CN111209674A - River channel sand modeling method and device and readable storage medium - Google Patents
River channel sand modeling method and device and readable storage medium Download PDFInfo
- Publication number
- CN111209674A CN111209674A CN202010022741.5A CN202010022741A CN111209674A CN 111209674 A CN111209674 A CN 111209674A CN 202010022741 A CN202010022741 A CN 202010022741A CN 111209674 A CN111209674 A CN 111209674A
- Authority
- CN
- China
- Prior art keywords
- network
- river channel
- sand body
- condition data
- channel sand
- 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
Images
Classifications
-
- 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
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Abstract
The invention discloses a river sand modeling method, equipment and a readable storage medium. The river sand modeling method comprises the following steps: acquiring basic parameters of a reservoir to be modeled; generating a plurality of first river channel sand body models and training condition data corresponding to the first river channel sand body models according to the acquired basic parameters; establishing a generating network; establishing a discrimination network; training the discrimination network and the generation network to obtain a generator; and inputting the well data into a generator to obtain a sand body model of the reservoir to be modeled. The invention has the beneficial effects that: a large number of first river channel sand body models and corresponding training condition data are generated through a target-based simulation method, then a generation network and a judgment network are trained through the first river channel sand body models and the training condition data, the generation network is trained in a large number to obtain a generator capable of generating a real river channel sand body model, and well data of a research area are input into the generator to obtain a sand body model of a reservoir to be modeled in the research area.
Description
Technical Field
The invention relates to the technical field of reservoir modeling, in particular to a river sand modeling method, equipment and a readable storage medium.
Background
At present, in the modeling research of river sand bodies, scholars at home and abroad propose a plurality of modeling methods. The traditional two-point statistics can realize the prediction of any reservoir, plays a great role in establishing a three-dimensional geological model of a continental facies sedimentary reservoir, and powerfully pushes the development of reservoir description to the direction of fineness and quantification. However, as research progresses, more and more scholars recognize that the two-point variation function can only represent two-point correlation, and it is difficult to depict the complex morphology of the reservoir (such as a curved river channel) and to depict different reservoir space configurations. Reservoir morphology can be well described based on the target method, and reservoir cause relation is revealed. It is based primarily on simulations based on predetermined target body morphology, such as direction, amplitude, width, and thickness of the river, but uses multiple iterations in the simulation process to match well data based on the target method, and may not generally satisfy all well data. The multipoint geostatistics modeling adopts the training images as main input parameters, can better respect well data than a target-based method, but has higher requirements on the training images, if the training images are required to have stationarity, the actual situation is less, and the phenomenon that the simulated river channel is interrupted is caused.
Disclosure of Invention
In view of the above, there is a need to provide a reservoir modeling method that can overcome the defect that the two-point variation function simulation method cannot reflect the real reservoir morphology, and can also overcome the problem that it is difficult to match well data based on the target simulation method.
The invention provides a river sand modeling method, which comprises the following steps:
s100, obtaining basic parameters of a reservoir to be modeled, wherein the basic parameters at least comprise river channel average width, river channel average curvature and river channel trend;
s200, generating a plurality of first river channel sand body models according to the acquired basic parameters, and generating a plurality of training condition data corresponding to the first river channel sand body models according to the first river channel sand body models;
s300, establishing a generating network according to each training condition data;
s400, establishing a discrimination network according to each training condition data and the corresponding first river channel sand body model;
s500, training the judging network and the generating network to obtain a generator;
s600, well data of the reservoir to be modeled are obtained, and the well data are input into a generator to obtain a sand body model of the reservoir to be modeled.
The invention also provides a device for modeling the river channel sand, which comprises a processor and a memory; the memory has stored thereon a computer readable program executable by the processor; the processor implements the steps of the river sand modeling method provided by the invention when executing the computer readable program.
The invention also provides a computer readable storage medium storing one or more programs, which are executable by one or more processors to implement the steps in the river sand modeling method provided by the invention.
