in the crack image, the average width of the cracks, the crack surface density, and the crack surface ratio according to the physical parameters;

[0085] Specifically, morphological crack connection, such as closing operation, can be used to connect adjacent target cracks. Furthermore, non-crack noise areas can be deleted.

[0086] The skeleton can be extracted from each connected domain, and basic physical parameters such as width, length, width, and area can be measured to further obtain the number of cracks, average crack width, crack surface density, and crack surface ratio. The average crack width is the total crack width in the sample divided by the number, the crack surface density is the ratio of the total crack length to the actual area of ​​the crack image, and the crack surface ratio is the ratio of the total crack area to the actual area of ​​the crack image.

[0087] S140, based on the number of fractures, average fracture width, fracture surface density and fracture surface ratio, the fracture network development intensity of the rock sample is obtained;

[0088] Specifically, Figure 2 As shown, the steps of obtaining the fracture network development strength of the rock sample include:

[0089] S210, calculate the first weight of the number of cracks, the second weight of the average crack width, the third weight of the crack surface density and the fourth weight of the crack surface ratio; S220, obtain the fracture network development intensity according to the product of the number of cracks and the first weight, the product of the average crack width and the second weight, the product of the fracture surface density and the third weight, and the product of the fracture surface ratio and the fourth weight.

[0090] S150, obtaining crack nodes of each complete crack structure, and performing node identification on each crack node to obtain a first number of first type nodes, a second number of second type nodes, and a third number of third type nodes, and extracting a node area and an area of ​​a filler in each crack node to obtain a node area and an area of ​​a filler in each crack node;

[0091] Specifically, node identification can use neural networks to identify the types of crack nodes, and the node types are first-type nodes (type I nodes), second-type nodes (type II nodes), and third-type nodes (type III nodes). Topological concepts are introduced to divide and count the types of crack nodes and the number of crack nodes. The node area is obtained by using the Harris corner point algorithm; the area information of the filling material in the node is obtained by using the spot detection algorithm.

[0092] Further, the crack nodes of each complete crack structure are obtained, and each crack node is identified and a local image containing only the crack node is cut out (for subsequent crack node area statistics and node filler area statistics), and the first number of the first type of nodes, the second number of the second type of nodes, and the third number of the third type of nodes are obtained; the gradient of the pixel points in the x and y directions is calculated for the cut local image containing only the crack nodes; the components of the calculated structure tensor are obtained, that is, the square of the horizontal gradient, the square of the vertical gradient, and the product of the horizontal and vertical gradients are calculated; Gaussian smoothing is performed on each component of the obtained structure tensor to reduce the influence of noise on the corner point detection results; and the corner point response function is calculated. The corner point response value of each pixel point is calculated using the smoothed components; in order to accurately locate the crack node, non-maximum suppression is required. Each pixel point is compared with other points in its neighborhood, and only the local maximum point is retained as a corner point candidate, and other values ​​are set to zero; the corner point response value is compared with a threshold, and the points with a response value less than the threshold are removed, and only the points with a response value greater than the threshold are retained as the final corner point. In this way, some noise points with small response values ​​can be filtered out. Extract the coordinates of the corner points from the response map after non-maximum suppression and thresholding, and count the area of ​​the crack nodes. Perform Gaussian blur processing on the local image that only contains the crack nodes to reduce the impact of noise on detection; smooth the image through a series of different scales to generate a series of smoothed images. At each scale, calculate the Laplace-Gaussian (LoG) operator response to the LoG response map (a result image generated during image processing) to detect the filler inside the crack node in the image; find the local extreme point in the scale space to obtain the confirmed spot position; compare the response value in the LoG response map with the preset threshold. Only points with response values ​​greater than the threshold are retained as spots; obtain the spot coordinates and area data from the image after thresholding and extreme value detection.

