CN107290800A - Log Forecasting Methodology before practical brill - Google Patents
Log Forecasting Methodology before practical brill Download PDFInfo
- Publication number
- CN107290800A CN107290800A CN201610191318.1A CN201610191318A CN107290800A CN 107290800 A CN107290800 A CN 107290800A CN 201610191318 A CN201610191318 A CN 201610191318A CN 107290800 A CN107290800 A CN 107290800A
- Authority
- CN
- China
- Prior art keywords
- log
- log data
- elastic parameter
- forecasting methodology
- probability
- 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
-
- 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
- G01V11/002—Details, e.g. power supply systems for logging instruments, transmitting or recording data, specially adapted for well logging, also if the prospecting method is irrelevant
Landscapes
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The present invention, which provides log Forecasting Methodology before log Forecasting Methodology before a kind of practical brill, the practical brill, to be included：Step 1, log data is counted, obtains treating the accurate prior probability model of inverting log data；Step 2, a series of elastic parameter is obtained by Earthquake Resilient inverting；Step 3, log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method, completes horizontal extrapolation of the log data from well point to space, obtain log data spatial sampling result；Step 4, with reference to prior distribution, conditional probability, the Bayes classifier using log data as class object is built, the posterior probability based on Bayesian Classification Arithmetic is calculated, that is, predicts log data spatial distribution.Log Forecasting Methodology can solve the problem of Logging Curves Forecasting Methodology is depended on unduly to well-log information before the practical brill, accurately predict log by elastic parameter, reduce the risk of exploration, reduce cost.
Description
Technical field
The present invention relates to In Oil Field Exploration And Development technical field, especially relate to log well before a kind of practical brill
Curve prediction method.
Background technology
Traditional log Predicting Technique is typically to extract seismic properties by seismic data, according to acquisition
Relation between seismic properties and log property, predicts log.People once thought, seismic properties
Lack clear and definite mathematical relationship between some log properties, or even the two may be without positive connection.But it is near
Nian Lai, the experimental results that geophysical work person is done show, geological data and some log properties
Between exist certain conventional thought be difficult to understand non-linear relation.Log property is predicted with seismic properties,
It is that certain characteristic for geological data is drawn between target data and seismic properties using method earlier
Conventional Due date Window, this method assumes there is linear dependence between log and seismic properties, by returning
Straight line is fitted, the coefficient of straight line is obtained with least square predicated error.The calculating side proposed in the recent period
Method is nonlinear transformation, i.e. neural net method.Because the relation between log response and seismic properties is past
It is past sufficiently complex, it is difficult to expressed with explicit function (except the time difference and density log data), and nerve net
Such issues that network is to solving has advantage, and it can complete complicated between inputting and exporting non-linear reflect
Penetrate.By being repeatedly combined to simple nonlinear function, it is possible to achieve the conversion of complicated function relation.
Particularly neutral net has more perfect learning functionality, unique adaptive ability, associative memory ability
And information processing manner etc., can be this by finding using the Artificial Neural Network of seismic multi-attribute
Non-linear relation connects seismic properties and aim curve.
However, the log Forecasting Methodology based on neutral net advantage need using enough well controls as
Condition, occurs neutral net or the unstable feelings of variogram in the case where well-log information is not enough
Condition, influence finally predicts the outcome；On the other hand, in practice it has proved that many lithology transitivity log properties and bullet
Property parameter there is close contact, it is a kind of under conditions of well logging is rare it is therefore necessary to propose, by ground
The elastic parameter that shake inverting is obtained predicts the new method of log.For this, we have invented a kind of new reality
Log Forecasting Methodology before brill, solves above technical problem.
The content of the invention
Logging Curves Forecasting Methodology can be solved to well-log information mistake it is an object of the invention to provide one kind
The problem of degree is relied on, provides for geological research personnel and is logged well before the practical brill of a set of simple and effective scheme
Curve prediction method.
The purpose of the present invention can be achieved by the following technical measures：Log prediction side before practical brill
Log Forecasting Methodology includes before method, the practical brill：Step 1, log data is counted, is treated
The accurate prior probability model of inverting log data；Step 2, obtain a series of by Earthquake Resilient inverting
Elastic parameter；Step 3, carry out that log data spatial distribution is discrete adopts using Monte Carlo sampling method
Sample, completes horizontal extrapolation of the log data from well point to space, obtains log data spatial sampling result；
Step 4, with reference to prior distribution, conditional probability, the Bayes point using log data as class object is built
Class device, calculates the posterior probability based on Bayesian Classification Arithmetic, that is, predicts log data spatial distribution.
