CA3152669A1 - Reservoir-based modeling method and device for pore network model - Google Patents
Reservoir-based modeling method and device for pore network model Download PDFInfo
- Publication number
- CA3152669A1 CA3152669A1 CA3152669A CA3152669A CA3152669A1 CA 3152669 A1 CA3152669 A1 CA 3152669A1 CA 3152669 A CA3152669 A CA 3152669A CA 3152669 A CA3152669 A CA 3152669A CA 3152669 A1 CA3152669 A1 CA 3152669A1
- Authority
- CA
- Canada
- Prior art keywords
- reservoir
- dimensional tensor
- data volume
- pore
- tensor data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000011148 porous material Substances 0.000 title claims abstract description 141
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000001228 spectrum Methods 0.000 claims abstract description 49
- 238000005481 NMR spectroscopy Methods 0.000 claims abstract description 19
- 238000003384 imaging method Methods 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims description 32
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 5
- 238000013527 convolutional neural network Methods 0.000 abstract 1
- 238000005755 formation reaction Methods 0.000 description 12
- 230000015572 biosynthetic process Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 10
- BVKZGUZCCUSVTD-UHFFFAOYSA-L Carbonate Chemical compound [O-]C([O-])=O BVKZGUZCCUSVTD-UHFFFAOYSA-L 0.000 description 6
- 230000035699 permeability Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 239000011435 rock Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004090 dissolution Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000002948 stochastic simulation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Computer Hardware Design (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Magnetic Resonance Imaging Apparatus (AREA)
- Image Analysis (AREA)
Abstract
A reservoir-based modeling method and device for a pore network model. The method comprises: obtaining a reservoir image obtained by means of electrical imaging well logging; obtaining a correlation length according to the reservoir image, and calculating a three-dimensional tensor convolution kernel according to the correlation length; obtaining a T2 spectrum of the reservoir obtained by nuclear magnetic resonance well logging, and obtaining pore-throat radius frequency distribution according to the T2 spectrum; forming an initial three-dimensional tensor data volume according to the pore-throat radius frequency distribution; generating, according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume and by means of a convolutional neural network forward propagation algorithm, a three-dimensional tensor data volume conforming to the pore-throat radius frequency distribution of the reservoir; constructing a disordered spatial structure of a pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the pore-throat radius frequency distribution of the reservoir and the disordered spatial structure.
Description
Reservoir-Based Modeling Method and Device For Pore Network Model Technical Field Embodiments of the present disclosure relate to, but are not limited to, the field of petroleum logging, in particular to a modeling method and a modeling device of a pore network model based on a reservoir.
Background With further development of oil and gas fields, distribution of oil and gas fluids (especially low permeability tight oil and gas reservoirs) underground is becoming more and more complicated, and geological modeling of oil reservoirs is an important part of oil reservoir research. Modeling methods are mainly divided into two aspects, namely a deterministic modeling and a stochastic modeling. The deterministic modeling is to provide deterministic prediction results for an unknown region between wells, while the stochastic modeling is to provide a variety of possible prediction results for an unknown region by applying a stochastic simulation method. However, in some reservoirs with strong heterogeneity, ultra-low permeability and tightness, especially in carbonate rock reservoirs with complex characteristics, strong heterogeneity and developed dissolution pores, due to limitation of interpolation accuracy and methods, there are still some deficiencies in analysis and research, and in some cases, there is also a lack of certain physical significance.
Summary The following is a summary of subject matters described in detail herein. This summary is not intended to limit the protection scope of the claims.
An embodiment of the present disclosure provides a modeling method of a pore network model based on a reservoir, including:
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length according to the reservoir image, and calculating a Date Recue/Date Received 2022-02-25 three-dimensional tensor convolution kernel according to the correlation length; wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network .. forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir and the disordered spatial structure.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
calculating the three-dimensional tensor convolution kernel E(h) according to the following formula:
E(h)= exp(-2h I Le) wherein h represents a distance from a spherical surface with ( Lõ,Ly , Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional L L L
coordinate system, L =. , LY =. , L = , and Lc represents the correlation length.
x 2 2 z 2 In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and the acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
Background With further development of oil and gas fields, distribution of oil and gas fluids (especially low permeability tight oil and gas reservoirs) underground is becoming more and more complicated, and geological modeling of oil reservoirs is an important part of oil reservoir research. Modeling methods are mainly divided into two aspects, namely a deterministic modeling and a stochastic modeling. The deterministic modeling is to provide deterministic prediction results for an unknown region between wells, while the stochastic modeling is to provide a variety of possible prediction results for an unknown region by applying a stochastic simulation method. However, in some reservoirs with strong heterogeneity, ultra-low permeability and tightness, especially in carbonate rock reservoirs with complex characteristics, strong heterogeneity and developed dissolution pores, due to limitation of interpolation accuracy and methods, there are still some deficiencies in analysis and research, and in some cases, there is also a lack of certain physical significance.
Summary The following is a summary of subject matters described in detail herein. This summary is not intended to limit the protection scope of the claims.
An embodiment of the present disclosure provides a modeling method of a pore network model based on a reservoir, including:
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length according to the reservoir image, and calculating a Date Recue/Date Received 2022-02-25 three-dimensional tensor convolution kernel according to the correlation length; wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network .. forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir and the disordered spatial structure.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
calculating the three-dimensional tensor convolution kernel E(h) according to the following formula:
E(h)= exp(-2h I Le) wherein h represents a distance from a spherical surface with ( Lõ,Ly , Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional L L L
coordinate system, L =. , LY =. , L = , and Lc represents the correlation length.
x 2 2 z 2 In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and the acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
2 Date Recue/Date Received 2022-02-25 acquiring n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology, where n is a positive integer; and summing the n T2 spectra, and converting an amplitude value of the summed T2 spectrum into the frequency distribution of the pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation is rm=cT2.;
wherein, r,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, the above method further has the following features:
forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
forming the initial three-dimensional tensor data volume by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of pore throat radius;
wherein the random function is the following lognormal distribution random function:
I
e-(Lnx-,11)2 I2cr2 f (x; P,a)¨
xcr-R- , wherein a mathematical expectation and a standard deviation a are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir by using the convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume includes:
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of the point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
In an exemplary embodiment, constructing the disordered spatial structure of the pore
In an exemplary embodiment, the preset quantitative relation is rm=cT2.;
wherein, r,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, the above method further has the following features:
forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
forming the initial three-dimensional tensor data volume by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of pore throat radius;
wherein the random function is the following lognormal distribution random function:
I
e-(Lnx-,11)2 I2cr2 f (x; P,a)¨
xcr-R- , wherein a mathematical expectation and a standard deviation a are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir by using the convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume includes:
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of the point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
In an exemplary embodiment, constructing the disordered spatial structure of the pore
3 Date Recue/Date Received 2022-02-25 network model according to the initial three-dimensional tensor data volume includes:
determining number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network including X, Y and Z directions according to the number of nodes and a spacing distance L between each two nodes;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
moving the coordinates (x,y,z) of each node according to the following formula:
(x,y,z) = [(i - OL rand( )%(0.540 -OL rand( )%(0.5L),(k ¨OL rand()%(0.5L)1 wherein, i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure includes:
assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to the nodes of the disordered spatial
determining number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network including X, Y and Z directions according to the number of nodes and a spacing distance L between each two nodes;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
moving the coordinates (x,y,z) of each node according to the following formula:
(x,y,z) = [(i - OL rand( )%(0.540 -OL rand( )%(0.5L),(k ¨OL rand()%(0.5L)1 wherein, i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure includes:
assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to the nodes of the disordered spatial
