CN111815773A - Three-dimensional complex geological model label manufacturing method suitable for machine learning algorithm - Google Patents
Three-dimensional complex geological model label manufacturing method suitable for machine learning algorithm Download PDFInfo
- Publication number
- CN111815773A CN111815773A CN202010847625.7A CN202010847625A CN111815773A CN 111815773 A CN111815773 A CN 111815773A CN 202010847625 A CN202010847625 A CN 202010847625A CN 111815773 A CN111815773 A CN 111815773A
- Authority
- CN
- China
- Prior art keywords
- model
- dimensional
- machine learning
- learning algorithm
- fault
- 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.)
- Granted
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 29
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 26
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 7
- 238000006073 displacement reaction Methods 0.000 claims description 27
- 238000000034 method Methods 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 22
- 230000015572 biosynthetic process Effects 0.000 claims description 15
- 238000005452 bending Methods 0.000 claims description 14
- 230000001629 suppression Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 238000002372 labelling Methods 0.000 claims 5
- 238000012545 processing Methods 0.000 abstract description 7
- 230000008021 deposition Effects 0.000 abstract description 4
- 238000011161 development Methods 0.000 description 13
- 238000005755 formation reaction Methods 0.000 description 12
- 230000004044 response Effects 0.000 description 10
- 235000019994 cava Nutrition 0.000 description 8
- 238000010276 construction Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000005553 drilling Methods 0.000 description 4
- 230000005764 inhibitory process Effects 0.000 description 4
- 230000001788 irregular Effects 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 239000004215 Carbon black (E152) Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000012530 fluid Substances 0.000 description 2
- 229930195733 hydrocarbon Natural products 0.000 description 2
- 150000002430 hydrocarbons Chemical class 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000945 filler Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V20/00—Geomodelling in general
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2219/00—Indexing scheme for manipulating 3D models or images for computer graphics
- G06T2219/004—Annotating, labelling
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Remote Sensing (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Graphics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention discloses a three-dimensional complex geological model label manufacturing method suitable for a machine learning algorithm, which is applied to the field of seismic data processing and aims at solving the problem that a three-dimensional geological model constructed in the prior art does not completely accord with stratum deposition and geomechanical rules.
Description
Technical Field
The invention belongs to the field of seismic data processing, and particularly relates to a three-dimensional geological model construction technology.
Background
The depth of exploration and development of oil and gas reservoirs is deeper and deeper, and the complexity of geological structures is higher and higher. The conventional oil and gas exploration and development method needs to go through a plurality of links, so that the research working efficiency is low, the exploration period is long, the description precision of an oil and gas reservoir is low, the cost is high, and the conventional oil and gas exploration and development method has become more and more compelling. Aiming at the difficult problems faced by the conventional oil and gas exploration and development, the machine learning algorithm provides a new thought and means for the exploration and development of the oil and gas reservoir with a complex geological structure. The application of machine learning to oil and gas exploration and development is becoming more and more popular among geophysicists and geologists in recent years. Whether the application success of the machine learning algorithm is closely related to the richness of the label data set, and the development data sets such as MNIST (LeCun et al, 2018) and ImageNet (Deng et al, 2009) promote the rapid progress of machine learning, so that the machine learning is exponentially increased in the last 10 years. Ideally, experts in the earth's world also need such geoscience reference data sets.
In the petroleum industry, seismic exploration is the primary means used for the exploration and development of complex geological formations of hydrocarbon reservoirs. The basis for extracting relevant parameters of a hydrocarbon reservoir of a complex geological structure from seismic data acquired by seismic exploration by utilizing machine learning is the need of defining seismic response characteristics of different geological structures. Drilling and formation parameter measurements are one of the most effective ways to establish relationships between different types of geological formations and seismic response characteristics. Under the influence of exploration and development cost, a data set obtained by drilling and formation parameter measurement far cannot meet the requirement of solving the exploration and development machine learning algorithm of the oil and gas reservoir with the complex geological structure. For processing images by applying a machine learning algorithm, predecessors often perform data sample expansion on labeled images by methods such as scaling, rotation, noise disturbance and the like. For seismic data, the method has the characteristics of large uncertainty factor, extremely complex internal structure, huge data volume, little known information and the like, and shows unique characteristics on the seismic data aiming at specific geological structure, stratum deposition and lithofacies, so that the data sample expansion method in the field of image processing is not suitable. Constructing seismic response datasets of different types of geological formations suitable for machine learning algorithms has been a difficult problem for geophysicists. Particularly for the dissolved-fluid reservoirs researched in Tarim basin in recent years, the reservoir space has the characteristics of uneven distribution, diversified forms, large longitudinal depth, extremely strong internal heterogeneity, and very complex reservoir space distribution, and the acquisition of the seismic response sample data set corresponding to the dissolved-fluid reservoir is more difficult and expensive. The construction of the seismic response data sets of different types of geological structures comprises construction of geological models conforming to actual conditions and seismic response numerical simulation. Seismic response modeling algorithms are already well established. Therefore, the core of the seismic response data set for constructing the complex geological structure model is how to construct the complex geological structure model which accords with the stratum sedimentary rule and geomechanics.