[0093] S160, obtaining a fracture network connectivity result according to the first number, the second number, the third number, the node area, and the area of ​​the filler in the node;

[0094] Specifically, in one embodiment, in the step of obtaining the fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area, and the filler area data in the node, the fracture network connectivity result is obtained based on the following formula:

[0095]

[0096] Among them, A, B, C represent the degrees of the first type of nodes, the second type of nodes, and the third type of nodes, respectively, A = 1, B = 3, C = 4; f is the number of cracks and f ≥ 2, g1 refers to the total number of the first type of nodes; g2 refers to the total number of the second type of nodes; g3 refers to the total number of the third type of nodes; S1 refers to the node area of ​​the first type of nodes; S1' refers to the node filling area of ​​each first type of node; S2 refers to the node area of ​​the second type of nodes; S2' refers to the node filling area of ​​each second type of node; S3 refers to the node area of ​​each third type of node; S3' refers to the node filling area of ​​each third type of node. (f-1) means that each crack has at most (f-1) nodes (type II or type III nodes) intersecting with it, 2f is the total number of type I nodes in the crack network, and 4×f×(f-1) refers to the optimal connectivity of the crack network, that is, all nodes are type III nodes with a degree of 4.

[0097] The closer the output result D of the above function is to 1, the better the connectivity of the fracture network is. The connectivity level is determined according to the following rules:

[0098] When 0.75<D≤1, the fracture network connectivity level is excellent;

[0099] When 0.5<D≤0.75, the fracture network connectivity level is good;

[0100] When 0.25<D≤0.5, the fracture network connectivity level is moderate;

[0101] When 0<D≤0.25, the fracture network connectivity level is poor.

[0102] S170, determining a comprehensive evaluation result of the rock fracture network according to the fracture network development intensity and fracture network connectivity results.

[0103] The above-mentioned shale oil and gas reservoir fracture network evaluation method extracts rough edges through edge detection, and further refines the accurate fracture edges through the shape characteristics of the fracture edges and the brightness difference of pixels, ensuring the accurate identification of fracture pixels, making the fracture boundaries clearer and enhancing the detection accuracy of fractures. And by connecting adjacent fractures, a complete fracture structure is obtained. In this way, the number, average width, surface density and porosity of fractures can be analyzed in detail, providing basic data for the comprehensive evaluation of fracture networks. In addition, by identifying and classifying fracture nodes and counting the fracture node area and the area of ​​fillings in the nodes, the number of different types of nodes is clarified, the number and node area of ​​different types of nodes and the area of ​​fillings in the nodes are clarified, providing a data basis for fracture network connectivity analysis, and helping to deeply understand the topological structure and connectivity characteristics of the fracture network. At the same time, the fracture network development intensity is combined with the fracture network connectivity results to obtain a comprehensive evaluation result of the rock fracture network. This comprehensive evaluation not only improves the accuracy of fracture evaluation, but also enhances the comprehensive understanding of the structural characteristics of rock fractures, which is helpful for the exploration and development of shale oil and gas.

[0104] In one embodiment, if Figure 3 As shown, the steps of calculating the first weight of the number of cracks, the second weight of the average width of the cracks, the third weight of the crack surface density and the fourth weight of the crack surface ratio include:

[0105] S310, performing standardization processing on the number of cracks, the average width of cracks, the density of crack surfaces and the surface ratio of crack surfaces to obtain standardized data;

[0106] S320, using an entropy weight method to process the standardized data to obtain a first weight, a second weight, a third weight and a fourth weight.

[0107] Specifically, the weights of these four indicators are set as w1, w2, w3, and w4; N′, G′, O′, and p′ are the standardized number of cracks, average crack width, crack surface density, and crack surface ratio data, respectively; the evaluation result of the crack network development intensity is S. Then the data is standardized, and the number of cracks, average crack width, crack surface density, and crack surface ratio are used as four indicators. Each indicator corresponds to a group of data. First, the standardized data of each group is obtained. The standardization formula is as follows:

[0108]

[0109] x' is the result after data standardization; x is the original value before data standardization; x min and x max They are the minimum and maximum values ​​in the group of data to be standardized.

[0110] Then calculate the proportion of each group's standardized data in the total standardized data of the group, the formula is as follows:

[0111]

[0112] Among them, p ij is the proportion of the standardized value of the i-th sample on the j-th index; x' ij is the standardized value of the i-th sample on the j-th index; ∑ n i＝1 x' ij is the sum of the standardized values ​​of all samples of the j-th indicator; n is the total number of data samples in each group; i represents the i-th sample data; j represents the j-th indicator.