The purpose of the present invention can be also achieved by the following technical measures：
In step 1, if prediction target is physical property curve, its prior probability model is multivariate Gaussian point
Cloth；If prediction target is lithology curve, its prior probability model is Markov chain model.
In step 1, the correction method of surroundings effecting is carried out to log data used in reservoir description, and to surveying
Well curve is standardized.
In step 2, the elastic parameter obtained includes velocity of longitudinal wave, Young's modulus, Poisson's ratio, shearing
Modulus, each elastic parameter has different rock physicses meanings.
In step 2, according to the log type to be predicted, preferably go out appropriate mutual uncorrelated
Different elastic parameters participate in curve prediction, it is ensured that the class conditional independence assumption of Bayesian Classification Arithmetic into
It is vertical.
In step 3, mathematical statistics, meter are carried out to log data sample data and the elastic parameter chosen
Calculate conditional probability；Log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method, is completed
Horizontal extrapolation of the log data from well point to space, the sample characteristics finally given must be with well point basic one
Cause, and information should more be enriched, the geological conditions not embodied comprising more well-log informations；With
Mathematics method is handled the elastic parameter of log data spatial sampling and selection, obtains elastic parameter phase
To the conditional probability of log data.
In step 4, the log prediction object function based on bayesian theory, curve prediction are set up
Object function is maximum a posteriori probability distribution of the log data under known elasticity Parameter Conditions, and formula is：
[w1,w2,w3]=argMax { P ([w1,w2,w3]j|[E1,E2])} (1)
Wherein, [w1,w2,w3] it is log data, [E1,E2] it is elastic parameter, j is the class of log data
Not Shuo, P () be probability density function,
Due to P ([E1,E2]) be certain area elastic parameter prior distribution, calculated according to Bayes's classification
The class conditional independence assumption of method, (1) formula is written as
[w1,w2,w3]=argMax { P (E1|[w1,w2,w3]j)×P(E2|[w1,w2,w3]j)×P([w1,w2,w3]j)} (2)
(2) formula is final curves predictive equation, wherein P ([w1,w2,w3]j) it is log data prior distribution,
P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jFor conditional probability of the elastic parameter to log data；
The bar of log data prior distribution, elastic parameter to log data in calculated curve prediction object function
Part probability, statistical analysis well-log information obtains prior distribution the P ([w of log data to be predicted1,w2,w3]j),
Using Monte Carlo sampling method carry out log data spatial distribution discrete sampling, obtain log data from
Dissipate stochastical sampling [w1i,w2i,w3i], horizontal extrapolation of the log data from well point to space is completed, according to well logging
Data and elastic parameter combined sampling space ([w1i,w2i,w3i],[E1i,E2i]), can be with through mathematical statistics
Ask for conditional probability P (E of the elastic parameter to log data1|[w1,w2,w3]j, P (E2|[w1,w2,w3]j,
By P ([w1,w2,w3]j)、P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jSubstitute into equation (2), meter
The maximum a posteriori probability distribution that log data belongs to a certain class is calculated, the numerical value of target logging trace is obtained.
Log Forecasting Methodology before practical brill in the present invention, based on Bayes's classification and Monte Carlo
Sampling techniques, set up the object function on curve prediction, using Bayes's classification based on bayesian theory
Algorithm for Solving log data Posterior probability distribution, it is ensured that method accuracy and effective performance solve conventional logging
The problem of curve prediction method is depended on unduly to well-log information, it is accurately bent by elastic parameter prediction well logging
Line, reduces the risk of exploration, reduces cost.
Brief description of the drawings
Fig. 1 for the present invention practical brill before log Forecasting Methodology a specific embodiment flow
Figure.
Embodiment
For enable the present invention above and other objects, features and advantages become apparent, it is cited below particularly go out
Preferred embodiment, and coordinate shown in accompanying drawing, it is described in detail below.
As shown in figure 1, flow charts of the Fig. 1 for log Forecasting Methodology before the practical brill of the present invention.
In step 101, log data is counted, obtains treating the accurate prior probability model of inverting log data.
Generally, if prediction target is physical property curve, its prior probability model is distributed for multivariate Gaussian；If prediction
Target is lithology curve, and its prior probability model is Markov chain model.
Due to influenceing the non-formation factor of log a lot, such as hole diameter, mud density and salinity, mud
Cake, borehole wall roughness, invaded zone, strata temperature and pressure, country rock and instrument external diameter, gap
Etc., in order to obtain preferable data process effects, it is necessary to carry out environmental correction to log.In length
During the exploration and development of phase, the log of all wells is it is difficult to ensure that be with same type of instrument, phase
With scale merit device and unified method measure, therefore certainly existed between the log data of each well with
Error based on scaling factors.Therefore, to log data used in reservoir description except the necessary ring of progress
Border influence correction is outer, it is necessary to which log is standardized.
