CN110412662A - Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning - Google Patents
Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning Download PDFInfo
- Publication number
- CN110412662A CN110412662A CN201910572785.2A CN201910572785A CN110412662A CN 110412662 A CN110412662 A CN 110412662A CN 201910572785 A CN201910572785 A CN 201910572785A CN 110412662 A CN110412662 A CN 110412662A
- Authority
- CN
- China
- Prior art keywords
- prediction
- seismic
- data
- thin
- deep learning
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V11/00—Prospecting or detecting by methods combining techniques covered by two or more of main groups G01V1/00 - G01V9/00
Abstract
The present invention provides a kind of method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning, the method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning includes the following steps: (1) that a variety of seismic properties of interval of interest are calculated according to the interpretation horizon of research interval, extracts the sandstone thickness and lithofacies classification information of drilling well；Step 2, sandstone thickness and lithofacies classification data using well point, and machine learning model of the various seismic attributes datas training based on supervision extracted；Step 3, the sandstone thickness prediction that trained machine learning model is used for no well area and lithofacies classification prediction.The method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning combines well point data and seismic attributes data, and with well point data constraint seismic attributes data, improves the precision of prediction of thin interbed sandstone reservoir.
Description
Technical field
The present invention relates to oil field development technical fields, especially relate to a kind of based on the thin of seismic multi-attribute deep learning
Alternating layers method for predicting reservoir.
Background technique
Important kind one of of the interbedded reservoir as subtle pool, the multiple gentle slope belt educated in continental rift lacustrine basin,
It is big with oil area, reserves abundance is medium famous, be increasingly subject to the attention of exploration and development.Interbedded reservoir is by different terrigenous deposits
Object composition, reservoir lithology is locally Sandy Silt based on packsand, is ligh-oil reservoir based on structural-stratigraphic trap.
Although the super stripping band shallow lacustrine reservoir potentiality in shore in stratum are big, High-quality Reservoir earthquake detailed predicting is difficult.Plane oil gas
Pool forming rules complicated and severe restricts subsequent exploration and development.It is mainly manifested in the following aspects: 1) the super stripping band shore Vlei phase in stratum
Reservoir is thin interbed deposition, and single layer 2-5m identifies that difficulty is big.2) the super stripping band shore Vlei phase Reservoir Lithofacies in stratum, rock electricity, structure are more
Sample results in longitudinal lithology combination rhythm complexity, and subsequent plane reservoir prediction difficulty is larger.3) the super stripping band shore Vlei phase in stratum
Reservoir integral thickness is little, is quickly thinned toward basin edge gentle slope belt reservoir, and the combination of the lithology rhythm is complicated on longitudinal direction, by supersequence circle
It is lower that face covers interference longitudinal resolution, internal Chao Bao ambiguity Chu, and reservoir interlayer reflecting interface is not under common seismic data
Clearly, multi-solution is strong, and stratum is super at present shells band shore Vlei phase reservoir prediction still based on common seismic data, but purpose
Layer resolution ratio is unsatisfactory for actual reservoir earthquake prediction needs, and the super stripping structure determination in stratum and reservoir prediction heavy dependence are opened up frequency and provided
The super stripping band explanatory processing method of shore Vlei phase different type seismic reservoir high-resolution of material, especially shortage stratum.4) stratum is super
The stripping band shallow lacustrine reservoir in shore is mostly dam deposition in sand-mud interbed beach other than the fan body of gentle slope, and the height of oil reservoir production capacity depends on
Reservoir situation, the sand group rank reservoir thickness seismic prediction technique precision of window is lower when big in the past, the reservoir and ground of different sequence grades
It is indefinite to shake relation on attributes, is unable to satisfy thickness in monolayer prediction actual needs.5) the super stripping band shore Vlei in different zones stratum mutually stores up
Layer is big at hiding difference, not formed specific aim Comprehensive Assessment Technology.
The super stripping in stratum becomes the major issue of current urgent need to resolve with shore Vlei phase reservoir fine earthquake prediction thus, because
It is necessary to be studied for this.Exploration practices prove that a variety of seismic properties of thin interbed sandstone reservoir and reservoir have complicated system
Count relationship.We have invented a kind of new method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning, solutions thus
It has determined the above technical problem.