Compared with the prior art, the technical scheme provided by the invention has the beneficial effects that: a large number of first river channel sand body models and corresponding training condition data are generated through a target-based simulation method, then a generation network and a judgment network are trained through the first river channel sand body models and the training condition data, the generation network is trained in a large number to obtain a generator capable of generating a vivid river channel sand body model, and well data of a research area are input into the generator to obtain a sand body model of a reservoir to be modeled in the research area.
Drawings
Fig. 1 is a schematic flow chart of an embodiment of a river sand modeling method provided by the present invention;
FIG. 2 is a first channel sand model of several reservoir regions of interest randomly generated by Petrel software;
fig. 3 is corresponding training condition data randomly generated by Petrel software for each first channel sand body model in fig. 2;
FIG. 4 is a schematic flow chart of step S300 in FIG. 1;
FIG. 5 is a schematic flow chart of step S400 in FIG. 1;
FIG. 6 is a schematic diagram illustrating a process of inputting training condition data and a corresponding first channel sand model into a judgment network;
fig. 7 is a schematic diagram of a process of inputting the second channel sand body model generated by the generation network into the judgment network for authenticity judgment;
FIG. 8 is a schematic flow chart of step S500 in FIG. 1;
FIG. 9 is well data for a study area;
FIG. 10 is a sand model of the reservoir to be modeled for the region of interest obtained by inputting the well data of FIG. 9 into the generator.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
This example uses the West lake sunken ancient near system hong Kong group as the study object. The hardware environment used in this experiment was an Nvidia GTX 1660 graphics card with 1408 stream processors (for machine learning computing), the software platform was pycharm, and the configured virtual environment was tenasorflow-gpu 1.11.0. The model adopts a PatchGAN structure in the discrimination process, namely, a discrimination layer of N x 1 is obtained through a plurality of convolution layers, wherein each element represents a corresponding authenticity judgment result, and the whole input authenticity judgment result is the average value of the N x N elements.
Referring to fig. 1, the present invention provides a river sand modeling method, including the following steps:
s100, obtaining basic parameters of a reservoir to be modeled, wherein the basic parameters at least comprise river channel average width, river channel average curvature and river channel trend; in this embodiment, according to the research results of the predecessors, the average width of the river of the reservoir is 500m, the average curvature of the river is 1.2, and the trend of the river is approximately east-west.
S200, generating a plurality of first river channel sand body models according to the acquired basic parameters, and generating a plurality of training condition data corresponding to the first river channel sand body models according to the first river channel sand body models; in this embodiment, 100 channel sand body models a are randomly generated by a Petrel software based target simulation method (see fig. 2, only 16 first channel sand body models are shown in fig. 2), and then a corresponding training condition data y is randomly generated for each first channel sand body model a by simulation software (see fig. 3).
S300, establishing a generating network according to each training condition data y; the method for establishing the generating network comprises the following steps (please refer to fig. 4):
s310, establishing a plurality of convolutional layers, coding each training condition data y, inputting the coded data into the convolutional layers for feature extraction, and obtaining a plurality of feature matrixes through convolution operation;
s320, establishing deconvolution layers with the same number as the convolution layers, connecting the convolution layers with the same dimensionality in the deconvolution layers, arranging the characteristic matrixes extracted in the step S310 into one-dimensional vectors, inputting the one-dimensional vectors into the deconvolution layers, and generating corresponding river sand model matrixes through deconvolution operation;
s330, setting a correction function for generating network parameters in the training process:
wherein theta is a parameter after network correctiongIn order to generate the parameters before the network modification,
is a loss gradient with respect to the parameter, and
wherein n is the total training times, D is the discrimination network, G is the generation network, z is the random vector, and i is the training time variable.