4 Date Recue/Date Received 2022-02-25 structure sequentially to establish the pore network model.
An embodiment of the present disclosure further provides a modeling device of a pore network model based on a reservoir, which includes a memory and a processor;
the memory is configured to store a program for modeling the pore network model of the reservoir;
the processor is configured to read and execute the program for modeling the pore network model of the reservoir, and execute the following modeling method:
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length according to the reservoir image, and calculating a three-dimensional tensor convolution kernel according to the correlation length; wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
calculating the three-dimensional tensor convolution kernel E(h) according to the
An embodiment of the present disclosure further provides a modeling device of a pore network model based on a reservoir, which includes a memory and a processor;
the memory is configured to store a program for modeling the pore network model of the reservoir;
the processor is configured to read and execute the program for modeling the pore network model of the reservoir, and execute the following modeling method:
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length according to the reservoir image, and calculating a three-dimensional tensor convolution kernel according to the correlation length; wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
calculating the three-dimensional tensor convolution kernel E(h) according to the
5 Date Recue/Date Received 2022-02-25 following formula:
E(h)= exp (-2h 1 Lc) wherein h represents a distance from a spherical surface with ( Lx , Ly , Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional L L L
coordinate system, L =. LY =. , L = , and Lc represents the correlation length.
z 2 In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
acquiring n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology, where n is a positive integer; and summing the n T2 spectra, and converting an amplitude value of the summed T2 spectrum into the frequency distribution of pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation is rm=cT2.;
wherein r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
forming the initial three-dimensional tensor data volume by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of the pore throat radius;
wherein the random function is the following lognormal distribution random function:
e-( Ln x¨,02 /26-2 xo--2-i-' wherein a mathematical expectation 1.t and a standard deviation a are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir by using the
E(h)= exp (-2h 1 Lc) wherein h represents a distance from a spherical surface with ( Lx , Ly , Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional L L L
coordinate system, L =. LY =. , L = , and Lc represents the correlation length.
z 2 In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
acquiring n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology, where n is a positive integer; and summing the n T2 spectra, and converting an amplitude value of the summed T2 spectrum into the frequency distribution of pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation is rm=cT2.;
wherein r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
forming the initial three-dimensional tensor data volume by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of the pore throat radius;
wherein the random function is the following lognormal distribution random function:
e-( Ln x¨,02 /26-2 xo--2-i-' wherein a mathematical expectation 1.t and a standard deviation a are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir by using the
6 Date Recue/Date Received 2022-02-25 convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume includes:
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir.
In an exemplary embodiment, constructing the disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume includes:
determining number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network including X, Y and Z directions according to the number of nodes and a spacing distance L between each node;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
moving the coordinates (x,y,z) of each node according to the following formula:
(x,y,z) = [(i ¨ OL rand( )%(0.54(j ¨OL rand( )%(0.5L),(k ¨OL
rand()%(0.5L)1
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir.
In an exemplary embodiment, constructing the disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume includes:
determining number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network including X, Y and Z directions according to the number of nodes and a spacing distance L between each node;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
moving the coordinates (x,y,z) of each node according to the following formula:
(x,y,z) = [(i ¨ OL rand( )%(0.54(j ¨OL rand( )%(0.5L),(k ¨OL
rand()%(0.5L)1
7 Date Recue/Date Received 2022-02-25 wherein i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, the establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir and the disordered spatial structure, includes:
assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to the nodes of the disordered spatial structure sequentially to establish the pore network model.
Other aspects will become apparent upon reading and understanding of the drawings and detailed description.
Brief Description of Drawings FIG. 1 is a schematic diagram of a modeling method of a pore network model based on a .. reservoir according to an embodiment of the present disclosure.
FIG. 2 is an example of an image of a sandstone formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 3 is an example of an image of a conglomerate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 4 is an example of an image of a carbonate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 5 is an example of a convolution kernel.
FIG. 6 is an example of a T2 spectrum obtained by nuclear magnetic resonance logging according to an embodiment of the present disclosure.
FIG. 7 is an example of a frequency distribution curve of a pore throat radius according to an embodiment of the present disclosure.
FIG. 8 is an example of an initial three-dimensional tensor data volume according to an embodiment of the present disclosure.
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, the establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of pore throat radius of the reservoir and the disordered spatial structure, includes:
assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to the nodes of the disordered spatial structure sequentially to establish the pore network model.
Other aspects will become apparent upon reading and understanding of the drawings and detailed description.
Brief Description of Drawings FIG. 1 is a schematic diagram of a modeling method of a pore network model based on a .. reservoir according to an embodiment of the present disclosure.
FIG. 2 is an example of an image of a sandstone formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 3 is an example of an image of a conglomerate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 4 is an example of an image of a carbonate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
FIG. 5 is an example of a convolution kernel.
FIG. 6 is an example of a T2 spectrum obtained by nuclear magnetic resonance logging according to an embodiment of the present disclosure.
FIG. 7 is an example of a frequency distribution curve of a pore throat radius according to an embodiment of the present disclosure.
FIG. 8 is an example of an initial three-dimensional tensor data volume according to an embodiment of the present disclosure.
8 Date Recue/Date Received 2022-02-25 FIG. 9 is a schematic diagram of an example of generating a three-dimensional tensor data volume conforming to a frequency distribution of pore throat radius according to an embodiment of the present disclosure.
FIG. 10a is an example of a structural diagram of an ordered spatial according to an embodiment of the present disclosure.
FIG. 10b is an example of a structural diagram of a disordered spatial according to an embodiment of the present disclosure.
FIG. 11 is an example of a pore network model according to an embodiment of the present disclosure.
FIG. 12 is a schematic diagram of a modeling device of a pore network model based on a reservoir according to an embodiment of the present disclosure.
Detailed Description Hereinafter, embodiments of the present disclosure will be described in detail with reference to accompanying drawings. It is to be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other randomly if there is no conflict.
FIG. 1 is a schematic diagram of a modeling method of a pore network model based on a reservoir according to an embodiment of the present disclosure. As shown in FIG. 1, the modeling method of this embodiment includes step S11- S16.
In S11, a reservoir image obtained by electrical imaging logging is acquired.
In S12, a correlation length is acquired according to the reservoir image, and a three-dimensional tensor convolution kernel is calculated according to the correlation length.
In S13, a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging is acquired, and a frequency distribution of a pore throat radius is acquired according to the T2 spectrum; and an initial three-dimensional tensor data volume is formed according to the frequency distribution of the pore throat radius.
In S14, a three-dimensional tensor data volume conforming to the frequency distribution
FIG. 10a is an example of a structural diagram of an ordered spatial according to an embodiment of the present disclosure.
FIG. 10b is an example of a structural diagram of a disordered spatial according to an embodiment of the present disclosure.
FIG. 11 is an example of a pore network model according to an embodiment of the present disclosure.