Summary of the invention
The super stripping band shore Vlei phase reservoir seismic prediction precision in stratum can be improved the object of the present invention is to provide one kind and answer
With specific aim, subsequent exploration and development is instructed, the interbedded reservoir based on seismic multi-attribute deep learning for reducing exploration risk is pre-
Survey method.
The purpose of the present invention can be achieved by the following technical measures: the thin interbed storage based on seismic multi-attribute deep learning
Layer prediction method, being somebody's turn to do the method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning includes the following steps: (1) according to research layer
The interpretation horizon of section calculates a variety of seismic properties of interval of interest, extracts the sandstone thickness and lithofacies classification information of drilling well；Step
Rapid 2, using the sandstone thickness and lithofacies classification data of well point, and the various seismic attributes datas training extracted is based on supervision
Machine learning model；Step 3, the sandstone thickness prediction that trained machine learning model is used for no well area and lithofacies classification
Prediction.
The purpose of the present invention can be also achieved by the following technical measures:
Step 1 includes:
S11, it is constrained using the layer position of seismic interpretation, extracts the seismic multi-attribute of interval of interest, including amplitude, frequency
These statistics class seismic properties of rate, phase；
S12, according to the corresponding relationship of earthquake and drilling well, count mesh using natural gamma and nutural potential logging profile
Interval sandstone thickness and lithofacies classify；
S13, seismic properties and sandstone thickness and lithofacies classification corresponding relationship are established；Well location point is extracted according to well point coordinate
The data of various seismic properties establish the database of well point position sandstone thickness and lithofacies classification and various seismic properties.
Step 2 includes:
S21, the statistical data for counting well point include that sandstone thickness and lithofacies are classified and various seismic properties, to data into
Row divides in proportion；
S22, selected depth mode of learning；
S23, framework multilayer deep learning model depth parameter, that is, model learning and error propagation number and level；
The end threshold values that S24, setting model are trained.
In the step s 21, a part of data are used for training machine learning model, the essence of a part of data verification training pattern
Degree, a part of data are used to test the precision of training pattern.
In step S22, is stacked based on multiple limited Boltzmann machines and form depth confidence network, be limited Boltzmann machine
As a kind of special feature extractor, the sand of encoding seismic multiattribute data and well point
In step S23, the longest path from input layer to output layer is set, model is once known by one layer of study
Know, the depth of study is initial data by the number of Level by level learning.
In step s 24, Definition Model learning error reaches how many requirements for meeting prediction, and model deconditioning simultaneously can
Predict unsupervised data；Error point is carried out using the practical lithofacies classification data and sandstone thickness data and the data of prediction of well point
Analysis, reduces the error of model prediction, control errors within permissible accuracy, sets the required precision of prediction and model is obtained
With application.
Include: in step 3
S31, short according to the second step selected deep learning model training time, the high model prediction of precision of prediction is without well area
Sandstone thickness and lithofacies classify, preferably the high training pattern of precision of prediction to no well area carry out lithofacies classification prediction and sandstone
Thickness distribution prediction.
The method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning in the present invention, is related to applied geophysics
It is distributed with the classification of well data associated prediction lithofacies and sandstone thickness, particularly relates to constrain seismic multi-attribute algorithm based on well data
It is a kind of using seismic multi-attribute and lithofacies classification and sandstone thickness distribution correlativity high-precision forecast sandstone be distributed side
Method.This is mutually hidden towards the super stripping band shore Vlei in stratum based on the method for prediction of reservoirs of thin interbeded research of seismic multi-attribute deep learning
Oil-gas reservoir target is studied from actual production demand and is suitble to the super stripping band shore Vlei phase different type reservoir high-resolution in stratum
The processing of earthquake target and fine earthquake prediction method solve the problems, such as the super stripping in stratum with shore Vlei phase reservoir prediction hardly possible.It is studied
Meaning is that the super stripping band shore Vlei phase reservoir seismic prediction precision in stratum can be improved and using specific aim, subsequent exploration is instructed to open
Hair reduces exploration risk.
