Tracing monitoring shaft-matrix-crack intelligent modeling method and system

Abstract

The invention discloses a tracer monitoring shaft-matrix-crack intelligent modeling method and a tracer monitoring shaft-matrix-crack intelligent modeling system, relates to the technical field of oil and gas reservoir tracer monitoring numerical simulation, and solves the problems that the total grid number of an actual model cannot be reduced and the operation speed cannot be increased while the oil and gas reservoir three-dimensional modeling precision cannot be ensured or even improved in the prior art. The method comprises the steps of obtaining a reservoir grid coordinate, performing encryption operation on the vicinity of a shaft to obtain a shaft grid coordinate point data body, and performing encryption operation on the vicinity of a fracturing crack to obtain a modeling matrix-fracturing crack grid coordinate point data body; the shaft grid coordinate point data body and the matrix-fracturing crack grid coordinate point data body are combined to serve as an overall grid coordinate point data body, and then clustering processing is carried out to obtain the coupling connection relation between unencrypted grids and encrypted grids in the overall grid coordinate point data body; according to the method, hierarchical multi-scale encrypted grid modeling is adopted, the model operation speed is increased, and therefore the accuracy of oil and gas well productivity analysis based on tracing monitoring is further improved.

Description

Technical Field

[0001] The invention relates to the technical field of numerical simulation for oil and gas reservoir tracer monitoring, and in particular to a method and system for intelligent modeling of a wellbore-matrix-fracture for tracer monitoring. Background Art

[0002] Oil and gas reservoir geological modeling is a key step in characterizing the spatial distribution and heterogeneity of underground reservoirs. The interval of field logging data points is generally 0.125m, while the grid size used in traditional oil and gas reservoir simulation is more than 10m. The logging data with an interval of 0.125m needs to be coarsened to more than 10m for production capacity analysis, which greatly reduces the accuracy and utilization of real logging data, and reduces the accuracy of subsequent production capacity analysis.

[0003] There are many related studies on modeling of oil and gas reservoirs in the existing technology, including CN105095986B, which published a method for predicting the overall production of multi-layer oil reservoirs; CN111222271B, which disclosed a numerical simulation method and system for fractures in oil reservoirs based on matrix-fracture non-steady-state crossflow; CN105528522A, which disclosed a method and device for calculating the continuous oil and gas reservoir resource based on spatial grids; In Volume 40 of 2021, Bulletin of Geological Science and Technology, Liu Yanfeng, Zhang Wenbiao, Duan Taizhong, etc. conducted a study on the progress of deep learning geological modeling of oil and gas reservoirs ; In Volume 3, 2018, Bulletin of Petroleum Science, Zhu Zhouyuan, Li Minghui, et al. conducted a study on high-order single chemical flooding simulation based on high-order format, and used high-order format method to improve the accuracy of chemical flooding simulation; In Volume 31, 2010, Acta Petrolei Sinica, Su Yunhe, Du Zhimin, et al. conducted a study on dynamic local grid encryption technology for high water-content reservoirs; However, the above studies were unable to reduce the total number of grids in the actual model while ensuring or even improving the accuracy of three-dimensional modeling of oil and gas reservoirs, and further improve the accuracy of oil and gas well productivity analysis based on tracer monitoring. Summary of the invention

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention provides a tracer monitoring wellbore-matrix-fracture intelligent modeling method and system, which solves the technical problem that the prior art is unable to reduce the total number of grids in the actual model while ensuring or even improving the accuracy of three-dimensional modeling of oil and gas reservoirs, thereby further improving the accuracy of oil and gas well productivity analysis.

[0005] A tracer monitoring wellbore-matrix-fracture intelligent modeling method, comprising:

[0006] Step 1: Collect basic parameters, well trajectory data, logging interpretation data and fracturing data;

[0007] Step 2: Process the data collected in step 1 to determine the maximum and minimum values ​​of the three-dimensional coordinates of the reservoir and establish a reservoir grid model to obtain the reservoir grid coordinates;

[0008] Specifically, the coordinate value of the center point of each grid in the reservoir grid model is used as the grid coordinate:

[0009] x 4i =x0+0.5l+il

[0010] y 4i =y0+0.5l+jl

[0011] z 4i =z0+0.5l+kl

[0012] Among them, i, j, k are the numbers of each grid in the three-dimensional block reservoir grid model in the XYZ directions, and (x0, y0, z0) are the coordinates of the starting point of the grid.