FIG. 12 is a schematic diagram of a modeling device of a pore network model based on a reservoir according to an embodiment of the present disclosure.
Detailed Description Hereinafter, embodiments of the present disclosure will be described in detail with reference to accompanying drawings. It is to be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other randomly if there is no conflict.
FIG. 1 is a schematic diagram of a modeling method of a pore network model based on a reservoir according to an embodiment of the present disclosure. As shown in FIG. 1, the modeling method of this embodiment includes step S11- S16.
In S11, a reservoir image obtained by electrical imaging logging is acquired.
In S12, a correlation length is acquired according to the reservoir image, and a three-dimensional tensor convolution kernel is calculated according to the correlation length.
In S13, a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging is acquired, and a frequency distribution of a pore throat radius is acquired according to the T2 spectrum; and an initial three-dimensional tensor data volume is formed according to the frequency distribution of the pore throat radius.
In S14, a three-dimensional tensor data volume conforming to the frequency distribution
9 Date Recue/Date Received 2022-02-25 of the pore throat radius of the reservoir is generated according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume by using a convolution neural network forward propagation algorithm.
In S15, a disordered spatial structure of the pore network model is constructed according to the initial three-dimensional tensor data volume.
In S16, the pore network model is established according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
Among the above steps, S13 may be performed before or after Sll and S12, and may be performed in parallel with Sll and S12. The sequential order of S14 and S15 is not limited, and they may be performed in parallel.
In an exemplary embodiment, the three-dimensional tensor convolution kernel E(h) may be calculated according to the following formula:
E(h)= exp (-2h 1 Lc) where h represents a distance from a spherical surface with (Lõ,L,,Lz) as a spherical center and a radius smaller than or equal to Lc to a spherical center in a three-dimensional L L L
coordinate system, L =. ' L' =. , L =. , and Lc represents the correlation length.
x 2 2 z 2 In an exemplary embodiment, n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology may be acquired, where n is a positive integer;
and the n T2 spectra are summed, and an amplitude value of the summed T2 spectrum is converted into the frequency distribution of the pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation may be rm=cT2.;
where r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, the initial three-dimensional tensor data volume may be formed by establishing, through a random function, a three-dimensional stable random field Date Recue/Date Received 2022-02-25 according to the frequency distribution of the pore throat radius, where the random function is the following lognormal distribution random function:
____________________________________________ e-i[nx-,02 /2c?
where a mathematical expectation Ia. and a standard deviation 6 are obtained by fitting the frequency distribution of the pore throat radius, and x represents the pore throat radius.
In an exemplary embodiment, tensor point multiplication may be performed on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, a point multiplication result is stacked according to an order in the initial three-dimensional tensor data volume, and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated.
In an exemplary embodiment, the number of nodes of the disordered spatial structure is determined according to data amount of the initial three-dimensional tensor data volume; a three-dimensional cube network including X, Y and Z directions is constructed according to the number of nodes and a spacing distance L between each two nodes; coordinates of each node in the three-dimensional cube network are calculated; it is determined whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and a tube bundle radius is distributed; it is determined whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and a tube bundle radius is distributed;
the coordinates of each node are moved based on a preset rule; the disordered spatial structure is generated according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, the coordinates (x,y,z) of each node may be moved according to the following formula:
(x,y,z) = [(i ¨1)L rand ()%(0.54(j ¨1)L rand ( )%(0.51,),(k ¨1)L rand ( )%(0.5L)]
where i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) Date Recue/Date Received 2022-02-25 represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir may be assigned to the nodes of the disordered spatial structure sequentially to establish the pore network model.
FIG. 2 is an image of a sandstone formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure. FIG. 3 is an image of a conglomerate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure. FIG. 4 is an image of a carbonate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
The above three formations (i.e. reservoirs) each have a thickness of 0.5m.
Some black masses in the figures are circled. An average value of the radii of one or more black masses is the correlation length L. A value of Lc may be identified and measured by ImageJ, it is not necessary to identify each black mass, only one or more black masses need to be identified by sampling to obtain their average value. It may be seen from FIG. 2 to FIG. 4, different reservoirs have different correlation lengths, generally for sandstone <
conglomerate <
carbonate. The three-dimensional tensor convolution kernel may be calculated by a covariance model in geostatistics according to the correlation length.
Among them, the three-dimensional tensor convolution kernel is E(h) = exp (-2h / Le ), Lc is the correlation length, which is related to lithology, and is obtained according to electrical imaging analysis; h represents a distance from a spherical surface with (Lõ,Ly,Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a L L three-dimensional coordinate system, L = , LY = , L2 L = .
For example, the x 2 2 2 three-dimensional convolution kernel (as shown in FIG. 5) may be achieved by the following program.
Lcx=Lc; Lcy=Lc; Lcz=Lc;
Kernel=zeros(Lcx,Lcy,Lcz); //Generate a zero matrix of Lc*Lc*Lc.
For i=1:Lcx For j=1:Lcy Date Recue/Date Received 2022-02-25 For k=1:Lcz distance=sqrt(power((Lc+1)/24,2.0)+
power((Lc+1)/2-j,2.0)+
power((Lc+1)/2-k,2.0));
kernel(Ij ,k)=exp(-2*di stance/Lc);
end end end.
FIG. 6 is an example of a T2 spectrum obtained by nuclear magnetic resonance logging according to an embodiment of the present disclosure. There are n T2 spectra in the figure. The n T2 spectra are obtained according to the following method: a formation (i.e.
reservoir) with a thickness of 0.5m is divided into n sub-formation (i.e. sub-reservoirs), and one T2 spectrum is obtained for each sub-formation by nuclear magnetic resonance logging. Among them, n is set by a user as required.
FIG. 7 is an example of a frequency distribution curve of a pore throat radius according to an embodiment of the present disclosure. A total T2 spectrum is obtained by accumulating n T2 spectra. An amplitude value of the T2 spectrum is converted into the frequency distribution of the pore throat radius by a quantitative relation. The quantitative relation may be rm=cT2., where r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer. A large number of statistics worldwide show that a pore throat length is generally distributed in a range of 50-200 microns, and the lower the permeability is, the longer the pore throat length has.
FIG. 8 is an example of an initial three-dimensional tensor data volume according to an embodiment of the present disclosure. The initial three-dimensional tensor data volume may be formed by establishing a three-dimensional stable random field through a random function (Log uniform or Log normal, etc.) according to the frequency distribution curve of the pore throat radius.
An average value (i.e. mathematical expectation) )t. and a standard deviation a of the frequency distribution of the pore throat radius may be obtained by fitting the frequency Date Recue/Date Received 2022-02-25 distribution curve of the pore throat radius by Matlab, thus the following lognormal distribution random function is obtained:
e-( Lnx¨p)2 /2crz .y0-7-1-where x represents a pore throat radius.
FIG. 9 is a schematic diagram of generating a three-dimensional tensor data volume conforming to a frequency distribution of a pore throat radius according to an embodiment of the present disclosure. A constructed three-dimensional tensor convolution kernel is used to slide through a three-dimensional tensor data volume, to perform tensor point multiplication calculation sequentially, wherein one point is obtained each time. Finally, result values of all channels are stacked in an original order to form a new three-dimensional tensor, which forms a three-dimensional tensor data volume that conforms to distribution frequency characteristics of the pore throat radius of the reservoir. After the data is input into python, a generated micro-pore-dissolution pore carbonate model image may be observed by vtk.