In step 102, a series of elastic parameter is obtained by Earthquake Resilient inverting, such as：Velocity of longitudinal wave,
Young's modulus, Poisson's ratio, modulus of shearing etc., each elastic parameter have different rock physicses meanings.Should
This preferably goes out appropriate mutual incoherent different elastic parameters according to the log type to be predicted
Curve prediction is participated in, that is, to ensure that the class conditional independence assumption of Bayesian Classification Arithmetic is set up.
In step 103, log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method,
Horizontal extrapolation of the log data from well point to space is completed, log data spatial sampling result is obtained.To surveying
Well data sample data and the elastic parameter chosen carry out mathematical statistics, design conditions probability.It is special using covering
The Caro method of sampling carries out log data spatial distribution discrete sampling, completes log data from well point to space
Horizontal extrapolation, the sample characteristics finally given must be basically identical with well point, and information should be more
It is abundant, the geological conditions not embodied comprising more well-log informations.It is empty to log data with mathematics method
Between sampling and choose elastic parameter handled, obtain elastic parameter general with respect to the condition of log data
Rate.
In step 104, with reference to prior distribution, conditional probability, build using log data as class object
Bayes classifier, calculates the posterior probability based on Bayesian Classification Arithmetic, that is, predicts log data space
Distribution.
Log based on bayesian theory predicts the foundation of object function.Based on Bayesian Classification Arithmetic
Log Forecasting Methodology, be the base in known object (log data) prior probability distribution to be sorted
The Posterior probability distribution that the object belongs to a certain class is solved on plinth, when posterior probability takes maximum, institute is right
The classification answered is exactly the classification belonging to the object.The object function of curve prediction is log data in known bullet
Property Parameter Conditions under maximum a posteriori probability distribution, the object function of such curve prediction is represented by
[w1,w2,w3]=argMax { P ([w1,w2,w3]j|[E1,E2])} (1)
Wherein, [w1,w2,w3] it is log data (such as interval transit time, density, natural gamma)；[E1,E2]
For elastic parameter (such as velocity of longitudinal wave, Poisson's ratio)；J is the classification number of log data；P () is general
Rate density function.
Due to P ([E1,E2]) be certain area elastic parameter prior distribution (constant), according to Bayes
The class conditional independence assumption of sorting algorithm, (1) formula can be written as
[w1,w2,w3]=argMax { P (E1|[w1,w2,w3]j)×P(E2|[w1,w2,w3]j)×P([w1,w2,w3]j)} (2)
(2) formula is final curves predictive equation.Wherein P ([w1,w2,w3]j) it is log data prior distribution；
P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jFor conditional probability of the elastic parameter to log data.
Log data prior distribution, elastic parameter are general to the condition of log data in curve prediction object function
The calculating of rate.
Statistical analysis well-log information can obtain prior distribution the P ([w of log data to be predicted1,w2,w3]j)。
Log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method, log data is obtained
Discrete Stochastic sampling [w1i,w2i,w3i], complete horizontal extrapolation of the log data from well point to space.According to
Log data and elastic parameter combined sampling space ([w1i,w2i,w3i],[E1i,E2i]), through mathematical statistics,
Conditional probability P (E of the elastic parameter to log data can be asked for1|[w1,w2,w3]j,
P(E2|[w1,w2,w3]j。
By P ([w1,w2,w3]j)、P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jSubstitute into equation (2), meter
The maximum a posteriori probability distribution that log data belongs to a certain class is calculated, the numerical value of target logging trace is obtained.
Claims (7)
1. log Forecasting Methodology before practical brill, it is characterised in that log is predicted before the practical brill
Method includes：
Step 1, log data is counted, obtains treating the accurate prior probability model of inverting log data；
Step 2, a series of elastic parameter is obtained by Earthquake Resilient inverting；
Step 3, log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method, completed
Horizontal extrapolation of the log data from well point to space, obtains log data spatial sampling result；
Step 4, with reference to prior distribution, conditional probability, the pattra leaves using log data as class object is built
This grader, calculates the posterior probability based on Bayesian Classification Arithmetic, that is, predicts log data spatial distribution.
2. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 1, if prediction target is physical property curve, its prior probability model is distributed for multivariate Gaussian；If pre-
Survey target is lithology curve, and its prior probability model is Markov chain model.
3. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 1, the correction method of surroundings effecting is carried out to log data used in reservoir description, and log is entered
Row standardization.
4. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 2, obtained elastic parameter includes velocity of longitudinal wave, Young's modulus, Poisson's ratio, modulus of shearing, respectively
Elastic parameter has different rock physicses meanings.
5. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 2, according to the log type to be predicted, preferably go out appropriate mutual incoherent different bullets
Property parameter participate in curve prediction, it is ensured that the class conditional independence assumption of Bayesian Classification Arithmetic is set up.
6. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 3, mathematical statistics is carried out to log data sample data and the elastic parameter chosen, design conditions are general
Rate；Log data spatial distribution discrete sampling is carried out using Monte Carlo sampling method, log data is completed
Horizontal extrapolation from well point to space, the sample characteristics finally given must be basically identical with well point, and
Information should more be enriched, the geological conditions not embodied comprising more well-log informations；With mathematics method
Elastic parameter to log data spatial sampling and selection is handled, and obtains the relative number of logging well of elastic parameter
According to conditional probability.
7. log Forecasting Methodology before practical brill according to claim 1, it is characterised in that in step
In rapid 4, the log prediction object function based on bayesian theory, curve prediction object function are set up
The maximum a posteriori probability for being log data under known elasticity Parameter Conditions distribution, formula is：
[w1,w2,w3]=argMax { P ([w1,w2,w3]j|[E1,E2])} (1)
Wherein, [w1,w2,w3] it is log data, [E1,E2] it is elastic parameter, j is the class of log data
Not Shuo, P () be probability density function,
Due to P ([E1,E2]) be certain area elastic parameter prior distribution, calculated according to Bayes's classification
The class conditional independence assumption of method, (1) formula is written as
[w1,w2,w3]=argMax { P (E1|[w1,w2,w3]j)×P(E2|[w1,w2,w3]j)×P([w1,w2,w3]j)} (2)
(2) formula is final curves predictive equation, wherein P ([w1,w2,w3]j) it is log data prior distribution,
P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jFor conditional probability of the elastic parameter to log data；
The bar of log data prior distribution, elastic parameter to log data in calculated curve prediction object function
Part probability, statistical analysis well-log information obtains prior distribution the P ([w of log data to be predicted1,w2,w3]j),
Using Monte Carlo sampling method carry out log data spatial distribution discrete sampling, obtain log data from
Dissipate stochastical sampling [w1i,w2i,w3i], horizontal extrapolation of the log data from well point to space is completed, according to well logging
Data and elastic parameter combined sampling space ([w1i,w2i,w3i],[E1i,E2i]), can be with through mathematical statistics
Ask for conditional probability P (E of the elastic parameter to log data1|[w1,w2,w3]j, P (E2|[w1,w2,w3]j,
By P ([w1,w2,w3]j)、P(E1|[w1,w2,w3]j, P (E2|[w1,w2,w3]jEquation (2) is substituted into, calculates and surveys
Well data belong to the maximum a posteriori probability distribution of a certain class, obtain the numerical value of target logging trace.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610191318.1A CN107290800A (en) | 2016-03-30 | 2016-03-30 | Log Forecasting Methodology before practical brill |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610191318.1A CN107290800A (en) | 2016-03-30 | 2016-03-30 | Log Forecasting Methodology before practical brill |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107290800A true CN107290800A (en) | 2017-10-24 |
Family
ID=60087923
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610191318.1A Pending CN107290800A (en) | 2016-03-30 | 2016-03-30 | Log Forecasting Methodology before practical brill |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107290800A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109614584A (en) * | 2018-11-16 | 2019-04-12 | 中国科学院计算技术研究所 | A kind of method for reconstructing of resource log data |
| 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 |
| CN110442834A (en) * | 2019-06-27 | 2019-11-12 | 中国石油天然气股份有限公司 | Method and device for constructing logging curve, computer equipment and readable storage medium |
| CN112132336B (en) * | 2020-09-22 | 2024-02-20 | 南京创蓝科技有限公司 | Quarterly prediction method for PM2.5 concentration |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102183788A (en) * | 2011-05-13 | 2011-09-14 | 上海石油天然气有限公司 | Method for synthesizing transverse-wave well-logging curve under stratum condition |