[0013] Step 3: using an unsupervised machine learning algorithm to perform an encryption operation on n reservoir grid points near the wellbore to obtain a wellbore grid coordinate point data volume, and using an unsupervised machine learning algorithm to perform an encryption operation on m reservoir grid points near the hydraulic fracture to obtain a modeling matrix-hydraulic fracture grid coordinate point data volume, wherein the encryption operation is implemented by a clustering processing algorithm;

[0014] Step 4: The wellbore grid coordinate point data volume and the matrix-fracturing fracture grid coordinate point data volume are combined as the overall grid coordinate point data volume, and the overall grid coordinate point data volume is clustered using an unsupervised machine learning algorithm to obtain the coupling connection relationship between the unencrypted grid, i.e., the coarse grid, and the encrypted grid, i.e., the fine grid. The overall grid coordinate point data volume and the coupling connection relationship between the coarse-fine grid are used as the reservoir model corresponding to the wellbore-fracture-matrix.

[0015] The unsupervised machine learning algorithm can be implemented using a clustering model, and the clustering models include: KNN, Kmeans, density clustering, hierarchical clustering, etc.

[0016] More specifically:

[0017] 1. Basic parameters: coarse grid size l 粗 , fine mesh size l 细 , formation temperature, original formation pressure, viscosity and density of formation oil and water, apparent molecular weight of gas, injection and production time, etc.;

[0018] 2. Well trajectory data: drilling depth d 1i , actual vertical depth z i , east-west bias x i , north-south direction y i ;

[0019] 3. Well logging interpretation data: Well logging depth d 2i , porosity Φ i , permeability ki , water saturation S wi ;

[0020] 4. Fracturing data: Fracturing section depth d 3i , the crack is half the length of the xi , crack height f yi , main fracture permeability k fi , permeability k of the fracture zone SRVi .

[0021] Furthermore, step 3 includes: searching for n grid points closest to all wellbore points within the reservoir grid coordinate range using the wellbore coordinate data as the search point and performing encryption operations to obtain a wellbore grid coordinate point data body; searching for m grid points closest to all fracturing fractures within the reservoir grid coordinate range using the fracturing fracture data body as the search point and performing encryption operations to obtain a matrix-fracturing fracture grid coordinate point data body.

[0022] Furthermore, the encryption operation includes: deleting the searched reservoir grid points, calculating the starting point of the local grid and the coordinates of all encrypted grids according to the size of the searched grid points and the encrypted grid parameters, and merging the retained reservoir grid point data volume with the encrypted grid coordinate point data volume to obtain a new grid coordinate point data volume. Specifically:

[0023] 1. Dig a hole. Dig out the selected reservoir grid point and delete its corresponding grid number and coordinate point value (x 4i ,y 4i , z 4i );

[0024] 2. Fill. According to the side length of the selected grid and the grid size parameters to be encrypted (such as the encryption multiple, the encrypted grid side length l 细 ), calculate the coordinates of the starting point of the local grid and the center point of all the encrypted grids;

[0025] 3. Merge. Merge the undeleted reservoir grid point data volume with the fine grid coordinate point data volume (i.e., the center point coordinates of all encrypted grids) to obtain a new grid coordinate point data volume, where the wellbore grid coordinate point data volume is represented by (x 4i ',y 4i ', z 4i '), the overall grid coordinate point data volume is represented by (x 4i ”,y 4i ”, z 4i ”).

[0026] Furthermore, in step 3, the encrypted grid parameters when the encryption operation is performed on the n reservoir grid points near the wellbore are different from the encrypted grid parameters when the encryption operation is performed on the m reservoir grid points near the fracturing cracks.

[0027] Furthermore, the calculation process of the wellbore coordinate data includes: establishing interpolation functions between the drilling depth and the actual vertical depth, the east-west deviation, and the north-south deviation as interpolation functions of the three well trajectories, specifically establishing the drilling depth d 1i and the actual vertical depth z i The interpolation function F between z (d 1i ) as an example:

[0028] A. Determine the complexity of the wellbore;

[0029] B1. For a simple wellbore trajectory, the wellbore trajectory is divided into multiple sections (such as vertical well section, inclined well section, horizontal section), and z is fitted in each section. i and d 1i The relationship between the two is fitted using a linear or nonlinear function:

[0030] z i =F z (d 1i )=fk(d 1i ),d 1i ∈[d 1k ,d 1k+1 ]

[0031] Where, k is the number of well sections;

[0032] It can also be used for complex wellbore trajectories, but manual segmentation is required.