FIG. 10a is an example of a diagram of an ordered spatial structure according to an embodiment of the present disclosure. FIG. 10b is an example of a diagram of a disordered spatial structure according to an embodiment of the present disclosure. A
construction process of the disordered spatial structure is as follows:
(1) The number of nodes of a model is specified (the number of nodes is determined according to a size and refinement degree of the model, and its value corresponds to the tensor data volume) to construct a simple three-dimensional cube grid of XxYxZ. Each node represents a pore, and the nodes are connected by throat channels. There are six throat channels connected around each node representing a pore in the established network.
Similarly, there are six pores connected around each throat channel. In each direction (i.e. X
direction, Y direction and Z direction), a spacing distance between two nodes is set to 1, the number of the nodes is set to d, and a side length of the model is (d-1) x 1.
(2) Coordinates of each node in the network model are calculated. A
calculation formula thereof is: (x, y, z)=[(i-1)1, (j -1)1, (k-1)11, where i, j and k in the formula are the node numbers in the X direction, Y direction and Z direction respectively, and their values are 1,2,3 ...
Date Recue/Date Received 2022-02-25 (3) A probability function with a probability of p is set in a program, and it is determined whether there is a tube bundle connected between each two adjacent nodes in the X, Y and Z
directions by a (pseudo) random number generator. The function is used to generate random numbers, thus a random probability is generated. For example, in C/C++
programming language, a rand () function may be used to generate random numbers, thus the random probability is generated. The specific C/C++ code is:
U' (rand( )%100 < p x 100) where, rand()%100 indicates that a computer randomly generates any integer in a range of 0-99.
When a connection probability p is 50%, there is a 50% possibility (probability) that the integers randomly generated by the rand() function are smaller than 50, and otherwise there is a 50% possibility (probability) that the integers randomly generated by the rand() function are greater than 50. Therefore, the expression may achieve a tube bundle connection with a probability p=50%, that is, when a number smaller than 50 is generated, the expression is true, and assignment of a tube bundle radius is performed (the tube bundle radius is assigned as r).
When a number no smaller than 50 is generated, the expression is false, and no operation is performed. Then, an ordered spatial structure is generated according to the three-dimensional cube network, each node, a result of determining whether there is the tube bundle connection and the tube bundle radius (as shown in FIG. 10a).
A connectivity of the network model may also be quantitatively described by a coordination number. A coordination number of conventional sandstone reservoir rocks is in a range of 4 to 6, and a coordination number of low permeability sandstone reservoir rocks is in a range of 3 to 4. A coordination number of 6 refers to the tube bundle connection between one node and other six nodes.
(4) A random network (i.e. a disordered spatial structure) is generated by moving coordinates of a node (adding a random number not exceeding a throat channel length at the left of the center point) (as shown in FIG. 10b), and the moved coordinates of the node are (x,y,z)=[(i-1)1+(rand() % (0.51),(j-1)1+(rand() % (0.51),(k-1)1+(rand() %
(0.51)1.
FIG. 11 is an example of a pore network model according to an embodiment of the Date Recue/Date Received 2022-02-25 present disclosure. A pore network model conforming to real geological characteristics of a reservoir may be constructed by sequentially assigning data in a three-dimensional tensor data volume conforming to frequency distribution characteristics of pore throat radius of the reservoir to nodes of a disordered spatial structure.
In the above modeling process, a workstation with 8 graphics cards and 512GB
memory may be used for multi-GPU calculation, thus the scale of the model (a range of about 100 meters, a thickness of about 10 meters, and a total number of pore nodes exceeding 1 billion) is expanded, and the pore network model of single-well oil and gas reservoir that conforms to the reservoir image and the characteristics of the pore throat is established.
The pore network model established in the embodiment of the present disclosure mainly relies on the methods of electrical imaging logging and nuclear magnetic resonance logging, in which the physical parameters of the reservoir are taken into account, and the problem that the established geological model lacks certain physical significance is solved to a certain extent.
Moreover, it solves the bottleneck problem of difficult in formation modeling with strong heterogeneity, and is also suitable for carbonate formations, which also provides a certain reference for modeling of low permeability tight oil and gas reservoirs.
FIG. 12 is a schematic diagram of a modeling device of a pore network model based on a reservoir according to an embodiment of the present disclosure. The modeling device of the pore network model based on the reservoir includes a memory and a processor.
The memory is configured to store a program for modeling the pore network model of the reservoir.
The processor is configured to read and execute the program for modeling the pore network model of the reservoir, and execute the following modeling method:
a reservoir image obtained by electrical imaging logging is acquired;
a correlation length is acquired according to the reservoir image, and a three-dimensional tensor convolution kernel is calculated according to the correlation length, wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
Date Recue/Date Received 2022-02-25 a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging is acquired, and a frequency distribution of a pore throat radius is acquired according to the T2 spectrum, and an initial three-dimensional tensor data volume is formed according to the frequency distribution of the pore throat radius;
a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
a disordered spatial structure of the pore network model is constructed according to the initial three-dimensional tensor data volume; and the pore network model is established according to the disordered spatial structure and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
the three-dimensional tensor convolution kernel E(h) is calculated according to the following formula:
E(h)= exp (-2h I Le) wherein, h represents a distance from a spherical surface with (1,, , Ly ,L,) as a spherical center and a radius smaller than or equal to I, to the spherical center in a L L
three-dimensional coordinate system, L = ' LY L = . , L2 = . , and Lc represents the x 2 2 2 correlation length.
In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
N T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology are acquired, where n is a positive integer; and Date Recue/Date Received 2022-02-25 the n T2 spectra are summed, and an amplitude value of the summed T2 spectrum is converted into the frequency distribution of the pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation is rm=cT2m;
where r,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
the initial three-dimensional tensor data volume may be formed by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of the pore throat radius, where the random function is the following lognormal distribution random function:
e-( Lnx¨p)2 /26-2 f (x; p,o)= _________________________________ xo-Aff where a mathematical expectation II and a standard deviation G are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume includes:
tensor point multiplication may be performed on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, results of the point multiplication are stacked according to an order in the initial three-dimensional tensor data volume, and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated.
In an exemplary embodiment, constructing the disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume includes:
the number of nodes of the disordered spatial structure is determined according to data Date Recue/Date Received 2022-02-25 amount of the initial three-dimensional tensor data volume;
a three-dimensional cube network including X, Y and Z directions is constructed according to the number of the nodes and a spacing distance L between each two nodes;
coordinates of each node in the three-dimensional cube network are calculated;
it is determined whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and a tube bundle radius is distributed;
it is determined whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and a tube bundle radius is distributed;
the coordinates of each node are moved based on a preset rule; and the disordered spatial structure is generated according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
the coordinates (x,y,z) of each node may be moved according to the following formula:
(x,y,z) = [(i ¨1)L rand( )%(0.51,),(j ¨1)L rand( )%(0.54(k ¨1)L rand( )%(0.5L)1 where i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, establishing the pore network model according to the three-dimensional tensor data volume conforming to the disordered spatial structure and the frequency distribution of the pore throat radius of the reservoir includes:
the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is assigned to the nodes of the disordered spatial structure sequentially to establish the pore network model.