[0113] Then, the entropy value of each standardized indicator is calculated as follows:

[0114]

[0115] Among them, e j is the entropy value of the jth standardized indicator

[0116] Then, calculate the weight according to the entropy value. The formula is as follows:

[0117]

[0118] Wherein, m represents the total number of indicators, and in this application, m=4.

[0119] Finally, construct the specific function:

[0120] S=w1·N′+w2·G′+w3·O′+w4·P′

[0121] In the above formula, w1 is the first weight, w2 is the second weight, w3 is the third weight, w4 is the fourth weight, N′ is the standardized number of cracks, G′ is the standardized average width of cracks, O′ is the standardized crack surface density, P′ is the standardized crack surface rate data; S is the evaluation result of the crack network development intensity.

[0122] The function result S represents the intensity of crack network development. The larger the value, the better the intensity of crack network development in the electron microscope image.

[0123] In one embodiment, if Figure 4 As shown, the step of identifying each crack node includes:

[0124] S410, labeling the rock image with node types to obtain a training set;

[0125] Specifically, multiple electron microscope images of shale are taken, such as Figure 5 As shown, node type annotation is performed to obtain the annotated image and node type label information; mosaic data enhancement is performed on the target image to produce and obtain a node type data set, i.e., a training set. Further, the step of node type annotation can be to select and mark the outer contour of the node on the image, and annotate the node type label information: the category name of the current node.

[0126] S420, training the network model using the training set, and performing node identification on each crack node using the trained network model.

[0127] Specifically, the network model can be a YOLOV5 network; the images and node type label information in the node type data set are input into the YOLOV5 network for training, and the optimized YOLOV5 network is obtained by training the network weight parameters, and the optimized YOLOV5 network is used to select and classify unknown nodes. Figure 6 As shown in the figure, traiN / boIII_loss and val / boIII_loss represent the mean loss function of the training set and the validation set, traiN / obj_loss and val / obj_loss represent the mean target detection loss of the training set and the validation set, traiN / cls_loss and val / cls_loss represent the mean classification loss of the training set and the validation set, Netrics / precisioN and Netrics / recall represent the accuracy and recall of the training set (accuracy can judge the correctness of the machine's overall judgment; recall indicates the probability of successful node recognition in practice), Netrics / NAp_0.5 represents the average accuracy of multiple categories, where 0.5 represents the threshold for determining IOU as a positive or negative sample, and Netrics / NAp_0.5:0.95 represents the average value after the threshold is 0.5:0.05:0.95.

[0128] After 200 epochs, the model reached a convergent state. During the model training process, its accuracy reached a high level; the recall rate basically reached 100%; the harmonic mean remained at around 80%; the average precision mean also basically remained at the level of 100%; the loss function values ​​of the training set and the validation set also remained at a low state, about 0.02; the mean target detection loss and the mean classification loss of the training set and the validation set were basically close to 0, Figure 7 This is the YOLOV5 node detection effect diagram. The overall training effect is in line with the expected effect.

[0129] In one embodiment, in the step of determining the comprehensive evaluation result of the rock fracture network according to the fracture network development intensity and the fracture network connectivity results, the comprehensive evaluation result is obtained based on the following formula:

[0130] Q = z1·S′+z2·D′;

[0131] Among them, Q is the comprehensive evaluation result of the fracture network; z1 is the weight of the fracture network development intensity; z2 is the weight of the fracture network connectivity; S′ is the standardized data of the fracture network development intensity S; D′ is the standardized data of the fracture network connectivity D. The larger the value of Q, the better the comprehensive evaluation of the fracture network.

[0132] In one embodiment, Figure 8 As shown, the first type of node (10) is an isolated terminal node on the crack, connecting only one crack branch; the second type of node (20) is a node generated by the semi-intersection of two cracks, connecting three crack branches; the third type of node (30) is a node generated by the complete penetration and intersection of two line segments, connecting four crack branches. The "degree" in this example is used to indicate the number of crack branches connected to the crack node. In this embodiment, the node types in the crack network are divided into three types, specifically including the first type of node (type I node), the second type of node (type II node), and the third type of node (type III node). Each type of node is defined as follows:

[0133] Type I nodes are terminal nodes that are connected to only one fracture branch in the fracture network and have a degree of 1 ( Figure 8 a);

[0134] Type II nodes are nodes generated by the semi-intersection of two cracks, that is, one crack intersects and terminates at another crack without protruding extension. This node connects three crack branches and has a degree of 3 ( Figure 8 b);

[0135] Type III nodes are nodes where two line segments completely penetrate and intersect each other. This node connects four crack branches and has a degree of 4 ( Figure 8 c).