Aiming at the difficult problem of fault space distribution identification based on seismic data and by a machine learning algorithm, Wuxin et al (2019) train by establishing a large amount of three-dimensional fault synthesis data as data samples, consider fault identification as a classification problem, apply the fault identification to actual seismic data, and show that the fault identification precision and efficiency are greatly improved. Wuxin et al (2020) also propose to construct a real geological structure model for training convolutional neural network training, thereby realizing seismic structure interpretation, and the processing of actual processing shows that a good effect is obtained. From the point of view of mathematical and statistical theory, the method can build countless geological structure models. However, from the perspective of sedimentology and geomechanics analysis, the three-dimensional geological model constructed by the method does not completely conform to the laws of stratum sedimentary and geomechanics.
Disclosure of Invention
In order to solve the technical problems, the invention provides a three-dimensional complex geological model label manufacturing method suitable for a machine learning algorithm, which is based on the theories of stratum sedimentology, tectonics, geological tectonics and the like, and takes the factors of stratum sedimentary law, stratum stress condition generated by faults, connection among karst caves, holes, cracks and faults and the like into consideration.
The technical scheme adopted by the invention is as follows: the three-dimensional complex geological model label manufacturing method suitable for the machine learning algorithm comprises the following steps:
s1, generating a three-dimensional horizontal lamellar model according to the simulated formation parameters;
s2, calculating displacement, and adding the displacement into the horizontal lamellar model to generate a three-dimensional fault model;
s3, performing stratum bending and stratum inclination on the basis of the three-dimensional fault model obtained in the step S2 to obtain a three-dimensional bending layered model;
and S4, adding the three-dimensional random hole-and-seam model into the three-dimensional bending lamellar model generated in the step S3 to generate a final three-dimensional geological model.
Further, the three-dimensional tomographic model of step S2 expresses a micro:
Vf(X,Y,Z)＝v(X,Y+Dy,Z+Dz)
wherein ,DyIs a y-direction displacement; dzIs the z-direction displacement.
Further, DyAnd DzThe general calculation of (a) is:
D＝α·(dmaxd)+η
wherein α is a reverse suppression coefficient, d is represented by (X)0,Y0,Z0) A three-dimensional Gaussian function centered as an inverse suppression variable, η is a constant, dmaxIs the maximum displacement.
Further, step S3 is specifically:
s31, adding a vertical displacement variable S on the basis of the three-dimensional fault model1(X, Y, Z) to obtain a curved lamellar model Vf(X,Y,Z+S1)；
S32, adding a formation dip variable S2(X, Y, Z) to produce a tilted curved lamellar model Vf(X,Y,Z+S1+S2)。
Further, S1The expression (X, Y, Z) is:
wherein ,
as a function of a linear scale, ZmaxIs the maximum value of the Z coordinate; c. Ck,dk,σkParameters of a two-dimensional Gaussian function; bkAre coefficients of a gaussian mixture function.
Further, S2The expression (X, Y, Z) is:
S2(X,Y,Z)＝aX+bY+c
wherein a and b are random numbers; c ═ aX0-bY0；
Further, step S3 includes performing cubic spline interpolation processing on the oblique curved layer model.
Further, the three-dimensional random hole and seam model in step S4 is generated by using a random medium theory.