Those skilled in the art may understand that all or part of the steps in the above method Date Recue/Date Received 2022-02-25 may be completed by instructing the relevant hardware through a program, which may be stored in computer-readable storage media, such as read-only memory, magnetic disk or optical disk, etc. Optionally, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Accordingly, the modules/units in the above embodiments may be implemented in the form of hardware or software functional modules.
The present disclosure is not limited to any specific combination of hardware and software.
There are many other embodiments of the present disclosure. Without departing from the spirit and essence of the present disclosure, those skilled in the art may make various respective changes and variations according to the present disclosure, but these respective changes and variations should be covered by the protection scope of the appended claims of the present disclosure.
Date Recue/Date Received 2022-02-25
In S15, a disordered spatial structure of the pore network model is constructed according to the initial three-dimensional tensor data volume.
In S16, the pore network model is established according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
Among the above steps, S13 may be performed before or after Sll and S12, and may be performed in parallel with Sll and S12. The sequential order of S14 and S15 is not limited, and they may be performed in parallel.
In an exemplary embodiment, the three-dimensional tensor convolution kernel E(h) may be calculated according to the following formula:
E(h)= exp (-2h 1 Lc) where h represents a distance from a spherical surface with (Lõ,L,,Lz) as a spherical center and a radius smaller than or equal to Lc to a spherical center in a three-dimensional L L L
coordinate system, L =. ' L' =. , L =. , and Lc represents the correlation length.
x 2 2 z 2 In an exemplary embodiment, n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology may be acquired, where n is a positive integer;
and the n T2 spectra are summed, and an amplitude value of the summed T2 spectrum is converted into the frequency distribution of the pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation may be rm=cT2.;
where r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, the initial three-dimensional tensor data volume may be formed by establishing, through a random function, a three-dimensional stable random field Date Recue/Date Received 2022-02-25 according to the frequency distribution of the pore throat radius, where the random function is the following lognormal distribution random function:
____________________________________________ e-i[nx-,02 /2c?
where a mathematical expectation Ia. and a standard deviation 6 are obtained by fitting the frequency distribution of the pore throat radius, and x represents the pore throat radius.
In an exemplary embodiment, tensor point multiplication may be performed on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, a point multiplication result is stacked according to an order in the initial three-dimensional tensor data volume, and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated.
In an exemplary embodiment, the number of nodes of the disordered spatial structure is determined according to data amount of the initial three-dimensional tensor data volume; a three-dimensional cube network including X, Y and Z directions is constructed according to the number of nodes and a spacing distance L between each two nodes; coordinates of each node in the three-dimensional cube network are calculated; it is determined whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and a tube bundle radius is distributed; it is determined whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and a tube bundle radius is distributed;
the coordinates of each node are moved based on a preset rule; the disordered spatial structure is generated according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, the coordinates (x,y,z) of each node may be moved according to the following formula:
(x,y,z) = [(i ¨1)L rand ()%(0.54(j ¨1)L rand ( )%(0.51,),(k ¨1)L rand ( )%(0.5L)]
where i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) Date Recue/Date Received 2022-02-25 represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir may be assigned to the nodes of the disordered spatial structure sequentially to establish the pore network model.
FIG. 2 is an image of a sandstone formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure. FIG. 3 is an image of a conglomerate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure. FIG. 4 is an image of a carbonate formation reservoir obtained by electrical imaging logging according to an embodiment of the present disclosure.
The above three formations (i.e. reservoirs) each have a thickness of 0.5m.
Some black masses in the figures are circled. An average value of the radii of one or more black masses is the correlation length L. A value of Lc may be identified and measured by ImageJ, it is not necessary to identify each black mass, only one or more black masses need to be identified by sampling to obtain their average value. It may be seen from FIG. 2 to FIG. 4, different reservoirs have different correlation lengths, generally for sandstone <
conglomerate <
carbonate. The three-dimensional tensor convolution kernel may be calculated by a covariance model in geostatistics according to the correlation length.
Among them, the three-dimensional tensor convolution kernel is E(h) = exp (-2h / Le ), Lc is the correlation length, which is related to lithology, and is obtained according to electrical imaging analysis; h represents a distance from a spherical surface with (Lõ,Ly,Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a L L three-dimensional coordinate system, L = , LY = , L2 L = .
For example, the x 2 2 2 three-dimensional convolution kernel (as shown in FIG. 5) may be achieved by the following program.
Lcx=Lc; Lcy=Lc; Lcz=Lc;
Kernel=zeros(Lcx,Lcy,Lcz); //Generate a zero matrix of Lc*Lc*Lc.
For i=1:Lcx For j=1:Lcy Date Recue/Date Received 2022-02-25 For k=1:Lcz distance=sqrt(power((Lc+1)/24,2.0)+
power((Lc+1)/2-j,2.0)+
power((Lc+1)/2-k,2.0));
kernel(Ij ,k)=exp(-2*di stance/Lc);
end end end.
FIG. 6 is an example of a T2 spectrum obtained by nuclear magnetic resonance logging according to an embodiment of the present disclosure. There are n T2 spectra in the figure. The n T2 spectra are obtained according to the following method: a formation (i.e.
reservoir) with a thickness of 0.5m is divided into n sub-formation (i.e. sub-reservoirs), and one T2 spectrum is obtained for each sub-formation by nuclear magnetic resonance logging. Among them, n is set by a user as required.
FIG. 7 is an example of a frequency distribution curve of a pore throat radius according to an embodiment of the present disclosure. A total T2 spectrum is obtained by accumulating n T2 spectra. An amplitude value of the T2 spectrum is converted into the frequency distribution of the pore throat radius by a quantitative relation. The quantitative relation may be rm=cT2., where r ,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer. A large number of statistics worldwide show that a pore throat length is generally distributed in a range of 50-200 microns, and the lower the permeability is, the longer the pore throat length has.
FIG. 8 is an example of an initial three-dimensional tensor data volume according to an embodiment of the present disclosure. The initial three-dimensional tensor data volume may be formed by establishing a three-dimensional stable random field through a random function (Log uniform or Log normal, etc.) according to the frequency distribution curve of the pore throat radius.
An average value (i.e. mathematical expectation) )t. and a standard deviation a of the frequency distribution of the pore throat radius may be obtained by fitting the frequency Date Recue/Date Received 2022-02-25 distribution curve of the pore throat radius by Matlab, thus the following lognormal distribution random function is obtained:
e-( Lnx¨p)2 /2crz .y0-7-1-where x represents a pore throat radius.
FIG. 9 is a schematic diagram of generating a three-dimensional tensor data volume conforming to a frequency distribution of a pore throat radius according to an embodiment of the present disclosure. A constructed three-dimensional tensor convolution kernel is used to slide through a three-dimensional tensor data volume, to perform tensor point multiplication calculation sequentially, wherein one point is obtained each time. Finally, result values of all channels are stacked in an original order to form a new three-dimensional tensor, which forms a three-dimensional tensor data volume that conforms to distribution frequency characteristics of the pore throat radius of the reservoir. After the data is input into python, a generated micro-pore-dissolution pore carbonate model image may be observed by vtk.