[0136] like Figure 8 The type I node in (a) connects a crack branch, so the degree is 1; Figure 8 The type II node in (b) connects three crack branches, so the degree is 3; Figure 8 The type III node in (c) connects four crack branches, so the degree is 4; the example of a crack node is shown in Figure 8 (d)- Figure 8 As shown in (f), the black material represents the filling inside the node, and the white represents the unfilled part inside the node. The higher the proportion of the white part, the better the connectivity of the node. Conversely, the higher the proportion of the black part, the more filling there is in the node, and the worse the connectivity of the node.

[0137] In one embodiment, the node area information is extracted based on the Harris corner detection algorithm. Fig. 9 The Harris corner point detection algorithm is used to process the cropped node image to locate the crack node, which is used to count the node area (the inside of the square box is the node located by the algorithm, and the area in this area is counted); the specific steps are:

[0138] Step 1: Calculate the gradient of each pixel in the crack node image in the x and y directions. The gradient represents the rate and direction of change of the image brightness. The Sobel operator can be used to calculate the gradient.

[0139] Use a horizontal Sobel operator to convolve the image and get the horizontal gradient of the image;

[0140] Convolve the image using a vertical Sobel operator to obtain the vertical gradient of the image;

[0141] Step 2: Calculate the components of the structure tensor, that is, calculate the square of the horizontal gradient, the square of the vertical gradient, and the product of the horizontal and vertical gradients;

[0142] Step 3: Gaussian smoothing. Perform Gaussian smoothing on each component obtained by the above calculation. Gaussian smoothing is a low-pass filter that can reduce the impact of noise on corner point detection results.

[0143] Step 4: Calculate the corner response function. Use the smoothed components to calculate the corner response value of each pixel. The calculation formula of the corner response function R is:

[0144] R = det(M) - k(trace(M)) 2

[0145] Among them, det(M) represents the determinant of the structure tensor M, trace(M) represents the trace of the structure tensor M, and k is an empirical constant, usually between 0.04 and 0.06, which will be adjusted according to the specific actual situation.

[0146] Step 5: Non-maximum suppression.

[0147] In order to accurately locate the corner points (in this invention, crack nodes), non-maximum suppression is required. Each pixel point is compared with other points in its neighborhood, and only the local maximum points are retained as corner point candidates, and other values ​​are set to zero.

[0148] Step 6: Thresholding;

[0149] Compare the corner point response value R with a threshold. Points with response values ​​less than the threshold are removed, and only points greater than the threshold are retained as the final corner points. This can filter out some noise points with small response values.

[0150] Step 7: Get the corner point coordinates;

[0151] The corner coordinates are extracted from the response image after non-maximum suppression and thresholding. These coordinates represent the positions of the corners detected in the image. Finally, the areas of the corners (crack nodes) are counted.

[0152] In one embodiment, a spot detection algorithm is used to extract the filler area information in the node. Fig.10 It is the filling result in the crack node extracted by the spot detection algorithm (the irregular curve is the filling, and its area is the filling area); the specific steps are:

[0153] Step 1', Gaussian blur. Apply a Gaussian blur filter to smooth the image and reduce the impact of noise on detection.

[0154] Step 2', select the scale space. In order to detect spots of different sizes (referring to the fillings in the nodes in this invention), the scale space theory is used. The scale space is to smooth the image through a series of different scales to form a set of images. The specific operation is as follows:

[0155] Multi-scale Gaussian smoothing: Perform multiple Gaussian smoothing on the image, using a different Gaussian kernel scale (σ value) each time to generate a series of smoothed images.

[0156] Step 3', calculate Laplace-Gaussian (LoG).

[0157] At each scale, the Laplacian of Gaussian (LoG) operator response is calculated. The LoG operator is used to detect spots in the image, and its response value reaches an extreme value at the center of the spot. The specific convolution operation is as follows: For each smoothed image, convolution is performed using the LoG operator to obtain a LoG response map (a result image generated during the image processing process).