[0033] B2. For complex wellbore trajectories, exponential or logarithmic functions are used for fitting:

[0034] z i =F z (d 1i )=a·ln(d 1i )+b+c

[0035] or

[0036] z i =F z (d 1i )=a·exp(b·d 1i )+c

[0037] Where a and b are fitting coefficients, and c is the spatial offset. c can be set manually (to adjust the offset of the well trajectory in space);

[0038] B21. On the basis of exponential or logarithmic function, the well inclination angle can be further added for correction to improve the fitting accuracy:

[0039] zi =F z (d 1i )=a·ln(d 1i )+b+c·cosθ i

[0040] or

[0041] z i =F z (d 1i )=a·exp(b·d 1i )+c·cosθ i

[0042] Similarly, establish the drilling depth d 1i and east-west bias x i The interpolation function x between i =F x (d 1i ), establish drilling depth d 1i and north-south deflection y i The interpolation function y between i =F y (d 1i );

[0043] The complete well trajectory coordinate values ​​in the reservoir, namely the wellbore coordinate data, are calculated by interpolation based on the three well trajectory interpolation functions.

[0044] Furthermore, obtaining the reservoir grid coordinates in step 2 includes:

[0045] Step 2.1: Substitute the logging depth value in the logging interpretation data into the three well trajectory interpolation functions to calculate the three-dimensional coordinates of the logging data points in the reservoir, and combine the calculated value with the logging interpretation data as a complete logging data body;

[0046] Step 2.2: Determine the range of the three-dimensional coordinates based on the coordinate values ​​of the logging points in the complete logging data volume and then calculate the three-dimensional maximum length of the reservoir;

[0047] Step 2.3: Set the side length of the reservoir grid and calculate the number of reservoir grids;

[0048] Step 2.4: Establish a reservoir grid model and calculate the coordinates of the center point of each grid in the reservoir grid model as the reservoir grid coordinates.

[0049] Furthermore, the fracturing section depth in the fracturing data is substituted into the three well trajectory interpolation functions to calculate the three-dimensional coordinates of the starting and ending points of the fracturing fracture in the reservoir, and the normal vector of the main fracture surface of the fracturing is calculated according to the three-dimensional coordinate values ​​of the starting and ending points of the well trajectory, and the above-mentioned calculated values ​​are combined with the fracturing data to obtain a complete fracturing fracture data volume. Specifically, the well trajectory starting point P1 = (x1, y1, 1) and the ending point P2 = (x2, y2, 1) are set, and the direction vector can be expressed as:

[0050]

[0051] Its length L can be calculated by the following formula:

[0052]

[0053] Final Normal Vector It can be expressed as a direction vector divided by its length:

[0054]

[0055] The above calculation results are combined with the fracturing crack data to obtain a complete fracturing crack data volume, including: the coordinate value of the fracturing crack initiation point (x 3i ,y 3i , z 3i ), the normal vector centered at the crack initiation point Cracked things half length f xi , crack height f yi , main fracture permeability k fi , permeability k of the fracture zone SRVi .

[0056] The subscript i in the data indicates that there are multiple data points and they are stored in a list in the computer.

[0057] Furthermore, the calculation formula of the extension range of the crack is as follows:

[0058]

[0059] Among them, (x 3i ,y 3i , z 3i ) is the coordinate data of the fracturing starting point, (x 4i ,y 4i , z 4i ) is the reservoir grid coordinate, l 粗 is the reservoir grid size, f xi is the half length of the crack, f yi is the crack height.

[0060] A tracer monitoring wellbore-matrix-crack intelligent modeling system is used to implement a tracer monitoring wellbore-matrix-crack intelligent modeling method, comprising a data body generation module, a data body processing module, a reservoir grid model construction module, and an encryption modeling module; the data body generation module is used to generate an original data body according to the collected data, the data body processing module is used to process the original data body to respectively obtain the well trajectory coordinates, the logging data body and the fracturing crack data body, the reservoir grid model construction module is used to establish a reservoir grid model according to the logging data body, and the encryption modeling module is used to cluster and encrypt the reservoir grid model based on the logging data body and the fracturing crack data body to obtain a grid system corresponding to the wellbore-crack-matrix.