FIG. 10a is an example of a diagram of an ordered spatial structure according to an embodiment of the present disclosure. FIG. 10b is an example of a diagram of a disordered spatial structure according to an embodiment of the present disclosure. A
construction process of the disordered spatial structure is as follows:
(1) The number of nodes of a model is specified (the number of nodes is determined according to a size and refinement degree of the model, and its value corresponds to the tensor data volume) to construct a simple three-dimensional cube grid of XxYxZ. Each node represents a pore, and the nodes are connected by throat channels. There are six throat channels connected around each node representing a pore in the established network.
Similarly, there are six pores connected around each throat channel. In each direction (i.e. X
direction, Y direction and Z direction), a spacing distance between two nodes is set to 1, the number of the nodes is set to d, and a side length of the model is (d-1) x 1.
(2) Coordinates of each node in the network model are calculated. A
calculation formula thereof is: (x, y, z)=[(i-1)1, (j -1)1, (k-1)11, where i, j and k in the formula are the node numbers in the X direction, Y direction and Z direction respectively, and their values are 1,2,3 ...
Date Recue/Date Received 2022-02-25 (3) A probability function with a probability of p is set in a program, and it is determined whether there is a tube bundle connected between each two adjacent nodes in the X, Y and Z
directions by a (pseudo) random number generator. The function is used to generate random numbers, thus a random probability is generated. For example, in C/C++
programming language, a rand () function may be used to generate random numbers, thus the random probability is generated. The specific C/C++ code is:
U' (rand( )%100 < p x 100) where, rand()%100 indicates that a computer randomly generates any integer in a range of 0-99.
When a connection probability p is 50%, there is a 50% possibility (probability) that the integers randomly generated by the rand() function are smaller than 50, and otherwise there is a 50% possibility (probability) that the integers randomly generated by the rand() function are greater than 50. Therefore, the expression may achieve a tube bundle connection with a probability p=50%, that is, when a number smaller than 50 is generated, the expression is true, and assignment of a tube bundle radius is performed (the tube bundle radius is assigned as r).
When a number no smaller than 50 is generated, the expression is false, and no operation is performed. Then, an ordered spatial structure is generated according to the three-dimensional cube network, each node, a result of determining whether there is the tube bundle connection and the tube bundle radius (as shown in FIG. 10a).
A connectivity of the network model may also be quantitatively described by a coordination number. A coordination number of conventional sandstone reservoir rocks is in a range of 4 to 6, and a coordination number of low permeability sandstone reservoir rocks is in a range of 3 to 4. A coordination number of 6 refers to the tube bundle connection between one node and other six nodes.
(4) A random network (i.e. a disordered spatial structure) is generated by moving coordinates of a node (adding a random number not exceeding a throat channel length at the left of the center point) (as shown in FIG. 10b), and the moved coordinates of the node are (x,y,z)=[(i-1)1+(rand() % (0.51),(j-1)1+(rand() % (0.51),(k-1)1+(rand() %
(0.51)1.
FIG. 11 is an example of a pore network model according to an embodiment of the Date Recue/Date Received 2022-02-25 present disclosure. A pore network model conforming to real geological characteristics of a reservoir may be constructed by sequentially assigning data in a three-dimensional tensor data volume conforming to frequency distribution characteristics of pore throat radius of the reservoir to nodes of a disordered spatial structure.
In the above modeling process, a workstation with 8 graphics cards and 512GB
memory may be used for multi-GPU calculation, thus the scale of the model (a range of about 100 meters, a thickness of about 10 meters, and a total number of pore nodes exceeding 1 billion) is expanded, and the pore network model of single-well oil and gas reservoir that conforms to the reservoir image and the characteristics of the pore throat is established.
The pore network model established in the embodiment of the present disclosure mainly relies on the methods of electrical imaging logging and nuclear magnetic resonance logging, in which the physical parameters of the reservoir are taken into account, and the problem that the established geological model lacks certain physical significance is solved to a certain extent.
Moreover, it solves the bottleneck problem of difficult in formation modeling with strong heterogeneity, and is also suitable for carbonate formations, which also provides a certain reference for modeling of low permeability tight oil and gas reservoirs.
FIG. 12 is a schematic diagram of a modeling device of a pore network model based on a reservoir according to an embodiment of the present disclosure. The modeling device of the pore network model based on the reservoir includes a memory and a processor.
The memory is configured to store a program for modeling the pore network model of the reservoir.
The processor is configured to read and execute the program for modeling the pore network model of the reservoir, and execute the following modeling method:
a reservoir image obtained by electrical imaging logging is acquired;
a correlation length is acquired according to the reservoir image, and a three-dimensional tensor convolution kernel is calculated according to the correlation length, wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
Date Recue/Date Received 2022-02-25 a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging is acquired, and a frequency distribution of a pore throat radius is acquired according to the T2 spectrum, and an initial three-dimensional tensor data volume is formed according to the frequency distribution of the pore throat radius;
a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
a disordered spatial structure of the pore network model is constructed according to the initial three-dimensional tensor data volume; and the pore network model is established according to the disordered spatial structure and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
In an exemplary embodiment, calculating the three-dimensional tensor convolution kernel according to the correlation length includes:
the three-dimensional tensor convolution kernel E(h) is calculated according to the following formula:
E(h)= exp (-2h I Le) wherein, h represents a distance from a spherical surface with (1,, , Ly ,L,) as a spherical center and a radius smaller than or equal to I, to the spherical center in a L L
three-dimensional coordinate system, L = ' LY L = . , L2 = . , and Lc represents the x 2 2 2 correlation length.
In an exemplary embodiment, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and acquiring the frequency distribution of the pore throat radius according to the T2 spectrum includes:
N T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology are acquired, where n is a positive integer; and Date Recue/Date Received 2022-02-25 the n T2 spectra are summed, and an amplitude value of the summed T2 spectrum is converted into the frequency distribution of the pore throat radius through a preset quantitative relation.
In an exemplary embodiment, the preset quantitative relation is rm=cT2m;
where r,, is a m-th pore throat radius, T2,, is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
In an exemplary embodiment, forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius includes:
the initial three-dimensional tensor data volume may be formed by establishing, through a random function, a three-dimensional stable random field according to the frequency distribution of the pore throat radius, where the random function is the following lognormal distribution random function:
e-( Lnx¨p)2 /26-2 f (x; p,o)= _________________________________ xo-Aff where a mathematical expectation II and a standard deviation G are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
In an exemplary embodiment, generating the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume includes:
tensor point multiplication may be performed on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, results of the point multiplication are stacked according to an order in the initial three-dimensional tensor data volume, and the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is generated.
In an exemplary embodiment, constructing the disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume includes:
the number of nodes of the disordered spatial structure is determined according to data Date Recue/Date Received 2022-02-25 amount of the initial three-dimensional tensor data volume;
a three-dimensional cube network including X, Y and Z directions is constructed according to the number of the nodes and a spacing distance L between each two nodes;
coordinates of each node in the three-dimensional cube network are calculated;
it is determined whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and a tube bundle radius is distributed;
it is determined whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and a tube bundle radius is distributed;
the coordinates of each node are moved based on a preset rule; and the disordered spatial structure is generated according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
In an exemplary embodiment, moving the coordinates of each node based on the preset rule includes:
the coordinates (x,y,z) of each node may be moved according to the following formula:
(x,y,z) = [(i ¨1)L rand( )%(0.51,),(j ¨1)L rand( )%(0.54(k ¨1)L rand( )%(0.5L)1 where i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
In an exemplary embodiment, establishing the pore network model according to the three-dimensional tensor data volume conforming to the disordered spatial structure and the frequency distribution of the pore throat radius of the reservoir includes:
the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir is assigned to the nodes of the disordered spatial structure sequentially to establish the pore network model.