[0158] Step 4', extreme value detection. Find local extreme value points in the scale space. Local extreme value points represent the center of spots at a certain scale. The specific operations are as follows:

[0159] Non-maximum suppression: In the LoG response map of each scale, find the local maximum points and record these points as candidate spots.

[0160] Multi-scale extreme value detection: Compare the LoG response maps at different scales to find the points with extreme values ​​at multiple scales and further confirm the spot location.

[0161] Step 5', thresholding. In order to remove noise and weak responses, the response value in the LoG response map is compared with the preset threshold. Only points with response values ​​greater than the threshold are retained as spots. The specific operation is as follows:

[0162] Set threshold: Choose an appropriate threshold and adjust it according to the characteristics of the image and the size of the spots.

[0163] Apply threshold: The values ​​in the LoG response map are compared with the threshold, and points smaller than the threshold are removed.

[0164] Step 6': extract the spot coordinates and area from the image after thresholding and extrema detection.

[0165] The above two algorithms can not only extract the area at the node and the area of ​​the filling inside the node, but are also applicable to the extraction of the entire crack and the filling area inside the crack. However, the connectivity at the crack node is the most critical in the connectivity evaluation of the entire crack network. Therefore, only the connectivity at the node is considered in this application.

[0166] In one embodiment, Fig.11 As shown, a fracture network evaluation method is provided, comprising the following steps: collecting shale sample specimens through a scanning electron microscope, graying the collected electron microscope specimens; increasing the contrast of the grayed specimens; performing Canny algorithm edge detection on the specimens after contrast increase; connecting adjacent fractures using morphology; finding all connected domains and deleting non-fracture noise areas; extracting a skeleton for each connected domain, and measuring basic physical parameters such as length and area; substituting data information such as fracture length and area into relevant formulas to calculate surface density and surface face rate for evaluating the development intensity of the fracture network; introducing the concept of topological structure to divide fracture nodes into three types: I, II, and III, as basic parameters of topological structure, combined with fracture node area and node filling area data information to characterize fracture network connectivity; annotating multiple images with node types as a data set for model training, and then inputting the images and node type data information in the data set into a YOLOV5 network for training, training network weight parameters, and obtaining an optimized YOLOV5 network;

[0167] The images processed by grayscale and Canny edge detection algorithm are input into the optimized YOLOV5 network for node recognition, and the number of three node types is counted and the node images are cropped; the Harris corner algorithm is used to extract the node area information from the docking point image, and the spot algorithm is used to extract the node filling area information from the docking point image. The data of the three types of nodes in the statistical shale electron microscope image samples are substituted into the constructed fracture network connectivity function model to complete the fracture network connectivity evaluation; the fracture network is comprehensively evaluated by combining the fracture network development intensity information and the fracture network connectivity results.

[0168] This patent has the following advantages: (1) The Canny edge detection algorithm based on machine vision is used to extract basic physical parameters such as crack length and width, which is not only efficient but also improves the accuracy of crack parameter results; (2) The entropy weight method is used to combine multiple basic crack parameter information to establish a mathematical model to ensure the objectivity and accuracy of the crack development intensity results; (3) The YOLOV5 network model is used to identify the type of crack nodes and count the corresponding number, which saves time and improves the accuracy compared with the traditional statistical module; (4) While introducing the topological structure, the influence of the area at the crack node and the area of ​​the filler in the node on the connectivity of the crack network is considered. The combination of the crack node topology data and the node area data makes the crack network connectivity results more accurate; (5) The Harris corner detection algorithm and the spot detection algorithm ensure the efficiency and accuracy of extracting the crack node area and the filler area data in the crack node.