S400, establishing a discrimination network D according to each training condition data y and the corresponding first river channel sand body model A; the method for establishing the discrimination network D includes the following steps (see fig. 5):
s410, establishing a plurality of convolutional layers, encoding each training condition data y and the corresponding first river channel sand body model A in pairs, inputting the encoded training condition data y and the corresponding first river channel sand body model A into the convolutional layers for feature extraction, and obtaining a plurality of feature matrixes through convolution operation;
s420, comparing the training condition data y with the feature matrix of the corresponding first channel sand body model a, and respectively calculating the true probability of each first channel sand body model a (see fig. 6);
s430, inputting the sand body G (y) model generated by the generation network G into the discrimination network D, and establishing a first loss function: v (G, D) ═ Ex,y[logD(y,x)]+Ex,y[log(1-D(y,G(y,z)))]Wherein y is training condition data, x is real data, G (y, z) represents a target domain model generated by the generation network G according to the training condition data y and a random vector, D (y, x) represents the probability of judging whether the real model is real by the D network, and D (y, G (y, z)) is the probability of judging whether the model G (y, z) generated by the generation network G is real by the network D;
s440 adds a second loss function L1 between the model G (y, z) generated by the generating network G and the real data to speed up the convergence of the model and improve the accuracy of the model1(G)＝Ex，y[||x-G(y，z)||1]Wherein y is training condition data, x is real data, and G (y, z) represents a target domain model generated by the generation network G according to the training condition data y and the random vector.
S500, training a discrimination network D and a generation network G to obtain a discriminator and a generator; referring to fig. 7 and 8, the method for training the discriminant network and the generator network includes the following steps:
s510, selecting training condition data y, and inputting the selected training condition data y into the generation network G, so that the generation network G generates a second river sand body model G (y) corresponding to the selected training condition data y;
s520, inputting the selected training condition data y and the corresponding first river channel sand body model A into a judging network D as judging conditions, inputting the corresponding second river channel sand body model G (y) into the judging network D as a judging object, judging the authenticity of the second river channel sand body model G (y) by the judging network D, and when the judging network D judges that the second river channel sand body model G (y) is false, adjusting the parameters of the generating network G according to feedback information, namely calculated loss values, and regenerating the second river channel sand body model G (y) corresponding to the selected training condition data y;
s530, according to the feedback information obtained in S520, adjusting the parameters of the discrimination network D, and improving the discrimination capability of the discrimination network;
s540 inputs the regenerated second channel sand body model G (y) into the discrimination network D for determination, and repeats steps S520 and S530 several times, thereby improving the accuracy of the determination network determination and improving the fidelity of the second channel sand body model generated by the generation network, so that through many times of confrontation training, the sand body model generated by the generation network G becomes more and more vivid, and the capability of the discrimination network to determine the authenticity of the sand body model becomes stronger and stronger, until the discrimination network D does not determine the authenticity of the second channel sand body model G (y), that is, the generation confrontation network reaches nash equilibrium, the determined probability is 0.5, finally the first loss function V (G, D) approaches a stable value, the second loss function L1 also gradually decreases, and converges to a stable value, and the established final generation function is: g ═ arg V (G, D) + λ L1(G), where λ is a hyperparameter controlling the ratio of L1 (G);
s550, repeating the steps S510 to S540 on the condition data y until all the training condition data y are trained.
S600, well data of the reservoir to be modeled are obtained, and the well data are input into a generator to obtain a sand body model of the reservoir to be modeled. In this embodiment, please refer to fig. 9 for the well data of the research area, which collectively includes 9 wells, where there are 6 wells drilled in the river, these data are used as the condition data for generating the reservoir model of the sand body of the river in the research area, and the sand body model satisfying the condition data can be generated through convolution and deconvolution operations (fig. 10).
The channel sand body model finally generated by the channel sand modeling method provided by the invention can meet all well data because the channel sand body model is obtained by inputting the well data of the research area into the generator, and meanwhile, the generator can reflect the real reservoir form because the generator is trained by a large number of first channel sand body models and corresponding training condition data. The technical scheme provided by the invention combines the idea of generating the countermeasure network by artificial intelligence deep learning, overcomes the defects of difficult conditioning and difficult recovery of the geometric form of sand bodies and the problem of difficult calculation of a variation function of a target-based modeling method, can well reflect the sedimentary characteristics of the reservoir and reproduce the characteristics of the complex form, the size and the like of the reservoir, provides another direction for the research of reservoir modeling, and better serves the production of oil fields.
It should be understood that, although the steps in the flowchart of fig. 1 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise.
The invention also provides a device for modeling the river channel sand, which comprises a processor and a memory; the memory has stored thereon a computer readable program executable by the processor; the processor implements the steps of the above-described channel sand modeling method when executing the computer-readable program, and the channel sand modeling method has been described in detail above, and thus is not described herein again.