The invention has the beneficial effects that: the invention comprehensively considers geological factors from the aspects of sedimentology, tectonics and geomechanics, simulates three-dimensional seismic data tags conforming to stratum sedimentary, geological stress conditions and geological rules, realizes the construction of a large number of three-dimensional geological models conforming to actual geological conditions, can enrich the existing seismic tag data, and combines a seismic wave field response simulation technology to provide conditions for extracting relevant characteristics of oil and gas reservoirs with complex geological structures from seismic data by utilizing a machine learning algorithm, reduce exploration cost and exploration period and improve the drilling success rate in the oil and gas exploration and development process.
Drawings
FIG. 1 is a flow chart of a protocol of the present invention;
FIG. 2 is a cross-sectional level model diagram;
FIG. 3 is a three-dimensional model building step;
FIG. 4 is a prior art fault model;
FIG. 5 is a fault model of the present invention;
FIG. 6 is a geological model of the present invention;
FIG. 7 is a forward model of the present invention.
Detailed Description
To facilitate understanding of the technical content of the present invention by those skilled in the art, the geologic modeling technique of the prior art is described below:
the existing geological model building method comprises the following steps:
the technology is mainly based on the solid geometry and surface fitting theory, the fault surface is fitted into a surface with obvious characteristics, and the three-dimensional geological model is constructed by stretching and extruding the surface. The realization of the technology comprises the following steps:
step 1: establishing a three-dimensional horizontal layered geological model r (X, Y, Z) with the value of a random number between [ -1,1 ]; and r (X, Y, Z) is the seismic longitudinal wave reflection coefficient of the stratum model.
Step 2: adding a geologic vertical displacement variable S1(X, Y, Z), which can be represented as:
wherein ,
a linear scale function is adopted, so that the vertical displacement is weakened from top to bottom; e is a natural base number; c. Ck,dk,σkIs a linear combination of two-dimensional Gaussian functions in a formula
The parameter (1) of (1); bkCoefficients that are linear combinations of gaussian functions; k is a subscript of a kth two-dimensional Gaussian function parameter; and N is the number of two-dimensional Gaussian functions.
And step 3: adding formation dip variables S2(X, Y, Z), which can be represented as:
S2(X,Y,Z)＝aX+bY+c (2)
wherein a and b are [ -0.25,0.25 [ ]]A random number in between; c ═ aX0-bY0So that the center lane (X)0,Y0Z) no offset; finally, a curved lamellar model r (X, Y, Z + S) is generated1+S2)；
And 4, step 4: in order to make the model more consistent with the actual stratum, faults are added into the model, wherein the fault generation has the following procedures:
determining the dip angle theta (theta belongs to 0 degree and 90 degrees) of a fault plane, and moving towards
Converting global coordinates to local coordinates (X, y, z) ═ R (X-X)0'；Y-Y0'；Z-Z0') in the local coordinate system with (X)0',Y0',Z0') is the origin of coordinates, runs in the x-direction, the tilt direction is the y-direction, the dip direction is the z-direction, where:
secondly, the fault surface is more complicated, the smooth fault surface is bent, points around the fault surface are randomly selected under a local coordinate system, and Green spline interpolation is carried out to obtain a new fault surface z (f (x, y));
defining a displacement function D in y directiony(x,y,z)：
(f(x,y)≤z≤γ+f(x,y)),Dy(x,y,z)＝λ·d(x,y；z＝0)·α(x,y,z) (4)
(f(x,y)-γ≤z≤f(x,y)),Dy(x,y,z)＝(λ-1)·d(x,y；z＝0)·α(x,y,z) (5)
Wherein gamma is a reverse inhibition radius, and lambda belongs to (0, 1); the d (x, y; z ═ 0) function spreads the displacement from the fault center point to the periphery, defined here as:
wherein ,
α (x, y, z) is such that the displacement does not diffuse too far in the model, which is defined as
Gamma is the reverse inhibition radius;
defining a displacement function D in the z directionz(x,y,z)：
Dz(x,y,z)＝f(x,y+Dy(x,y,z))-f(x,y) (7)
Finally, the fault model (x, y + D) is generatedy,z+Dz) And then converted from the local coordinate system (X, Y, Z) to the global coordinate system (X, Y, Z)
The invention will be described in detail below with reference to the accompanying drawings 1-7:
the success of the application of the machine learning algorithm depends on the completeness of the tag data set. A large number of label data sets with complete characteristics can guarantee the effectiveness of the machine learning algorithm, overfitting is prevented, and the robustness of the algorithm is improved. Under the influence of exploration and development cost, seismic response data sets corresponding to different geological structure models calibrated by drilling and formation parameter measurement far cannot meet the requirement for solving the problem of machine learning algorithm training of exploration and development of complex geological structure oil and gas reservoirs. Although countless geological structure models can be built by the existing geological model building technology, the three-dimensional geological model built by the prior art does not completely accord with the laws of stratum deposition and geomechanics from the perspective of theories such as sedimentology, geomechanics and the like.