[0169] In one embodiment, a rock fracture network evaluation device is provided, comprising:

[0170] The first edge detection module is used to obtain a crack image of a rock sample, and perform edge detection processing on the crack image to obtain a crack edge, and determine a preset threshold value corresponding to pixels on the crack edge according to shape characteristics of the crack edge;

[0171] A second edge detection module is used to expand or contract the crack edge according to the brightness of the pixels on the crack edge in the crack image and a preset threshold, and to determine the target crack according to the crack edge;

[0172] A connection module is used to connect adjacent target cracks to obtain a complete crack structure, and is also used to obtain physical parameters of each complete crack structure, and obtain the number of cracks in the crack image, the average width of the cracks, the crack surface density and the crack surface ratio according to the physical parameters;

[0173] The first evaluation module is used to obtain the fracture network development intensity of the rock sample based on the number of fractures, the average width of fractures, the fracture surface density and the fracture surface ratio;

[0174] A node identification module is used to obtain the crack nodes of each complete crack structure, and perform node identification on each crack node to obtain a first number of first type nodes, a second number of second type nodes, and a third number of third type nodes, and extract the node area and the area of ​​the filling inside the node from each crack node to obtain the node area and the area of ​​the filling inside the node of each crack node;

[0175] A second evaluation module is used to obtain the fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area and the filling area in the node;

[0176] The third evaluation module is used to determine the comprehensive evaluation results of the rock fracture network based on the fracture network development intensity and fracture network connectivity results.

[0177] For the specific definition of the shale oil and gas reservoir fracture network evaluation device, please refer to the definition of the shale oil and gas reservoir fracture network evaluation method mentioned above, which will not be repeated here. Each module in the above-mentioned shale oil and gas reservoir fracture network evaluation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0178] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps of the above method when executing the computer program.

[0179] In one embodiment, a computer-readable storage medium is further provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0180] When the embodiments of the present application are specifically implemented, reference may be made to the above-mentioned embodiments, which have corresponding technical effects.

[0181] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0182] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. The above is only a specific implementation method of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will comply with the widest range consistent with the principles and novel features applied for herein.

Claims

1. A method for evaluating a shale oil and gas reservoir fracture network, characterized in that: include: Acquire a crack image of a rock sample, perform edge detection processing on the crack image to obtain a crack edge, and determine a preset threshold corresponding to pixels on the crack edge according to shape characteristics of the crack edge; According to the brightness of the pixels on the crack edge in the crack image and the preset threshold, the crack edge is expanded or contracted, and the target crack is determined according to the crack edge; adjacent target cracks are connected to obtain a complete crack structure, and physical parameters of each complete crack structure are obtained, and the number of cracks, average crack width, crack surface density and crack surface ratio in the crack image are obtained according to the physical parameters; Based on the number of cracks, the average width of cracks, the crack surface density and the crack surface ratio, the crack network development strength of the rock sample is obtained; Acquire the crack nodes of each complete crack structure, and perform node identification on each crack node to obtain a first number of first type nodes, a second number of second type nodes, and a third number of third type nodes, and extract the node area and the area of ​​the filling in the node from each crack node to obtain the node area and the area of ​​the filling in the node of each crack node; Obtaining a fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area, and the area of ​​the filler in the node; A comprehensive evaluation result of the rock fracture network is determined based on the fracture network development intensity and the fracture network connectivity results.

2. The shale oil and gas reservoir fracture network evaluation method according to claim 1, characterized in that: The step of expanding or contracting the crack edge according to the brightness of the pixels on the crack edge in the crack image and the preset threshold value comprises: Acquire the actual brightness of the pixels on the crack edge in the crack image and the corresponding background brightness, and obtain the brightness difference according to the actual brightness and the background brightness; Confirming the pixel whose brightness difference is greater than the corresponding preset threshold as a first crack pixel, and confirming the outward neighboring pixels of the first crack pixel as pixels at the edge of the crack, until the brightness difference of the outward neighboring pixels is less than the corresponding preset threshold; Confirming the pixel whose brightness difference is less than the corresponding preset threshold as a second crack pixel, and confirming the inner neighboring pixel of the second crack pixel as a pixel at the edge of the crack, until the brightness difference of the inner neighboring pixel is less than the corresponding preset threshold; The step of determining a target crack according to the crack edge comprises: The first crack pixel, the second crack pixel, and pixels within the crack edge are determined as target cracks.

3. The shale oil and gas reservoir fracture network evaluation method according to claim 2, characterized in that: Also includes the steps: Acquire the brightness of adjacent pixels whose distance from any pixel on the crack edge is a preset value; The corresponding background brightness is determined according to the brightness of the adjacent pixels.

4. The shale oil and gas reservoir fracture network evaluation method according to claim 1, characterized in that: The step of obtaining the fracture network development strength of the rock sample comprises: Calculate a first weight of the number of cracks, a second weight of the average width of the cracks, a third weight of the crack surface density, and a fourth weight of the crack surface ratio; The fracture network development intensity is obtained according to the product of the number of fractures and the first weight, the product of the average width of the fractures and the second weight, the product of the fracture surface density and the third weight, and the product of the fracture surface ratio and the fourth weight.

5. The shale oil and gas reservoir fracture network evaluation method according to claim 4, characterized in that: The steps of calculating the first weight of the number of cracks, the second weight of the average width of the cracks, the third weight of the crack surface density and the fourth weight of the crack surface ratio include: The number of cracks, the average width of the cracks, the surface density of the cracks and the surface ratio of the cracks are standardized to obtain standardized data; The standardized data is processed using an entropy weight method to obtain the first weight, the second weight, the third weight, and the fourth weight.

6. The shale oil and gas reservoir fracture network evaluation method according to claim 1, characterized in that: In the step of obtaining the fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area and the filler area data in the node, the fracture network connectivity result is obtained based on the following formula: Among them, A, B, and C represent the degrees of the first type of nodes, the second type of nodes, and the third type of nodes, respectively, A=1, B=3, and C=4; f is the number of cracks and f≥2, g1 refers to the total number of the first type of nodes; g2 refers to the total number of the second type of nodes; g3 refers to the total number of the third type of nodes; S1 refers to the node area of ​​the first type of nodes; S1 ’ S2 refers to the node area of ​​each first-type node; S2 refers to the node area of ​​the second-type node; ’ S3 refers to the node area of ​​each second type node; S3 refers to the node area of ​​each third type node; ’ Refers to the in-node filling area of ​​each third-type node.

7. The shale oil and gas reservoir fracture network evaluation method according to claim 6, characterized in that: The first type of nodes, the second type of nodes and the third type of nodes are obtained by dividing by introducing a topological concept; The node area is obtained by using the Harris corner point algorithm; The filling area information in the node is obtained by processing with a spot detection algorithm.

8. The shale oil and gas reservoir fracture network evaluation method according to claim 1, characterized in that: The step of performing node identification on each of the crack nodes comprises: Label the rock images by node type to obtain a training set; The training set is used to train the network model, and the trained network model is used to perform node identification on each of the crack nodes.

9. The shale oil and gas reservoir fracture network evaluation method according to claim 1, characterized in that: In the step of determining the comprehensive evaluation result of the rock fracture network according to the fracture network development intensity and the fracture network connectivity result, the comprehensive evaluation result is obtained based on the following formula: ； in, is the comprehensive evaluation result of the fracture network; z1 is the weight of the fracture network development intensity; z2 is the weight of the fracture network connectivity; is the standardized data of fracture network development intensity S; is the normalized data of fracture network connectivity D.

10. A shale oil and gas reservoir fracture network evaluation device, characterized in that: include: A first edge detection module is used to obtain a crack image of a rock sample, and perform edge detection processing on the crack image to obtain a crack edge, and determine a preset threshold corresponding to pixels on the crack edge according to shape characteristics of the crack edge; A second edge detection module, configured to expand or contract the crack edge according to the brightness of pixels on the crack edge in the crack image and the preset threshold, and determine a target crack according to the crack edge; A connection module, used to connect adjacent target cracks to obtain a complete crack structure, and also used to obtain physical parameters of each complete crack structure, and obtain the number of cracks in the crack image, the average width of the cracks, the crack surface density and the crack surface ratio according to the physical parameters; A first evaluation module is used to obtain the fracture network development strength of the rock sample based on the number of fractures, the average width of the fractures, the fracture surface density and the fracture surface ratio; a node identification module, for obtaining the crack nodes of each of the complete crack structures, and performing node identification on each of the crack nodes to obtain a first number of first-type nodes, a second number of second-type nodes, and a third number of third-type nodes, and extracting the node area and the area of ​​the filling material in the node from each of the crack nodes to obtain the node area and the area of ​​the filling material in the node of each of the crack nodes; A second evaluation module is used to obtain a fracture network connectivity result according to the first quantity, the second quantity, the third quantity, the node area, and the filling area in the node; The third evaluation module is used to determine the comprehensive evaluation result of the rock fracture network according to the fracture network development intensity and the fracture network connectivity result.

Images