In summary, the invention generates a large number of first river sand body models and corresponding training condition data through a target-based simulation method, trains the generation network and the discrimination network through the first river sand body models and the training condition data, obtains a generator capable of generating a vivid river sand body model after the generation network is trained in a large number, and inputs well data of a research area into the generator to obtain a sand body model of a reservoir to be modeled in the research area.
Of course, it will be understood by those skilled in the art that all or part of the processes of the methods of the above embodiments may be implemented by a computer program instructing relevant hardware (such as a processor, a controller, etc.), and the program may be stored in a computer readable storage medium, and when executed, the program may include the processes of the above method embodiments. The storage medium may be a memory, a magnetic disk, an optical disk, etc.
The above-described embodiments of the present invention should not be construed as limiting the scope of the present invention. Any other corresponding changes and modifications made according to the technical idea of the present invention should be included in the protection scope of the claims of the present invention.
Claims (9)
1. A river sand modeling method is characterized by comprising the following steps:
s100, obtaining basic parameters of a reservoir to be modeled, wherein the basic parameters at least comprise river channel average width, river channel average curvature and river channel trend;
s200, generating a plurality of first river channel sand body models according to the acquired basic parameters, and generating a plurality of training condition data corresponding to the first river channel sand body models according to the first river channel sand body models;
s300, establishing a generating network according to each training condition data;
s400, establishing a discrimination network according to each training condition data and the corresponding first river channel sand body model;
s500, training the judging network and the generating network to obtain a generator;
s600, well data of the reservoir to be modeled are obtained, and the well data are input into a generator to obtain a sand body model of the reservoir to be modeled.
2. The method for modeling channel sand of claim 1, wherein the step S300 comprises:
s310, establishing a plurality of convolutional layers, coding each training condition data y, inputting the coded data into the convolutional layers for feature extraction, and obtaining a plurality of feature matrixes through convolution operation;
s320, establishing deconvolution layers with the same number as the convolution layers, connecting the convolution layers with the same dimensionality in the deconvolution layers, arranging the characteristic matrixes extracted in the step S310 into one-dimensional vectors, inputting the one-dimensional vectors into the deconvolution layers, and generating corresponding river sand model matrixes through deconvolution operation;
s330, a correction function for generating network parameters in the training process is set.
3. The method for modeling riverway sand of claim 2, wherein the step S330 is performed in step S330The correction function for the generated network is: theta is equal to thetag- ▽, where θ is the network corrected parameter, θgTo generate the parameters before network correction, ▽ is the loss gradient for the parameters, and the loss gradient
Wherein n is the total training times, D is the discrimination network, G is the generation network, z is the random vector, and i is the training time variable.
4. The method of modeling channel sand of claim 1, wherein the step S400 comprises:
s410, establishing a plurality of convolutional layers, encoding each training condition data and the corresponding first river channel sand body model in pairs, inputting the encoded training condition data and the corresponding first river channel sand body model into the convolutional layers for feature extraction, and obtaining a plurality of feature matrixes through convolution operation;
s420, comparing the training condition data with the corresponding characteristic matrix of the first river channel sand body model, and respectively calculating the real probability of each first river channel sand body model;
s430, inputting the sand body model generated by the generated network into a discrimination network, and establishing a first loss function;
s440 adds a second loss function between the model generated by the generating network and the real data to speed up the model convergence and improve the accuracy of the model.
5. The method for modeling channel sand of claim 4, wherein in step S430, the first loss function is: v (G, D) ═ Ex,y[logD(y,x)]+Ex,y[log(1-D(y,G(y,z)))]Wherein y is training condition data, x is real data, G (y, z) represents a target domain model generated by the generation network G according to the training condition data y and a random vector, D (y, x) represents the probability of judging whether the real model is real by the D network, and D (y, G (y, z)) is a model G (y, z) generated by judging the generation network G by the network D) Probability of being true or not.
6. The method for modeling channel sand of claim 4, wherein in step S440, the second loss function is: l is1(G)＝Ex,y[||x-G(y,z)||1]Wherein y is training condition data, x is real data, and G (y, z) represents a target domain model generated by the generation network G according to the training condition data y and the random vector z.