Therefore, the method is based on the theories of stratum sedimentology, tectonic geology, geological tectonic mechanics and the like, takes the stratum sedimentary law, the stratum stress condition generated by faults, the relation between karst caves, holes, cracks and the faults and the like into consideration, is suitable for the three-dimensional complex geological model label manufacturing method of the machine learning algorithm, and realizes the construction of a large number of three-dimensional geological models which accord with the actual geological conditions (such as stratum bending, faults, cracks, irregular geometric shapes of the karst caves, filler properties and the like).
The establishment of a large number of three-dimensional geological models can enrich the characteristics of seismic data samples, and provide more complete information for machine learning to extract the relevant characteristics of complex geological structure oil and gas reservoirs from seismic data. The geological model is quite complex to build, and the method can approach the actual seismic data to a certain extent. As shown in fig. 1, the process of building a three-dimensional geological model is:
a1, establishing a three-dimensional horizontal laminar velocity model v (X, Y, Z), wherein the value of the model v is a random value of a simulated corresponding lithofacies;
a2, as shown in FIG. 2, defines an angle θ (θ ∈ (0 °,90 °)) and an angle
And a center point (X)0,Y0,Z0) A fault plane f (X, Y, Z) is determined. Wherein
For the trend of the fault, according to the geometric knowledge, the dip angle of the fault is
Then, a fault is generated in the space:
Dy(X,Y,Z)＝Dsin(θ) (8)
Dz(X,Y,Z)＝Dcos(θ) (9)
wherein ,Dy(X, Y, Z) is displacement in the Y direction; dz(X, Y, Z) is a Z-direction displacement; the displacement variable D is calculated as: d ═ α · (D)maxd)+η：
Wherein α is a reverse suppression coefficient; d is (X)0,Y0,Z0) A three-dimensional gaussian function as a center, i.e., an inverse suppression variable; λ ∈ (0, 1); n (n)>1) The greater n is the power of the equation, the more curved the resulting fault and the range of curvatureThe smaller the circumference, in most cases the degree of bending and extent of the fault is also relatively weak; gamma is the reverse inhibition radius; eta is a constant, | eta #>dmax，dmaxIs the maximum displacement.
Therefore, the fault model Vf(X, Y, Z) is generated as follows:
Vf(X,Y,Z)＝v(X,Y+Dy,Z+Dz) (12)
in the formula ,DyIs a y-direction displacement; dzIs a z-direction displacement;
a3, in order to make the model more approximate to the real data, the fault model VfCarrying out stratum bending, stratum inclination and adding a hole model in the model on the basis of (X, Y, Z):
first, add a vertical displacement variable S1(X, Y, Z), in which case a curved layer model V is obtainedf(X, Y, Z + S1) that enables the previously generated flat fault plane to become a curved surface, reducing the prior art green spline interpolation step, making the model building step more simplified, which can be expressed as:
wherein ,
the vertical displacement is weakened from top to bottom by a linear scale function, and the larger beta is, the higher the formation bending degree is; c. Ck,dk,σkIs a linear combination of two-dimensional Gaussian functions in a formula
The parameter (1) of (1); bkCoefficients that are linear combinations of gaussian functions;
next, a formation dip variable S is added2(X, Y, Z), which can be represented as:
S2(X,Y,Z)＝aX+bY+c (14)
wherein a and b are [ -0.25,0.25 [ ]]A random number in between; c ═ aX0-bY0So that the center lane (X)0,Y0Z) no offset, (X)0,Y0,Z0) As a central channel (X)0,Y0A point on Z);
then producing a tilted curved lamellar model Vf(X,Y,Z+S1+S2) In order to make the model line smoother, cubic spline interpolation is carried out on the model;
a4, combining the connection between the hole and the fault in the actual situation according to the position of the fault, generating a three-dimensional random hole and hole model V by the random medium theoryrand(X, Y, Z) is added to the model Vf(X,Y,Z+S1+S2) And (5) generating a final three-dimensional geological model.