| CN103485768A (en) * | 2012-06-13 | 2014-01-01 | 中国石油天然气集团公司 | Method for forming acoustic logging curve |
| CN104516017A (en) * | 2013-09-29 | 2015-04-15 | 中国石油化工股份有限公司 | Carbonate rock physical parameter seismic inversion method |
-
2016
- 2016-03-30 CN CN201610191318.1A patent/CN107290800A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102183788A (en) * | 2011-05-13 | 2011-09-14 | 上海石油天然气有限公司 | Method for synthesizing transverse-wave well-logging curve under stratum condition |
| CN103485768A (en) * | 2012-06-13 | 2014-01-01 | 中国石油天然气集团公司 | Method for forming acoustic logging curve |
| CN104516017A (en) * | 2013-09-29 | 2015-04-15 | 中国石油化工股份有限公司 | Carbonate rock physical parameter seismic inversion method |
Non-Patent Citations (6)
| Title |
|---|
| MARIT ULVMOEN ET AL.: "Improved resolution in Bayesian lithology/fluid inversion from prestack seismic data and well observations: Part 1-Methodology", 《GEOPHYSICS》 * |
| MIGUEL BOSCH ET AL.: "Petrophysical seismic inversion conditioned to well-log data: Methods and application to a gas reservoir", 《GEOPHYSICS》 * |
| 印兴耀等: "基于弹性阻抗的储层物性参数预测方法", 《地球物理学报》 * |
| 王丽萍等: "塔中奥陶系碳酸盐岩缝洞型储层贝叶斯叠前反演预测研究", 《石油物探》 * |
| 胡华锋等: "基于贝叶斯分类的储层物性参数联合反演方法", 《石油物探》 * |
| 赵军龙: "《测井资料处理与解释》", 31 January 2012, 北京：石油工业出版社 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109614584A (en) * | 2018-11-16 | 2019-04-12 | 中国科学院计算技术研究所 | A kind of method for reconstructing of resource log data |
| CN109614584B (en) * | 2018-11-16 | 2020-10-27 | 中国科学院计算技术研究所 | Reconstruction method of resource logging data |
| 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 |
| CN110442834A (en) * | 2019-06-27 | 2019-11-12 | 中国石油天然气股份有限公司 | Method and device for constructing logging curve, computer equipment and readable storage medium |
| CN110442834B (en) * | 2019-06-27 | 2023-12-26 | 中国石油天然气股份有限公司 | Method, device, computer equipment and readable storage medium for constructing logging curve |
| CN112132336B (en) * | 2020-09-22 | 2024-02-20 | 南京创蓝科技有限公司 | Quarterly prediction method for PM2.5 concentration |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Caers | Modeling uncertainty in the earth sciences | |
| Evans et al. | Groundwater Level Mapping Tool: An open source web application for assessing groundwater sustainability | |
| Ortiz et al. | Indicator simulation accounting for multiple-point statistics | |
| Zhou et al. | A pattern‐search‐based inverse method | |
| CN108596780B (en) | A kind of Reservoir type division methods and system based on multiple information | |
| Bodin et al. | Probabilistic surface reconstruction from multiple data sets: An example for the Australian Moho | |
| Shi et al. | Brittleness index prediction in shale gas reservoirs based on efficient network models | |
| He et al. | Assessing hydrological model predictive uncertainty using stochastically generated geological models | |
| CN103088803B (en) | Method for identifying collapse factors caused by karst collapse | |
| CN107290800A (en) | Log Forecasting Methodology before practical brill | |
| Chen et al. | Global-search distributed-gauss-newton optimization method and its integration with the randomized-maximum-likelihood method for uncertainty quantification of reservoir performance | |
| Wong et al. | Multiple permeability predictions using an observational learning algorithm | |
| CN113610945A (en) | Ground stress curve prediction method based on hybrid neural network | |
| Dutta et al. | Value of information analysis for subsurface energy resources applications | |
| Obergfell et al. | Estimation of average diffuse aquifer recharge using time series modeling of groundwater heads | |
| Chilton et al. | Uncovering the Controls on Fluvial Bedrock Erodibility and Knickpoint Expression: A High‐Resolution Comparison of Bedrock Properties Between Knickpoints and Non‐Knickpoint Reaches | |
| Pham et al. | Missing sonic log prediction using convolutional long short-term memory | |
| Jiang et al. | History matching for geological carbon storage using data-space inversion with spatio-temporal data parameterization | |
| Deswarte et al. | Sequential model aggregation for production forecasting | |
| Sarkheil et al. | The fracture network modeling in naturally fractured reservoirs using artificial neural network based on image loges and core measurements | |
| Jaquet et al. | Estimation of volcanic hazards using geostatistical models | |
| US11156740B2 (en) | Electrofacies determination | |
| CN106547028A (en) | The method and apparatus of prediction shale reservoir TOC | |
| Zhang et al. | A new zonation algorithm with parameter estimation using hydraulic head and subsidence observations | |
| Rawlinson et al. | Gaussian process modeling of well logs |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20171024 |