7. The method of modeling channel sand of claim 1, wherein the step S500 comprises:
s510, selecting training condition data, and inputting the selected training condition data into the generation network, so that the generation network generates a second river channel sand body model corresponding to the selected training condition data;
s520, inputting the selected training condition data and the corresponding first river channel sand body model and second river channel sand body model into a discrimination network, judging the authenticity of the second river channel sand body model by the discrimination network, and when the discrimination network judges that the second river channel sand body model is false, adjusting the parameters of the generation network according to feedback information, namely the calculated loss value, and regenerating the second river channel sand body model corresponding to the selected training condition data;
s530, according to the feedback information obtained in S520, adjusting the parameters of the discrimination network and improving the discrimination capability of the discrimination network;
s540, inputting the second river channel sand body model regenerated in S520 into the discrimination network, and repeating the step S520 and the step S530 for a plurality of times, so that the accuracy of judgment of the discrimination network is improved, and the fidelity of the second river channel sand body model generated by the generation network is improved until the discrimination network cannot judge the authenticity of the second river channel sand body model;
s550, repeating the steps S510 to S540 on all the condition data until all the training condition data are trained.
8. The device for modeling the river channel sand is characterized by comprising a processor and a memory;
the memory has stored thereon a computer readable program executable by the processor;
the processor, when executing the computer readable program, implements the steps in the method of river sand modeling according to any one of claims 1-3.
9. A computer readable storage medium, storing one or more programs, which are executable by one or more processors to implement the steps of the method of modeling channel sand according to any one of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010022741.5A CN111209674B (en) | 2020-01-09 | 2020-01-09 | River channel sand modeling method and device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010022741.5A CN111209674B (en) | 2020-01-09 | 2020-01-09 | River channel sand modeling method and device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111209674A true CN111209674A (en) | 2020-05-29 |
| CN111209674B CN111209674B (en) | 2023-04-11 |
Family
ID=70786027
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010022741.5A Active CN111209674B (en) | 2020-01-09 | 2020-01-09 | River channel sand modeling method and device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111209674B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114969895A (en) * | 2022-04-26 | 2022-08-30 | 长江大学 | River channel profile morphology modeling method based on trend constraint |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108665546A (en) * | 2018-05-17 | 2018-10-16 | 四川大学 | A kind of multiple spot geological statistics three-dimensional modeling method of combination deep learning |
| CN109446735A (en) * | 2018-12-18 | 2019-03-08 | 中国石油大学(北京) | A kind of generation method, equipment and the system of modeling logging data |
| US20190080205A1 (en) * | 2016-09-16 | 2019-03-14 | Deep Learning Analytics, LLC | Systems and Methods for Deep Model Translation Generation |
| US20190107642A1 (en) * | 2017-10-11 | 2019-04-11 | Beyond Limits, Inc. | System for improved reservoir exploration and production |
| CN110070174A (en) * | 2019-04-10 | 2019-07-30 | 厦门美图之家科技有限公司 | A kind of stabilization training method generating confrontation network |
| CN110414601A (en) * | 2019-07-30 | 2019-11-05 | 南京工业大学 | Photovoltaic module method for diagnosing faults, system and equipment based on depth convolution confrontation network |
-
2020
- 2020-01-09 CN CN202010022741.5A patent/CN111209674B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190080205A1 (en) * | 2016-09-16 | 2019-03-14 | Deep Learning Analytics, LLC | Systems and Methods for Deep Model Translation Generation |
| US20190107642A1 (en) * | 2017-10-11 | 2019-04-11 | Beyond Limits, Inc. | System for improved reservoir exploration and production |
| CN108665546A (en) * | 2018-05-17 | 2018-10-16 | 四川大学 | A kind of multiple spot geological statistics three-dimensional modeling method of combination deep learning |