As shown in fig. 3, a reverse fault is simulated, the fault generates weak dragging under the stress condition, and the dragging accords with the actual geological law. Then, the cracks can be seen in the three-dimensional random karst caves, holes and cracks generated according to the random medium theory, the cracks penetrate through the karst caves and are randomly distributed in the whole three-dimensional space, and the karst caves, the holes and the cracks can be seen in the three-dimensional geological model and are distributed around the faults according to the geological relationship between the karst caves, the holes and the cracks and the faults, so that the constructed geological model is more complex and has the characteristics according with the stratum deposition and the geomechanical law.
The technical effects of the invention are illustrated below with reference to specific examples:
a fault model with three layers of stratums is generated by the existing method and the method, then the random medium theory is used for filling karst cave, holes and cracks in the fault model, and finally a three-dimensional geological model and a three-dimensional convolution model are obtained.
As shown in fig. 4, the prior art simulates a normal fault, and when the plate descends on the normal fault under consideration of stress factors, the ground dragging around the fault should be bent upwards, so that it can be seen from fig. 4 that the actual geological law is violated by the slight dragging of the fault. The inclined bending lamellar model is shown in figure 5, the problem in figure 4 is corrected in figure 5, the corrected fault micro-dragging can be seen to be more in line with geological rules, and factors such as connection among karst caves, holes, cracks and faults are considered. As shown in fig. 6, the solution cavity, the hole and the fracture are added on the basis of fig. 5, and in fig. 6, it can be seen that the fracture passes through the solution cavity, and the generated random solution cavity, the hole and the fracture are distributed around the fault, so as to simulate the more real formation condition,
FIG. 7 is a three-dimensional convolution model obtained by adding random noise after the convolution of a three-dimensional geological model and 15HZ rake wavelets, wherein a fault plane is an irregular curved surface after the stratum is bent, so that the model is more in line with the actual geological law.
The invention has the following advantages:
(1) constructing a displacement function for generating a fault, so that the generation of the fault is more consistent with a geological stress rule;
(2) constructing a reverse inhibition function with a controllable fault pulling range, so that the pulling of the fault under the action of stress is controlled within a certain range;
(3) irregular fault surfaces can be obtained by carrying out stratum bending after the fault is generated, so that the fault surfaces are more consistent with the actual geological condition, and the Green spline interpolation step in the prior art is omitted, so that the model construction is simpler and more effective;
(4) the vertical bending variable of the stratum is changed, so that the bending degree of the stratum is more convenient to control;
(5) the random medium theory is introduced to be combined with the fault model, so that the representation information of the constructed geological model is richer;
it will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Various modifications and alterations to this invention will become apparent to those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims (8)
1. The three-dimensional complex geological model label manufacturing method suitable for the machine learning algorithm is characterized by comprising the following steps of:
s1, generating a three-dimensional horizontal lamellar model according to the simulated formation parameters;
s2, calculating displacement, and adding the displacement into the horizontal lamellar model to generate a three-dimensional fault model;
s3, performing stratum bending and stratum inclination on the basis of the three-dimensional fault model obtained in the step S2 to obtain a three-dimensional bending layered model;
and S4, adding the three-dimensional random hole-and-seam model into the three-dimensional bending lamellar model generated in the step S3 to generate a final three-dimensional geological model.
2. The method for labeling the three-dimensional complex geological model suitable for the machine learning algorithm according to claim 1, wherein the three-dimensional fault model expression in step S2 is as follows:
Vf(X,Y,Z)＝v(X,Y+Dy,Z+Dz)
wherein ,DyIs a y-direction displacement; dzIs the z-direction displacement.
3. The method for labeling the three-dimensional complex geological model applied to the machine learning algorithm according to claim 2, characterized in that DyAnd DzThe general calculation of (a) is:
D＝α·(dmaxd)+η
wherein α is a reverse suppression coefficient, d is represented by (X)0,Y0,Z0) A three-dimensional Gaussian function centered as an inverse suppression variable, η is a constant, dmaxIs the maximum displacement.