Those skilled in the art may understand that all or part of the steps in the above method Date Recue/Date Received 2022-02-25 may be completed by instructing the relevant hardware through a program, which may be stored in computer-readable storage media, such as read-only memory, magnetic disk or optical disk, etc. Optionally, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Accordingly, the modules/units in the above embodiments may be implemented in the form of hardware or software functional modules.
The present disclosure is not limited to any specific combination of hardware and software.
There are many other embodiments of the present disclosure. Without departing from the spirit and essence of the present disclosure, those skilled in the art may make various respective changes and variations according to the present disclosure, but these respective changes and variations should be covered by the protection scope of the appended claims of the present disclosure.
Date Recue/Date Received 2022-02-25
Claims (10)
1. A modeling method of a pore network model based on a reservoir, comprising:
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length and a reservoir thickness according to the reservoir image, and calculating a three-dimensional tensor convolution kernel according to the correlation length and the reservoir thickness, wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
acquiring a reservoir image obtained by electrical imaging logging;
acquiring a correlation length and a reservoir thickness according to the reservoir image, and calculating a three-dimensional tensor convolution kernel according to the correlation length and the reservoir thickness, wherein the correlation length represents an average value of radii of preselected black masses in the reservoir image;
acquiring a T2 spectrum of a reservoir obtained by nuclear magnetic resonance logging, and acquiring a frequency distribution of a pore throat radius according to the T2 spectrum;
forming an initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius;
generating a three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir by using a convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume;
constructing a disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume; and establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure.
2. The method of claim 1, wherein, calculating the three-dimensional tensor convolution kernel according to the correlation length and the reservoir thickness comprises:
calculating the three-dimensional tensor convolution kernel E(h) according to following formula:
E(h)= exp(-2h/ Le) wherein h represents a distance from a spherical surface with ( L,,L,,Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional Date Recue/Date Received 2022-02-25 coordinate system, L L , L = , and Lc represents the correlation length.
x 2 2 z 2
calculating the three-dimensional tensor convolution kernel E(h) according to following formula:
E(h)= exp(-2h/ Le) wherein h represents a distance from a spherical surface with ( L,,L,,Lz) as a spherical center and a radius smaller than or equal to Lc to the spherical center in a three-dimensional Date Recue/Date Received 2022-02-25 coordinate system, L L , L = , and Lc represents the correlation length.
x 2 2 z 2
3. The method of claim 1, wherein, acquiring the T2 spectrum of the reservoir obtained by nuclear magnetic resonance logging, and the acquiring the frequency distribution of the pore throat radius according to the T2 spectrum comprises:
acquiring n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology, where n is a positive integer; and summing the n T2 spectra, and converting an amplitude value of the summed T2 spectrum into the frequency distribution of the pore throat radius through a preset quantitative relation.
acquiring n T2 spectra of n sub-reservoirs of the reservoir obtained by nuclear magnetic resonance logging technology, where n is a positive integer; and summing the n T2 spectra, and converting an amplitude value of the summed T2 spectrum into the frequency distribution of the pore throat radius through a preset quantitative relation.
4. The method of claim 3, wherein, the preset quantitative relation is rni=cT2m;
wherein, rm is a m-th pore throat radius, T2m is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
wherein, rm is a m-th pore throat radius, T2m is a m-th amplitude value of the T2 spectrum, c is a preset conversion coefficient and m is a positive integer.
5. The method of claim 3, wherein, forming the initial three-dimensional tensor data volume according to the frequency distribution of the pore throat radius comprises:
forming the initial three-dimensional tensor data volume by establishing a three-dimensional stable random field through a random function according to the frequency distribution of the pore throat radius;
wherein the random function is following lognormal distribution random function:
fe-(Lnx-,0212O-2 xo--NITT
wherein a mathematical expectation t.t and a standard deviation 6 are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
forming the initial three-dimensional tensor data volume by establishing a three-dimensional stable random field through a random function according to the frequency distribution of the pore throat radius;
wherein the random function is following lognormal distribution random function:
fe-(Lnx-,0212O-2 xo--NITT
wherein a mathematical expectation t.t and a standard deviation 6 are obtained by fitting the frequency distribution of the pore throat radius, and x represents a pore throat radius.
6. The method of claim 3, wherein, generating the three-dimensional tensor data volume conforming to the frequency Date Recue/Date Received 2022-02-25 distribution of the pore throat radius of the reservoir by using the convolution neural network forward propagation algorithm according to the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume comprises:
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of the point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
performing tensor point multiplication on the three-dimensional tensor convolution kernel and the initial three-dimensional tensor data volume sequentially, stacking results of the point multiplication according to an order in the initial three-dimensional tensor data volume to generate the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir.
7. The method of claim 1, wherein, constructing the disordered spatial structure of the pore network model according to the initial three-dimensional tensor data volume, comprises:
determining the number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network comprising X, Y and Z directions according to the number of nodes and a spacing distance L between each two nodes;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
determining the number of nodes of the disordered spatial structure according to data amount of the initial three-dimensional tensor data volume;
constructing a three-dimensional cube network comprising X, Y and Z directions according to the number of nodes and a spacing distance L between each two nodes;
calculating coordinates of each node in the three-dimensional cube network;
determining whether there is a tube bundle connected between each two adjacent nodes in the X direction according to a first preset probability function, and distributing a tube bundle radius;
determining whether there is a tube bundle connected between each two adjacent nodes in the Y direction according to a second preset probability function, and distributing a tube bundle radius;
moving the coordinates of each node based on a preset rule; and generating the disordered spatial structure according to the three-dimensional cube network, a result of determining whether there is the tube bundle connection, the distributed tube bundle radius and the moved coordinates of the node.
8. The method of claim 7, wherein, moving the coordinates of each node based on the preset rule comprises:
Date Recue/Date Received 2022-02-25 moving the coordinates (x,y,z) of each node according to following formula:
(x,y,z) =[(i ¨1)L rand ( )%(0.5L),(j ¨
rand ( )%(0.54(k ¨1)L rand( )%(0.5L)]
wherein, i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
Date Recue/Date Received 2022-02-25 moving the coordinates (x,y,z) of each node according to following formula:
(x,y,z) =[(i ¨1)L rand ( )%(0.5L),(j ¨
rand ( )%(0.54(k ¨1)L rand( )%(0.5L)]
wherein, i is a node number in the X direction, j is a node number in the Y
direction, k is a node number in the Z direction, i, j and k are integers greater than 0, and rand()%(0.5L) represents randomly generating of any integer within a range of 0.5L.
9. The method of claim 1, wherein, establishing the pore network model according to the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir and the disordered spatial structure comprises:
1 0 assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to nodes of the disordered spatial structure sequentially to establish the pore network model.
1 0 assigning the three-dimensional tensor data volume conforming to the frequency distribution of the pore throat radius of the reservoir to nodes of the disordered spatial structure sequentially to establish the pore network model.
10. A modeling device of a pore network model based on a reservoir, comprising: a memory and a processor;
the memory is configured to store a program for modeling the pore network model of the reservoir;
the processor is configured to read and execute the program for modeling the pore network model of the reservoir, and perform the modeling method of any one of claims 1 to 9.
Date Recue/Date Received 2022-02-25
the memory is configured to store a program for modeling the pore network model of the reservoir;
the processor is configured to read and execute the program for modeling the pore network model of the reservoir, and perform the modeling method of any one of claims 1 to 9.
Date Recue/Date Received 2022-02-25
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010681232.3 | 2020-07-15 | ||
| CN202010681232.3A CN111950193B (en) | 2020-07-15 | 2020-07-15 | Modeling method and device of pore network model based on reservoir |
| PCT/CN2020/126212 WO2022011893A1 (en) | 2020-07-15 | 2020-11-03 | Reservoir-based modeling method and device for pore network model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA3152669A1 true CA3152669A1 (en) | 2022-01-20 |
Family
ID=73340601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA3152669A Pending CA3152669A1 (en) | 2020-07-15 | 2020-11-03 | Reservoir-based modeling method and device for pore network model |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN111950193B (en) |
| CA (1) | CA3152669A1 (en) |
| WO (1) | WO2022011893A1 (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114283254B (en) * | 2021-12-31 | 2022-09-16 | 西南石油大学 | Core digital pore network model construction method based on nuclear magnetic resonance data |
| CN115290930B (en) * | 2022-08-12 | 2023-03-14 | 西南石油大学 | Digital core random construction method considering unconventional reservoir multi-component characteristics |
| CN115755610B (en) * | 2022-11-21 | 2023-09-01 | 西安石油大学 | Water injection throughput development numerical simulation system |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012006604A2 (en) * | 2010-07-08 | 2012-01-12 | Geco Technology B.V. | Methods and Devices for Transformation of Collected Data for Improved Visualization Capability |
| CN103822865B (en) * | 2014-03-20 | 2016-05-04 | 中国石油大学(华东) | A kind of high-resolution three-dimension digital cores modeling method |
| CN104375204A (en) * | 2014-11-21 | 2015-02-25 | 中国海洋石油总公司 | Method and device for analyzing anisotropism of reservoir |
| FR3034894B1 (en) * | 2015-04-09 | 2018-08-10 | Services Petroliers Schlumberger | OIL STORAGE SATURATION TANK AND PERMEABILITY MODELING |
| FR3038408B1 (en) * | 2015-06-30 | 2017-08-25 | Services Petroliers Schlumrberger | MODELING THE SATURATION AND PERMEABILITY OF PETROLEUM FIELD RESERVOIR |
| CN107728231B (en) * | 2017-10-18 | 2019-04-12 | 科为联合(青岛)能源技术研究院有限公司 | A kind of prediction nuclear magnetic resonance log lateral relaxation time T2The method of distribution |
| CN108629072B (en) * | 2018-03-12 | 2019-07-09 | 山东科技大学 | Convolutional neural networks study and prediction technique towards the distribution of earthquake oil and gas reservoir |
| CN110362842A (en) * | 2018-04-09 | 2019-10-22 | 长江大学 | Random pore network model modeling method based on various shapes pore throat |
| CN109655394B (en) * | 2018-12-21 | 2022-05-27 | 中国海洋石油集团有限公司 | Nuclear magnetic resonance T2 spectrum permeability calculation method under constraint of throat structure parameters |
| CN110009150B (en) * | 2019-04-02 | 2019-11-19 | 中国石油大学(华东) | The compact reservoir physical parameter intelligent Forecasting of data-driven |
| CN110866912B (en) * | 2019-11-15 | 2022-04-19 | 成都理工大学 | Shale streak layer heterogeneity data processing method based on imaging logging image texture |
| CN110988998A (en) * | 2019-12-24 | 2020-04-10 | 王颜 | Reservoir porosity prediction method based on Hilbert spectrum and BP neural network |
| CN111189862B (en) * | 2020-01-09 | 2021-06-11 | 中国石油大学(北京) | Method, device and system for constructing digital core |
-
2020
- 2020-07-15 CN CN202010681232.3A patent/CN111950193B/en active Active
- 2020-11-03 WO PCT/CN2020/126212 patent/WO2022011893A1/en active Application Filing
- 2020-11-03 CA CA3152669A patent/CA3152669A1/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| CN111950193A (en) | 2020-11-17 |
| CN111950193B (en) | 2023-03-14 |
| WO2022011893A1 (en) | 2022-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA3152669A1 (en) | Reservoir-based modeling method and device for pore network model | |
| WO2022011894A1 (en) | Convolutional neural network-based modeling method and device for pore network model | |
| Dowd et al. | A comparison of methods for the stochastic simulation of rock fractures | |
| EP2629123A2 (en) | Simulation model optimization | |
| US9229129B2 (en) | Reservoir geobody calculation | |
| AU2017202784B2 (en) | Gridless simulation of a fluvio-deltaic environment | |
| US10387583B2 (en) | Rotations from gradient directions | |
| US8630832B2 (en) | Estimation of lithological properties of a geological zone | |
| CN105205239A (en) | Method and device for modeling reservoir physical property parameter | |
| US11768982B1 (en) | Density abrupt interface inversion method and system based on machine learning constraints | |
| Li et al. | Refined modeling and identification of complex rock blocks and block-groups based on an enhanced DFN model | |
| CN110927793B (en) | Reservoir prediction method and system based on sequential random fuzzy simulation | |
| Krotz et al. | Variable resolution Poisson-disk sampling for meshing discrete fracture networks | |
| Rasera et al. | A conflict-free, path-level parallelization approach for sequential simulation algorithms | |
| Wu et al. | Research on fault cutting algorithm of the three-dimensional numerical manifold method | |
| Shi et al. | Data-driven multi-stage sampling strategy for a three-dimensional geological domain using weighted centroidal voronoi tessellation and IC-XGBoost3D | |
| Fu et al. | Spatial topology identification of three-dimensional complex block system of rock masses | |
| US11300706B2 (en) | Designing a geological simulation grid | |
| CN107356971A (en) | Geological data rule method and device | |
| GB2584447A (en) | Apparatus method and computer-program product for processing geological data | |
| GB2584449A (en) | Apparatus method and computer-program product for calculating a measurable geological metric | |
| Dantas et al. | A constrained version for the stereology inverse problem: Honoring power law and persistences of the fracture traces exposed on arbitrary surfaces | |
| Verscheure et al. | History matching of a realistic stochastic fault model using fractal geometry and the gradual deformation method | |
| Mohajerani et al. | An efficient computational model for simulating stress-dependent flow in three-dimensional discrete fracture networks | |
| CN118015219B (en) | Geological model generation method, device and equipment based on qualitative kriging interpolation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request |
Effective date: 20220428 |
|
| EEER | Examination request |
Effective date: 20220428 |
|
| EEER | Examination request |
Effective date: 20220428 |
|
| EEER | Examination request |
Effective date: 20220428 |
|
| EEER | Examination request |
Effective date: 20220428 |
|
| EEER | Examination request |
Effective date: 20220428 |