| CN109446735A (en) * | 2018-12-18 | 2019-03-08 | 中国石油大学(北京) | A kind of generation method, equipment and the system of modeling logging data |
| CN110070174A (en) * | 2019-04-10 | 2019-07-30 | 厦门美图之家科技有限公司 | A kind of stabilization training method generating confrontation network |
| CN110414601A (en) * | 2019-07-30 | 2019-11-05 | 南京工业大学 | Photovoltaic module method for diagnosing faults, system and equipment based on depth convolution confrontation network |
Non-Patent Citations (6)
| Title |
|---|
| DARIO A. B. OLIVEIRA , RODRIGO S. FERREIRA , REINALDO SILVA, AND EMILIO VITAL BRAZIL: "Interpolating Seismic Data With Conditional Generative Adversarial Networks" * |
| EMILIEN DUPONT，TUANFENG ZHANG，PETER TILKE，LIN LIANG，WILLIAM BAILEY: "Generating Realistic Geology Conditioned on Physical Measurements with Generative Adversarial Networks" * |
| ÉRIC GUÉRIN, JULIE DIGNE, ÉRIC GALIN, ADRIEN PEYTAVIE, CHRISTIAN WOLF, BEDRICH BENES, BENOÎT MARTINEZ: "Interactive example-based terrain authoring with conditional generative adversarial networks" * |
| RODRIGO EXTERKOETTER，FERNANDO BORDIGNON，LEANDRO PASSOS DE FIGUEIREDO，MAURO ROISENBERG，BRUNO BARBOSA RODRIGUES: "Petroleum Reservoir Connectivity Patterns Reconstruction using Deep Convolutional Generative Adversarial Networks" * |
| SHING CHAN，AHMED H. ELSHEIKH: "Parametric generation of conditional geological realizations using generative neural networks" * |
| 魏庆，李向阳: "基于条件生成式对抗网络的地震数据重构算法" * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114969895A (en) * | 2022-04-26 | 2022-08-30 | 长江大学 | River channel profile morphology modeling method based on trend constraint |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111209674B (en) | 2023-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108334816B (en) | Multi-pose face recognition method based on contour symmetric constraint generation type countermeasure network | |
| CN109611087B (en) | Volcanic oil reservoir parameter intelligent prediction method and system | |
| CN111563841B (en) | High-resolution image generation method based on generation countermeasure network | |
| CN113052371B (en) | Residual oil distribution prediction method and device based on deep convolutional neural network | |
| CN111523713B (en) | Method and device for predicting saturation distribution of residual oil in oil field | |
| CN111612906B (en) | Method and system for generating three-dimensional geological model and computer storage medium | |
| CN112489164B (en) | Image coloring method based on improved depth separable convolutional neural network | |
| CN105354873A (en) | Mode density function simulation algorithm for three-dimensional reconstruction of porous media | |
| CN109146061A (en) | The treating method and apparatus of neural network model | |
| Liu et al. | Multilevel strategies and geological parameterizations for history matching complex reservoir models | |
| CN110516724A (en) | Visualize the high-performance multilayer dictionary learning characteristic image processing method of operation scene | |
| CN110991621A (en) | Method for searching convolutional neural network based on channel number | |
| CN117290721A (en) | Digital twin modeling method, device, equipment and medium | |
| CN111209674B (en) | River channel sand modeling method and device and readable storage medium | |
| CN112541256A (en) | Deep learning dimensionality reduction reconstruction-based strong heterogeneous reservoir history fitting method | |
| CN112016956B (en) | Ore grade estimation method and device based on BP neural network | |
| CN113554104B (en) | Image classification method based on deep learning model | |
| CN114897884A (en) | No-reference screen content image quality evaluation method based on multi-scale edge feature fusion | |
| CN115063847A (en) | Training method and device for facial image acquisition model | |
| CN112560326B (en) | Method and device for determining pressure field | |
| CN108875163B (en) | Method and system for evaluating three-dimensional fracture network connectivity | |
| CN112199887A (en) | Semi-supervised resistivity inversion method and system based on countermeasure generation network and pseudo-labeling | |
| Zhao et al. | 3-D point cloud normal estimation based on fitting algebraic spheres | |
| CN112949719B (en) | Well testing interpretation proxy model generation method based on GAN | |
| CN117436350B (en) | Fracturing horizontal well pressure prediction method based on deep convolution generation countermeasure network |
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 |