4. The method for making the label of the three-dimensional complex geological model suitable for the machine learning algorithm according to claim 1, wherein the step S3 is specifically as follows:
s31, adding a vertical displacement variable S on the basis of the three-dimensional fault model1(X, Y, Z) to obtain a curved lamellar model Vf(X,Y,Z+S1)；
S32, adding a formation dip variable S2(X, Y, Z) to produce a tilted curved lamellar model Vf(X,Y,Z+S1+S2)。
5. The method for labeling the three-dimensional complex geological model applied to the machine learning algorithm according to claim 4, characterized in that S1The expression (X, Y, Z) is:
wherein ,
as a function of a linear scale, ZmaxIs the maximum value of the Z coordinate; c. Ck,dk,σkParameters of a two-dimensional Gaussian function; bkAre coefficients of a gaussian mixture function.
6. The method for labeling the three-dimensional complex geological model applied to the machine learning algorithm according to claim 4, characterized in that S2The expression (X, Y, Z) is:
S2(X,Y,Z)＝aX+bY+c
wherein a and b are random numbers; c ═ aX0-bY0。
7. The method for labeling a three-dimensional complex geological model suitable for use in a machine learning algorithm according to claim 1, wherein step S3 further comprises performing cubic spline interpolation on the tilted curved layered model.
8. The method for making the label of the three-dimensional complex geological model suitable for the machine learning algorithm according to claim 1, wherein the three-dimensional random hole and seam hole model of step S4 is generated by using a random medium theory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847625.7A CN111815773B (en) | 2020-08-21 | 2020-08-21 | Three-dimensional complex geologic model label manufacturing method suitable for machine learning algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010847625.7A CN111815773B (en) | 2020-08-21 | 2020-08-21 | Three-dimensional complex geologic model label manufacturing method suitable for machine learning algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111815773A true CN111815773A (en) | 2020-10-23 |
| CN111815773B CN111815773B (en) | 2023-08-25 |
Family
ID=72859531
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010847625.7A Active CN111815773B (en) | 2020-08-21 | 2020-08-21 | Three-dimensional complex geologic model label manufacturing method suitable for machine learning algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111815773B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112766321A (en) * | 2020-12-31 | 2021-05-07 | 中国地质调查局成都地质调查中心 | Geological feature detection and identification method and system based on deep learning |
| CN113253342A (en) * | 2021-05-11 | 2021-08-13 | 电子科技大学 | Method for constructing complex three-dimensional seismic model label by combining logging data |
| CN114355449A (en) * | 2022-01-05 | 2022-04-15 | 电子科技大学 | Vector median constrained structure-oriented three-dimensional seismic image enhancement method |
| CN114966840A (en) * | 2022-05-24 | 2022-08-30 | 哈尔滨工业大学 | Three-dimensional seismic wave field model building method based on general solution neural network |
| CN115469361A (en) * | 2022-09-28 | 2022-12-13 | 北京中恒利华石油技术研究所 | Three-dimensional geological modeling method for clastic rock stratum |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11898435B2 (en) * | 2020-09-25 | 2024-02-13 | Halliburton Energy Services, Inc. | Correcting borehole images using machine-learning models |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886641A (en) * | 2014-04-18 | 2014-06-25 | 重庆市勘测院 | Three-dimensional geological model constructing and integrating method for mountain city area |
| CN109388817A (en) * | 2017-08-04 | 2019-02-26 | 中国石油化工股份有限公司 | A kind of Reservoir Fracture three-dimensional modeling method |
| CN110060342A (en) * | 2019-02-22 | 2019-07-26 | 中铁工程设计咨询集团有限公司 | A kind of three-dimension curved surface approximating method |
| CN110705168A (en) * | 2019-10-11 | 2020-01-17 | 内蒙古科技大学 | Simulation method of structural stress field |
| US20200158898A1 (en) * | 2017-05-23 | 2020-05-21 | Schlumberger Technology Corporation | Analogue Facilitated Seismic Data Interpretation System |
-
2020
- 2020-08-21 CN CN202010847625.7A patent/CN111815773B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886641A (en) * | 2014-04-18 | 2014-06-25 | 重庆市勘测院 | Three-dimensional geological model constructing and integrating method for mountain city area |