[0061] The beneficial effects of the present invention include:

[0062] 1. The algorithm is fast and efficient based on block grids and has strong adaptability to complex wellbore trajectories and fracture extensions;

[0063] 2. The hierarchical grid is used to encrypt the wellbore-fracture-matrix grid structure, which can make fuller use of the logging data points to conduct hierarchical and detailed modeling of the reservoir around the wellbore, providing basic data for the accuracy of the production capacity simulation analysis;

[0064] 3. For the first time, it is proposed to use unsupervised machine learning algorithms to perform local grid encryption processing and establish coarse-fine grid coupling relationship after spatial coordinate point relocation of field data such as well trajectory data, logging interpretation data, and fracturing data;

[0065] 4. This method can be used not only for multi-scale modeling at the coarse-fine grid level, but also for multi-scale modeling at the coarse-medium-fine grid level or above. Among them, fine grids can be used for encrypted modeling near the wellbore, medium-scale grids can be used for encrypted modeling of the matrix of the fracturing cracks and the surrounding areas, and the rest are coarse-scale grids. To achieve this hierarchical encrypted modeling, you only need to adjust the size of the encrypted grid in the "filling" step according to the selected grid;

[0066] 5. The use of hierarchical multi-scale encrypted grid modeling can greatly reduce the total number of grids in the actual model while ensuring or even improving the modeling accuracy, further improving the accuracy of oil and gas well productivity analysis based on tracer monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 Schematic diagram of the coarse grid and its coupled connection relationship.

[0068] Figure 2 Schematic diagram of the coarse grid and fine grid after the encryption operation.

[0069] Figure 3Schematic diagram of the coarse-fine grid coupling connection relationship.

[0070] Figure 4 3D rendering of coarse-fine grid multi-scale modeling of the actual reservoir model.

[0071] Figure 5 Planar rendering of coarse-fine grid multi-scale modeling for an actual reservoir model.

[0072] Figure 6 It is a multi-scale schematic diagram of the coarse-fine grid wellbore-matrix-fracture fracture of the actual reservoir model. DETAILED DESCRIPTION

[0073] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0074] Wellbore-matrix-fracture adaptive modeling method for fracturing directional wells, such as Figure 1-3 As shown, including:

[0075] Step 1: Collect basic parameters, well trajectory data, logging interpretation data and fracturing data; specifically:

[0076] 1. Basic parameters: coarse grid size l 粗 , fine mesh size l 细 , formation temperature, original formation pressure, viscosity and density of formation oil and water, apparent molecular weight of gas, injection and production time, etc.;

[0077] 2. Well trajectory data: drilling depth d 1i , actual vertical depth z i , east-west bias x i , north-south direction y i ;

[0078] 3. Well logging interpretation data: Well logging depth d 2i , porosity Φ i , permeability k i , water saturation S wi ;

[0079] 4. Fracturing data: Fracturing section depth d 3i , the crack is half the length of the xi, crack height f yi , main fracture permeability k fi , permeability k of the fracture zone SRVi .

[0080] The subscript i in the data indicates that there are multiple data points and they are stored in a list in the computer.

[0081] Step 2: Process the data collected in step 1 to determine the maximum and minimum values ​​of the three-dimensional coordinates of the reservoir and establish a reservoir grid model to obtain the reservoir grid coordinates;

[0082] Step 3: using an unsupervised machine learning algorithm to perform an encryption operation on n reservoir grid points near the wellbore to obtain a wellbore grid coordinate point data volume, and using an unsupervised machine learning algorithm to perform an encryption operation on m reservoir grid points near the hydraulic fracture to obtain a modeling matrix-hydraulic fracture grid coordinate point data volume, wherein the encryption operation is implemented by a clustering processing algorithm;

[0083] Step 4: The wellbore grid coordinate point data volume and the matrix-fracturing fracture grid coordinate point data volume are combined as the overall grid coordinate point data volume, and the overall grid coordinate point data volume is clustered using an unsupervised machine learning algorithm to obtain the coupling connection relationship between the unencrypted grid, i.e., the coarse grid, and the encrypted grid, i.e., the fine grid. The overall grid coordinate point data volume and the coupling connection relationship between the coarse-fine grid are used as the reservoir model corresponding to the wellbore-fracture-matrix.

[0084] The unsupervised machine learning algorithm can be implemented using a clustering model, and the clustering models include: KNN, Kmeans, density clustering, hierarchical clustering, etc.

[0085] In another embodiment, the calculation process of the wellbore coordinate data includes: establishing interpolation functions between the drilling depth and the actual vertical depth, the east-west deviation, and the north-south deviation as interpolation functions of the three well trajectories, specifically establishing the drilling depth d 1i and the actual vertical depth z i The interpolation function F between z (d 1i ) as an example:

[0086] A. Determine the complexity of the wellbore trajectory;

[0087] B1. For a simple wellbore trajectory, the wellbore trajectory is divided into multiple sections (such as vertical well section, inclined well section, horizontal section), and z is fitted in each section. i and d 1i The relationship between the two is fitted using a linear or nonlinear function:

[0088] z i =F z (d 1i)=fk(d 1i ),d 1i ∈[d 1k ,d 1k+1 ]

[0089] Where, k is the number of well sections;

[0090] It can also be used for complex wellbore trajectories, but manual segmentation is required.

[0091] B2. For complex wellbore trajectories, exponential or logarithmic functions are used for fitting:

[0092] z i =F z (d 1i )=a·ln(d 1i )+b+c

[0093] or

[0094] z i =F z (d 1i )=a·exp(b·d 1i )+c

[0095] Where a and b are fitting coefficients, and c is the spatial offset. c can be set manually (to adjust the offset of the well trajectory in space);

[0096] B21. On the basis of exponential or logarithmic function, the well inclination angle can be further added for correction to improve the fitting accuracy:

[0097] z i =F z (d 1i )=a·ln(d 1i )+b+c·cosθ i

[0098] or

[0099] z i =F z (d 1i )=a·exp(b·d 1i )+c·cosθ i

[0100] Similarly, establish the drilling depth d 1i and east-west bias x i The interpolation function x between i =F x (d 1i ), establish drilling depth d 1i and north-south deflection y i The interpolation function y between i =Fy (d 1i );

[0101] The complete well trajectory coordinate values ​​in the reservoir, namely the wellbore coordinate data, are calculated by interpolation based on the three well trajectory interpolation functions.

[0102] In another embodiment, the acquisition of reservoir grid coordinates includes: collecting logging interpretation data to calculate a complete logging data body, determining the range of three-dimensional coordinates based on the logging point coordinate values ​​in the complete logging data body and then calculating the three-dimensional maximum length of the reservoir, and then setting the side length of the reservoir grid to calculate the number of reservoir grids, and finally establishing a reservoir grid model and calculating the center point coordinates of each grid in the reservoir grid model as the reservoir grid coordinates. The details are as follows:

[0103] The logging depth value d in the logging interpretation data 2i , this value is brought into the interpolation function of the three well trajectories to calculate the three-dimensional (x 2i ,y 2i , z 2i ) coordinates. At this time, the logging depth value d 2i ,Right now

[0104] x 2i =F x (d 2i )

[0105] y 2i =F y (d 2i )

[0106] z 2i =F z (d 2i )

[0107] The above calculation results are combined with the logging interpretation data to obtain a complete logging data volume for subsequent modeling: the coordinate value of the logging point (x 2i ,y 2i , z 2i ), porosity Φ i , permeability k i , water saturation S wi .

[0108] According to the coordinate value of the logging point in the logging data volume (x 2i ,y 2i , z 2i ), respectively determine the maximum and minimum values ​​of the XYZ three-dimensional coordinates, and calculate the three-dimensional XYZ maximum length of the reservoir (L x , L y , L z ):

[0109] L x =max(x i )-min(x i )

[0110] L y =max(y i )-min(y i )

[0111] L z =max(z i )-min(z i )

[0112] Set the side length of the coarse grid to l 粗 , then we get the XYZ grid number (n x , n y , n z )for:

[0113] n x =L x / l 粗

[0114] n y =L y / l 粗

[0115] n z =L z / l 粗

[0116] Based on the above data, a reservoir grid model can be established, and the coordinate values ​​of the center point of each grid in the reservoir grid model can be calculated as grid coordinates:

[0117] x 4i =x0+0.5l+il

[0118] y 4i =y0+0.5l+jl

[0119] z 4i =z0+0.5l+kl

[0120] Among them, i, j, k are the numbers of each grid in the three-dimensional block reservoir grid model in the XYZ directions, and (x0, y0, z0) are the coordinates of the starting point of the grid.

[0121] In another embodiment, the fracturing crack data is collected and the fracturing section depth is substituted into three well trajectory interpolation functions to calculate the three-dimensional coordinates of the fracturing crack initiation point in the reservoir, and the normal vector of the fracturing main fracture surface is calculated according to the three-dimensional coordinate values ​​of the starting point and the ending point of the well trajectory, as follows: the starting point P1 of the well trajectory is set to (x1, y1, 1) and the ending point P2 is set to (x2, y2, 1), then the direction vector can be expressed as:

[0122]

[0123] Its length L can be calculated by the following formula:

[0124]

[0125] Final Normal Vector It can be expressed as a direction vector divided by its length:

[0126]

[0127] The above calculation results are combined with the fracturing crack data to obtain a complete fracturing crack data volume, including: the coordinate value of the fracturing crack initiation point (x 3i ,y 3i , z 3i ), the normal vector centered at the crack initiation point Cracked things half length f xi , crack height f yi , main fracture permeability k fi , permeability k of the fracture zone SRVi .

[0128] In another embodiment, step 3 includes: searching for n grid points closest to all wellbore points within the reservoir grid coordinate range using the wellbore coordinate data as the search point and performing encryption operations to obtain a wellbore grid coordinate point data body; searching for m grid points closest to all fracturing fractures within the reservoir grid coordinate range using the fracturing fracture data body as the search point and performing encryption operations to obtain a matrix-fracturing fracture grid coordinate point data body.

[0129] In this embodiment, the clustering model adopts the KNN model, and the specific process is as follows:

[0130] The reservoir grid coordinates (x 4i ,y 4i , z 4i ) data points are imported into the KNN model, with the wellbore (well trajectory) coordinate data (x 1i ,y 1i , z 1i) is the search point, and the K value range can be set to 10 to 50. The search result is the coordinate numbers of the 10 to 50 grid points closest to all wellbore points, including the wellbore point.

[0131] The reservoir grid coordinates (x 4i ,y 4i , z 4i ) data points are imported into the KNN model as Bool i (Extension range of the hydraulic fracture)×(x 4i ,y 4i , z 4i The non-zero data points obtained by multiplying the corresponding data bodies are the search points, and the K value range can be set to 10 to 50. The search results are the coordinate numbers of the 10 to 50 grid points closest to all the hydraulic fractures, including the hydraulic fractures and the matrix SRV areas.

[0132] Where Bool i The calculation formula of crack extension range is as follows:

[0133]

[0134] Among them, (x 3i ,y 3i , z 3i ) is the coordinate of the fracturing starting point, (x 4i ,y 4i , z 4i ) is the reservoir grid coordinate, l 粗 is the reservoir grid size, f xi is the half length of the crack, f yi is the crack height.

[0135] In another embodiment, the encryption operation includes: deleting the searched reservoir grid points, calculating the starting point of the local grid and the coordinates of all encrypted grids according to the size of the searched grid points and the encrypted grid parameters, and merging the retained reservoir grid point data volume with the encrypted grid coordinate point data volume to obtain a new grid coordinate point data volume. Specifically:

[0136] 1. Dig a hole. Dig out the selected reservoir grid point and delete its corresponding grid number and coordinate point value (x 4i ,y 4i , z 4i );

[0137] 2. Fill. According to the side length of the selected grid and the grid size parameters to be encrypted (such as the encryption multiple, the encrypted grid side length l 细 ), calculate the coordinates of the starting point of the local grid and the center point of all the encrypted grids;

[0138] 3. Merge. Merge the undeleted reservoir grid point data volume with the fine grid coordinate point data volume (i.e., the center point coordinates of all encrypted grids) to obtain a new grid coordinate point data volume, where the wellbore grid coordinate point data volume is represented by (x 4i ',y 4i ', z 4i '), the overall grid coordinate point data volume is represented by (x 4i ”,y 4i ”, z 4i ”).

[0139] In this embodiment, the encrypted grid parameters when the encryption operation is performed on the n reservoir grid points near the wellbore are the same as the encrypted grid parameters when the encryption operation is performed on the m reservoir grid points near the fracturing cracks. At this time, the generated reservoir model includes multi-scale modeling of two levels of coarse-fine grids.

[0140] In another embodiment, in step 3, the encrypted grid parameters when performing the encryption operation on the n reservoir grid points near the wellbore are different from the encrypted grid parameters when performing the encryption operation on the m reservoir grid points near the hydraulic fracture. The reservoir model generated at this time includes multi-scale modeling of three levels or multiple levels of coarse-medium-fine grids. Among them, fine grids can be used for encrypted modeling near the wellbore, medium-scale grids can be used for encrypted modeling of the matrix of the hydraulic fracture and its surrounding areas, and the rest are coarse-scale grids. To achieve this hierarchical encrypted modeling, it is only necessary to adjust the size of the encrypted grid in the "filling" step according to the selected grid.

[0141] Specific modeling effects such as Figure 4-6 As shown in the figure, the block grid is used as the basis for construction. The algorithm is fast and efficient and has strong adaptability to complex wellbore trajectory-fracture extension. The hierarchical grid is used to encrypt the wellbore-fracture-matrix grid structure, which can make more full use of logging data and well trajectory data to perform hierarchical and detailed modeling of the reservoir around the wellbore. The hierarchical multi-scale encrypted grid modeling can greatly reduce the total number of grids in the actual model while ensuring or even improving the modeling accuracy, further improving the accuracy of subsequent oil and gas well productivity analysis.

[0142] In another embodiment, a tracer monitoring wellbore-matrix-crack intelligent modeling system is provided, which is used to implement a fracturing directional wellbore-matrix-crack adaptive modeling method, including a data body generation module, a data body processing module, a reservoir grid model construction module, and an encryption modeling module; the data body generation module is used to generate an original data body according to the collected data, the data body processing module is used to process the original data body to respectively obtain the well trajectory coordinates, the logging data body and the fracturing crack data body, the reservoir grid model construction module is used to establish a reservoir grid model according to the logging data body, and the encryption modeling module is used to cluster and encrypt the reservoir grid model based on the logging data body and the fracturing crack data body to obtain a grid system corresponding to the wellbore-crack-matrix.

[0143] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring, characterized in that: include: Step 1: Collect basic parameters, well trajectory data, logging interpretation data and fracturing data; Step 2: Process the data collected in step 1 to determine the maximum and minimum values ​​of the three-dimensional coordinates of the reservoir, and establish a reservoir grid model to obtain the reservoir grid coordinates; Step 3: using an unsupervised machine learning algorithm to perform an encryption operation on n reservoir grid points near the wellbore to obtain a wellbore grid coordinate point data volume, and using an unsupervised machine learning algorithm to perform an encryption operation on m reservoir grid points near the hydraulic fracture to obtain a modeling matrix-hydraulic fracture grid coordinate point data volume, wherein the encryption operation is implemented by a clustering processing algorithm; Step 4: The wellbore grid coordinate point data volume and the matrix-fracturing fracture grid coordinate point data volume are combined as the overall grid coordinate point data volume, and the overall grid coordinate point data volume is clustered using an unsupervised machine learning algorithm to obtain the coupling connection relationship between the unencrypted grid, i.e., the coarse grid, and the encrypted grid, i.e., the fine grid. The overall grid coordinate point data volume and the coupling connection relationship between the coarse-fine grid are used as the reservoir model corresponding to the wellbore-fracture-matrix.

2. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 1, characterized in that: The step 3 comprises: using the wellbore coordinate data as the search point within the reservoir grid coordinate range to search for n grid points closest to all wellbore points and perform encryption operations to obtain a wellbore grid coordinate point data body; using the fracturing fracture data body as the search point within the reservoir grid coordinate range to search for m grid points closest to all fracturing fractures and perform encryption operations to obtain a matrix-fracturing fracture grid coordinate point data body.

3. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 2, characterized in that: The encryption operation includes: deleting the searched reservoir grid points, calculating the starting point of the local grid and the coordinates of all encrypted grids according to the size of the searched grid points and the encrypted grid parameters, and merging the retained reservoir grid point data body with the encrypted grid coordinate point data body to obtain a new grid coordinate point data body.

4. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 3, characterized in that: In step 3, the encrypted grid parameters when the encryption operation is performed on the n reservoir grid points near the wellbore are different from the encrypted grid parameters when the encryption operation is performed on the m reservoir grid points near the fracturing cracks.

5. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 2, characterized in that: The calculation process of the wellbore coordinate data includes: establishing interpolation functions between the drilling depth and the actual vertical depth, the east-west deviation, and the north-south deviation as interpolation functions of the three well trajectories, wherein the drilling depth d 1i and the actual vertical depth z i The interpolation function F between z (d 1i ) is as follows: A. Determine the complexity of the wellbore; B1. If it is a simple wellbore trajectory, divide the wellbore trajectory into multiple sections including vertical well section, inclined well section and horizontal section, and use linear or nonlinear function to fit in each section: z i =F z (d 1i )=fk(d 1i ),d 1i ∈[d 1k ,d 1k+1 ] Where, k is the number of well sections; B2. If the wellbore trajectory is complex, use exponential or logarithmic function for fitting: z i =F z (d 1i )=a·ln(d 1i )+b+c or z i =F z (d 1i )=a·exp(b·d 1i )+c Where a and b are fitting coefficients, and c is the spatial offset; B21. Add the well inclination angle to make corrections based on the exponential or logarithmic function: z i =F z (d 1i )=a·ln(d 1i )+b+c·cosθ i or z i =F z (d 1i )=a·exp(b·d 1i )+c·cosθ i Similarly, establish the drilling depth d 1i and east-west bias x i The interpolation function x between i =F x (d 1i ), drilling depth d 1i and north-south deflection y i The interpolation function y between i =F y (d 1i ); The complete well trajectory coordinate values ​​in the reservoir, namely the wellbore coordinate data, are calculated by interpolation based on the three well trajectory interpolation functions.

6. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 5, characterized in that: Acquiring reservoir grid coordinates in step 2 includes: Step 2.1: Substitute the logging depth value in the logging interpretation data into the three well trajectory interpolation functions to calculate the three-dimensional coordinates of the logging data points in the reservoir, and combine the calculated value with the logging interpretation data as a complete logging data body; Step 2.2: Determine the range of the three-dimensional coordinates based on the coordinate values ​​of the logging points in the complete logging data volume and then calculate the three-dimensional maximum length of the reservoir; Step 2.3: Set the side length of the reservoir grid and calculate the number of reservoir grids; Step 2.4: Establish a reservoir grid model and calculate the coordinates of the center point of each grid in the reservoir grid model as the reservoir grid coordinates.

7. The method for intelligent modeling of wellbore-matrix-fracture for tracing monitoring according to claim 5, characterized in that: The fracturing section depth in the fracturing data is substituted into the three well trajectory interpolation functions to calculate the three-dimensional coordinates of the fracturing start and end points in the reservoir, and the normal vector of the main fracturing fracture surface is calculated according to the three-dimensional coordinate values ​​of the starting point and the end point of the well trajectory. The above-mentioned calculated value is combined with the fracturing data to obtain a complete fracturing fracture data volume.

8. A method for intelligent modeling of wellbore-matrix-fractures for tracing monitoring according to claim 7, characterized in that: The calculation formula of the extension range of the crack is as follows: Among them, (x 3i ,y 3i , z 3i ) is the coordinate data of the fracturing starting point, (x 4i ,y 4i , z 4i ) is the reservoir grid coordinate, l is the reservoir grid size, f xi is the half length of the crack, f yi is the crack height.

9. The method for intelligent modeling of wellbore-matrix-fracture for tracing monitoring according to claim 1, characterized in that: The basic parameters include: coarse grid size l 粗 , fine mesh size l 细 , formation temperature, original formation pressure, viscosity and density of formation oil and water, apparent molecular weight of gas, injection and production time; The well trajectory data includes: drilling depth d 1i , actual vertical depth z i , east-west bias x i , north-south direction y i ; The logging interpretation data includes: logging depth d 2i , porosity Φ i , permeability k i , water saturation S wi ; The fracturing data include: fracturing section depth d 3i , the crack is half the length of f xi , crack height f yi , main fracture permeability k fi , permeability k of the fracture zone SRVi .

10. A tracing monitoring wellbore-matrix-fracture intelligent modeling system, characterized in that: A method for implementing a tracer monitoring wellbore-matrix-crack intelligent modeling method as described in any one of claims 1 to 9, comprising a data body generation module, a data body processing module, a reservoir grid model construction module, and an encryption modeling module; the data body generation module is used to generate an original data body according to the collected data, the data body processing module is used to process the original data body to respectively obtain well trajectory coordinates, logging data body and fracturing crack data body, the reservoir grid model construction module is used to establish a reservoir grid model according to the logging data body, and the encryption modeling module is used to cluster and encrypt the reservoir grid model based on the logging data body and the fracturing crack data body to obtain a grid system corresponding to the wellbore-crack-matrix.

Images