The method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning in the present invention, including the use of log
It divides and determines thin sandstone thickness；And each attribute of earthquake and sandstone thickness, and the phase with lithology are statisticallyd analyze at well point
It closes in relational process, using the data of well point as label data, the different deep learning model of training simultaneously predicts sandstone thickness
With lithology breakdown, preferably training pattern algorithm model with high accuracy is prediction model, carries out the thin-inter bed reservoir prediction in work area, is wrapped
It includes sandstone thickness prediction and lithology breakdown is predicted.Should be based on well point statistical data and seismic properties correlation analysis, and test difference
Algorithm model, preferably precision higher prediction model prediction thin interbed sandstone distribution, this method combines well point data and ground
Attribute data is shaken, and with well point data constraint seismic attributes data, improves the precision of prediction of thin interbed sandstone reservoir.
Detailed description of the invention
Fig. 1 is a specific embodiment of the method for prediction of reservoirs of thin interbeded of the invention based on seismic multi-attribute deep learning
Flow chart；
Fig. 2 is the training pattern lithofacies nicety of grading figure calculated in a specific embodiment of the invention；
Fig. 3 is training sandstone thickness neural network prediction model precision analysis figure in a specific embodiment of the invention；
Fig. 4 is the schematic diagram of the lithofacies distribution of deep learning prediction in a specific embodiment of the invention；
Fig. 5 is the schematic diagram of the sandstone thickness distribution of deep learning prediction in a specific embodiment of the invention.
Specific embodiment
To enable above and other objects, features and advantages of the invention to be clearer and more comprehensible, preferably implementation is cited below particularly out
Example, and cooperate shown in attached drawing, it is described in detail below.
As shown in FIG. 1, FIG. 1 is the streams of the method for prediction of reservoirs of thin interbeded of the invention based on seismic multi-attribute deep learning
Cheng Tu.
S101, a variety of seismic properties of interval of interest are calculated according to the interpretation horizon of research interval and extract the sand of drilling well
Rock thickness and lithofacies classification information:
S11, it is constrained using the layer position of seismic interpretation, extracts the seismic multi-attribute of interval of interest, including amplitude, frequency
These statistics class seismic properties of rate, phase；
S12, according to the corresponding relationship of earthquake and drilling well, count mesh using natural gamma and nutural potential logging profile
Interval sandstone thickness and lithofacies classify；
S13, seismic properties and sandstone thickness and lithofacies classification corresponding relationship are established；Well location point is extracted according to well point coordinate
The data of various seismic properties establish the database of well point position sandstone thickness and lithofacies classification and various seismic properties；
S102, the sandstone thickness and lithofacies classification data of well point and the seismic attributes data training inhomogeneity of well point are utilized
The machine learning model based on supervision of type:
S21, the statistical data for analyzing well point include that sandstone thickness and lithofacies are classified and various seismic properties, to data into
Row divides in proportion, and a portion data (70%) are used for training machine learning model, a part of data (20%) verifying instruction
Practice the precision of model, a part of data (10%) are used to test the precision of training pattern；
S22, selected depth mode of learning type: being stacked based on multiple limited Boltzmann machines and form depth confidence network,
Limited Boltzmann machine is as a kind of special feature extractor, the sandstone thickness and rock of encoding seismic multiattribute data and well point
Result is used for the classification or recurrence of supervised learning through overfitting by phase distributional class；
S23, set multilayer deep learning model depth parameter, that is, model learning and error propagation number and level.
The longest path from input layer to output layer is set, model obtains a knowledge by one layer of study, then the depth learnt can
It is interpreted as number of the initial data by Level by level learning.
The end threshold values that S24, setting model are trained, i.e. Definition Model learning error reach how many and meet wanting for prediction
It asks, model deconditioning simultaneously can predict unsupervised data.Using the practical lithofacies classification data and sandstone thickness data of well point
Error analysis is carried out with the data of prediction, reduces the error of model prediction, control errors within permissible accuracy, setting is pre-
The required precision of survey simultaneously applies model.
S103, the sandstone thickness prediction that trained machine learning model is used for no well area and lithofacies classification prediction.
S31, the training pattern established according to well point sandstone thickness and lithofacies classification and various seismic properties, are preferably predicted
Training pattern with high accuracy carries out lithofacies classification prediction and sandstone thickness forecast of distribution to no well area.
Fig. 2 is to be analyzed using method of the invention the model accuracy that lithofacies are predicted, orbicular spot represents prediction correctly, poor
Point represents prediction error, wherein number 0 represents mud stone sample, 1 represents sandstone phase, as can be seen from the figure model prediction accuracy compared with
Height, accuracy rate reaches 80% or more according to statistics.Fig. 3 is the neural network prediction model precision analysis to well point sandstone thickness, from
It can be seen that the correlation of training set reaches 0.79 on prediction model, the correlation for verifying collection reaches 0.67, and overall model is pre-
It surveys correlation and reaches 0.69, illustrate that the predictive ability of the machine learning model is strong, precision is high.
Fig. 4 is based on the lithofacies classification chart of seismic multi-attribute deep learning prediction, and Fig. 5 is that seismic multi-attribute deep learning is pre-
The sandstone thickness distribution map of survey compares two figures as can be seen that sand lithofacies distribution and sand thickness distribution coincide, and this method can be quasi-
The really sandstone reservoir in prediction thin interbed.
The limitation that technical solution of the present invention is not limited to the above specific embodiments, it is all to do according to the technique and scheme of the present invention
Technology deformation out, falls within the scope of protection of the present invention.
Claims (8)
1. the method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning, which is characterized in that seismic multi-attribute should be based on
The method for prediction of reservoirs of thin interbeded of deep learning includes:
Step 1, a variety of seismic properties that interval of interest is calculated according to the interpretation horizon of research interval, the sandstone for extracting drilling well are thick
Degree and lithofacies classification information；
Step 2, sandstone thickness and lithofacies classification data using well point, and the various seismic attributes datas training extracted are based on
The machine learning model of supervision；
Step 3, the sandstone thickness prediction that trained machine learning model is used for no well area and lithofacies classification prediction.
2. the method for prediction of reservoirs of thin interbeded according to claim 1 based on seismic multi-attribute deep learning, feature exist
In step 1 includes:
S11, it is constrained using the layer position of seismic interpretation, extracts the seismic multi-attribute of interval of interest, including amplitude, frequency, phase
These statistics class seismic properties of position；
S12, according to the corresponding relationship of earthquake and drilling well, count target zone using natural gamma and nutural potential logging profile
The sandstone thickness and lithofacies of section are classified；
S13, seismic properties and sandstone thickness and lithofacies classification corresponding relationship are established；The various of well location point are extracted according to well point coordinate
The data of seismic properties establish the database of well point position sandstone thickness and lithofacies classification and various seismic properties.
3. the method for prediction of reservoirs of thin interbeded according to claim 1 based on seismic multi-attribute deep learning, feature exist
In step 2 includes:
S21, the statistical data for counting well point include that sandstone thickness and lithofacies are classified and various seismic properties, to data carry out by
Ratio cut partition；
S22, selected depth mode of learning；
S23, framework multilayer deep learning model depth parameter, that is, model learning and error propagation number and level；
The end threshold values that S24, setting model are trained.
4. the method for prediction of reservoirs of thin interbeded according to claim 3 based on seismic multi-attribute deep learning, feature exist
In, in the step s 21, a part of data be used for training machine learning model, the precision of a part of data verification training pattern, one
Partial data is used to test the precision of training pattern.
5. the method for prediction of reservoirs of thin interbeded according to claim 3 based on seismic multi-attribute deep learning, feature exist
In, in step S22, is stacked based on multiple limited Boltzmann machines and form depth confidence network, limited Boltzmann machine conduct
The sandstone thickness of special feature extractor, encoding seismic multiattribute data and well point a kind of and lithofacies distribution class, through overfitting,
Result is used for the classification or recurrence of supervised learning.
6. the method for prediction of reservoirs of thin interbeded according to claim 3 based on seismic multi-attribute deep learning, feature exist
In in step S23, longest path of the setting from input layer to output layer, model obtains a knowledge by one layer of study, learns
The depth of habit is initial data by the number of Level by level learning.
7. the method for prediction of reservoirs of thin interbeded according to claim 3 based on seismic multi-attribute deep learning, feature exist
In in step s 24, Definition Model learning error reaches how many requirements for meeting prediction, and model deconditioning simultaneously can be predicted
Unsupervised data；Error analysis is carried out using the practical lithofacies classification data and sandstone thickness data and the data of prediction of well point,
The error for reducing model prediction sets the required precision of prediction and model is able to control errors within permissible accuracy
Using.
8. the method for prediction of reservoirs of thin interbeded according to claim 1 based on seismic multi-attribute deep learning, feature exist
In including: in step 3
S31, precision of prediction high model prediction sand without well area short according to the second step selected deep learning model training time
Rock thickness and lithofacies are classified, and preferably the high training pattern of precision of prediction carries out lithofacies classification prediction and sandstone thickness to no well area
Forecast of distribution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910572785.2A CN110412662A (en) | 2019-06-27 | 2019-06-27 | Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910572785.2A CN110412662A (en) | 2019-06-27 | 2019-06-27 | Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110412662A true CN110412662A (en) | 2019-11-05 |
Family
ID=68358344
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910572785.2A Pending CN110412662A (en) | 2019-06-27 | 2019-06-27 | Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110412662A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111472763A (en) * | 2019-12-30 | 2020-07-31 | 北京国双科技有限公司 | Stratum thickness prediction method and device |
| CN111695729A (en) * | 2020-06-08 | 2020-09-22 | 浙江大学 | Aftershock prediction method based on DMAP model |
| CN114167498A (en) * | 2021-11-30 | 2022-03-11 | 中海石油(中国)有限公司 | Well logging and seismic attribute combined step-by-step clustering analysis method |
| CN115421181A (en) * | 2022-07-27 | 2022-12-02 | 北京超维创想信息技术有限公司 | Three-dimensional geological model phase control attribute modeling method based on deep learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140309937A1 (en) * | 2010-03-05 | 2014-10-16 | Vialogy Llc | Active Noise Injection Computations for Improved Predictability in Oil and Gas Reservoir Characterization and Microseismic Event Analysis |
| CN108921058A (en) * | 2018-06-19 | 2018-11-30 | 厦门大学 | Fish identification method, medium, terminal device and device based on deep learning |
| US20190064378A1 (en) * | 2017-08-25 | 2019-02-28 | Wei D. LIU | Automated Seismic Interpretation Using Fully Convolutional Neural Networks |
| CN109685127A (en) * | 2018-12-17 | 2019-04-26 | 郑州云海信息技术有限公司 | A kind of method and system of parallel deep learning first break pickup |
| CN109709612A (en) * | 2018-11-07 | 2019-05-03 | 中国石油大学(北京) | Sand-body distribution prediction technique, device, equipment and readable storage medium storing program for executing |
| CN109765609A (en) * | 2019-01-30 | 2019-05-17 | 中国石油大学(北京) | A kind of Sand-body Prediction method and system based on target zone Yu adjacent layer seismic properties |
| CN109799533A (en) * | 2018-12-28 | 2019-05-24 | 中国石油化工股份有限公司 | A kind of method for predicting reservoir based on bidirectional circulating neural network |
-
2019
- 2019-06-27 CN CN201910572785.2A patent/CN110412662A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140309937A1 (en) * | 2010-03-05 | 2014-10-16 | Vialogy Llc | Active Noise Injection Computations for Improved Predictability in Oil and Gas Reservoir Characterization and Microseismic Event Analysis |
| US20190064378A1 (en) * | 2017-08-25 | 2019-02-28 | Wei D. LIU | Automated Seismic Interpretation Using Fully Convolutional Neural Networks |
| CN108921058A (en) * | 2018-06-19 | 2018-11-30 | 厦门大学 | Fish identification method, medium, terminal device and device based on deep learning |
| CN109709612A (en) * | 2018-11-07 | 2019-05-03 | 中国石油大学(北京) | Sand-body distribution prediction technique, device, equipment and readable storage medium storing program for executing |
| CN109685127A (en) * | 2018-12-17 | 2019-04-26 | 郑州云海信息技术有限公司 | A kind of method and system of parallel deep learning first break pickup |
| CN109799533A (en) * | 2018-12-28 | 2019-05-24 | 中国石油化工股份有限公司 | A kind of method for predicting reservoir based on bidirectional circulating neural network |
| CN109765609A (en) * | 2019-01-30 | 2019-05-17 | 中国石油大学(北京) | A kind of Sand-body Prediction method and system based on target zone Yu adjacent layer seismic properties |
Non-Patent Citations (3)
| Title |
|---|
| ZHI ZHONG，等: "A deep learning approach to anomaly detection in geological carbon sequestration sites using pressure measurements", 《JOURNAL OF HYDROLOGY》 * |
| 胡越，等: "关于深度学习的综述与讨论", 《智能系统学报》 * |
| 郑阳: "基于深度学习的岩性识别研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111472763A (en) * | 2019-12-30 | 2020-07-31 | 北京国双科技有限公司 | Stratum thickness prediction method and device |
| CN111695729A (en) * | 2020-06-08 | 2020-09-22 | 浙江大学 | Aftershock prediction method based on DMAP model |
| CN114167498A (en) * | 2021-11-30 | 2022-03-11 | 中海石油(中国)有限公司 | Well logging and seismic attribute combined step-by-step clustering analysis method |
| CN114167498B (en) * | 2021-11-30 | 2023-09-15 | 中海石油(中国)有限公司 | Step-by-step cluster analysis method combining logging and seismic attribute |
| CN115421181A (en) * | 2022-07-27 | 2022-12-02 | 北京超维创想信息技术有限公司 | Three-dimensional geological model phase control attribute modeling method based on deep learning |
| CN115421181B (en) * | 2022-07-27 | 2023-10-20 | 北京超维创想信息技术有限公司 | Three-dimensional geological model phase control attribute modeling method based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qi et al. | Neural network prediction of carbonate lithofacies from well logs, Big Bow and Sand Arroyo Creek fields, Southwest Kansas | |
| CN110412662A (en) | Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning | |
| US10578767B2 (en) | Conditional process-aided multiple-points statistics modeling | |
| CN102918423A (en) | Method for seismic hydrocarbon system analysis | |
| CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
| CN111080021B (en) | Sand body configuration CMM neural network prediction method based on geological information base | |
| WO2009134685A2 (en) | System and method for interpretation of well data | |
| Eskandari et al. | Reservoir Modelling of Complex Geological Systems--A Multiple-Point Perspective | |
| Tan et al. | Evaluation of complex petroleum reservoirs based on data mining methods | |
| Liu et al. | NMR log response prediction from conventional petrophysical logs with XGBoost-PSO framework | |
| CN110609327B (en) | Carbonate reservoir facies prediction method and device based on pre-stack seismic attributes | |
| Kang et al. | Managing uncertainty in geological scenarios using machine learning-based classification model on production data | |
| Ikpeka et al. | Application of machine learning models in predicting initial gas production rate from tight gas reservoirs | |
| CN116168224A (en) | Machine learning lithology automatic identification method based on imaging gravel content | |
| Huang et al. | Quantitative analysis of the main controlling factors of oil saturation variation | |
| Dramsch | Machine Learning in 4D Seismic Data Analysis | |
| Liu et al. | Machine learning assisted recovery of subsurface energy: a review | |
| Chikhi et al. | Probabilistic neural method combined with radial-bias functions applied to reservoir characterization in the Algerian Triassic province | |
| Kuroda et al. | Analysis of porosity, stratigraphy, and structural delineation of a Brazilian carbonate field by machine learning techniques: A case study | |
| Emelyanova et al. | Unsupervised identification of electrofacies employing machine learning | |
| CN112379435A (en) | Phase-controlled karst type seam hole aggregate carving method and device | |
| Wang et al. | Quantitative evaluation of unconsolidated sandstone heavy oil reservoirs based on machine learning | |
| Wang et al. | Auto recognition of carbonate microfacies based on an improved back propagation neural network | |
| Carrasquilla et al. | Using facies, data mining and artificial intelligence concepts in the evaluation of a carbonate reservoir in Campos basin, Southeastern Brazil | |
| Fan et al. | Quantitative Prediction of Low-Permeability Sandstone Grain Size Based on Conventional Logging Data by Deep Neural Network-Based BP Algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191105 |
|
| RJ01 | Rejection of invention patent application after publication |