| US20200158898A1 (en) * | 2017-05-23 | 2020-05-21 | Schlumberger Technology Corporation | Analogue Facilitated Seismic Data Interpretation System |
| CN109388817A (en) * | 2017-08-04 | 2019-02-26 | 中国石油化工股份有限公司 | A kind of Reservoir Fracture three-dimensional modeling method |
| CN110060342A (en) * | 2019-02-22 | 2019-07-26 | 中铁工程设计咨询集团有限公司 | A kind of three-dimension curved surface approximating method |
| CN110705168A (en) * | 2019-10-11 | 2020-01-17 | 内蒙古科技大学 | Simulation method of structural stress field |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112766321A (en) * | 2020-12-31 | 2021-05-07 | 中国地质调查局成都地质调查中心 | Geological feature detection and identification method and system based on deep learning |
| CN112766321B (en) * | 2020-12-31 | 2024-05-17 | 中国地质调查局成都地质调查中心 | Geological feature detection and recognition method and system based on deep learning |
| CN113253342A (en) * | 2021-05-11 | 2021-08-13 | 电子科技大学 | Method for constructing complex three-dimensional seismic model label by combining logging data |
| CN114355449A (en) * | 2022-01-05 | 2022-04-15 | 电子科技大学 | Vector median constrained structure-oriented three-dimensional seismic image enhancement method |
| CN114966840A (en) * | 2022-05-24 | 2022-08-30 | 哈尔滨工业大学 | Three-dimensional seismic wave field model building method based on general solution neural network |
| CN114966840B (en) * | 2022-05-24 | 2023-04-07 | 哈尔滨工业大学 | Model building method of three-dimensional seismic wave field based on general solution neural network |
| CN115469361A (en) * | 2022-09-28 | 2022-12-13 | 北京中恒利华石油技术研究所 | Three-dimensional geological modeling method for clastic rock stratum |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111815773B (en) | 2023-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111815773B (en) | Three-dimensional complex geologic model label manufacturing method suitable for machine learning algorithm | |
| US8838425B2 (en) | Generating facies probablity cubes | |
| US10838108B2 (en) | Method and system for interpolating discontinuous functions in a subsurface model | |
| US10466388B2 (en) | System and method for editing geological models by switching between volume-based models and surface-based structural models augmented with stratigraphic fiber bundles | |
| EP3631528A1 (en) | Constructing structural models of the subsurface | |
| US20130246031A1 (en) | Constructing Geologic Models From Geologic Concepts | |
| US20180217283A1 (en) | Assignment of Systems Tracts | |
| CN111612906B (en) | Method and system for generating three-dimensional geological model and computer storage medium | |
| CN109884710B (en) | Micro-logging tomography method aiming at excitation well depth design | |
| US10387583B2 (en) | Rotations from gradient directions | |
| CN107894618A (en) | A kind of full waveform inversion gradient preprocess method based on model smoothing algorithm | |
| CN112394404A (en) | Progressive reservoir fine characterization method | |
| CN115308799B (en) | Method and system for identifying free gas structure of seismic imaging | |
| CN115220101A (en) | Modeling method for deep clastic rock sand body sediment structure | |
| CN109143337A (en) | A kind of lithologic sand body reservoir modeling methodologies | |
| CN114755721A (en) | Unet-based river channel sand body top surface prediction method | |
| CN113253342A (en) | Method for constructing complex three-dimensional seismic model label by combining logging data | |
| CN111751886A (en) | Shale gas reservoir fracture modeling method based on microseism monitoring data | |
| CN115469361B (en) | Clastic rock stratum three-dimensional geological modeling method | |
| CN115292771B (en) | Pseudo 2.5D simulation method for resistivity logging while drilling response | |
| CN115963569A (en) | Method and device for establishing phase-control initial model in phase-control inversion | |
| CN114022631A (en) | Three-dimensional geological model automatic reconstruction method based on depth convolution generation countermeasure network | |
| CN117422840A (en) | Geological modeling method based on deep learning | |
| CN116935001A (en) | Depleted hydrocarbon reservoir CO 2 Geological modeling method for geological storage full stratum | |
| Aydin | A Bayesian Framework for Quantifying Fault Network Uncertainty Using a Marked Point Process Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |