CN110009150B - The compact reservoir physical parameter intelligent Forecasting of data-driven - Google Patents
The compact reservoir physical parameter intelligent Forecasting of data-driven Download PDFInfo
- Publication number
- CN110009150B CN110009150B CN201910261627.5A CN201910261627A CN110009150B CN 110009150 B CN110009150 B CN 110009150B CN 201910261627 A CN201910261627 A CN 201910261627A CN 110009150 B CN110009150 B CN 110009150B
- Authority
- CN
- China
- Prior art keywords
- data
- network model
- physical parameter
- machine learning
- reservoir physical
- 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.)
- Active
Links
- 238000010801 machine learning Methods 0.000 claims abstract description 45
- 238000000034 method Methods 0.000 claims abstract description 42
- 230000008569 process Effects 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 17
- 230000000644 propagated effect Effects 0.000 claims abstract description 11
- 238000003860 storage Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 238000006467 substitution reaction Methods 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 12
- 238000004422 calculation algorithm Methods 0.000 description 10
- 230000015572 biosynthetic process Effects 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 6
- 238000011161 development Methods 0.000 description 3
- 230000000704 physical effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000035699 permeability Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000011148 porous material Substances 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Animal Husbandry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Agronomy & Crop Science (AREA)
- Mining & Mineral Resources (AREA)
- General Health & Medical Sciences (AREA)
- Marine Sciences & Fisheries (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
Present disclose provides a kind of compact reservoir physical parameter intelligent Forecastings of data-driven, construct learning sample collection using 3-D seismics Wave data and well-log information, pre-process to learning sample collection;The machine learning network model that building convolution is mixed with inner product operator, uses convolution operation to pretreated 3-D seismics Wave data, inner product operation is used to additional space position, along network propagated forward;The training that machine learning network model is carried out using learning sample collection, until the network model for obtaining meeting error requirements;Centered on sampled point, multiple data are acquired in a plurality of directions as sampled data, using sampled data as the input variable for the machine learning network model trained, compact reservoir physical parameter is predicted in realization.Disclosure substitution relies on the inversion method of empirical model, realizes the compact reservoir physical parameter intelligent predicting of data-driven, solves the problems, such as small-sample learning between well logging-earthquake.
Description
Technical field
This disclosure relates to the compact reservoir physical parameter of Exploration of Oil And Gas technical field more particularly to a kind of data-driven
Intelligent Forecasting.
Background technique
Only there is provided background technical informations relevant to the disclosure for the statement of this part, it is not necessary to so constitute first skill
Art.
With stepping up for oil-gas exploration and development degree, compact reservoir has become the emphasis of oil-gas exploration.With conventional storage
Layer is different, and the pore structure of compact reservoir is more complicated, heterogeneity is stronger, needs more, finer reservoir characterization parameter.
High-quality lithology, effecive porosity, stratum brittleness, permeability etc. are all important the physical parameter of compact reservoir, these parameters
Accurate Prediction is distributed predicting oil/gas and oil-gas exploration and development is instructed to be of great significance.
The physical property property on stratum is implied in seismic wave field, but multiple due to existing between formation physical property and seismic wave characteristic
Miscellaneous response mechanism, the two are not one-to-one relationship, can not establish determining forward model.And seismic inversion algorithms are necessary
Based on forward model, do not determine that the inverting of forward model can not carry out.Therefore reservoir physical parameter prediction at present
It is to obtain formation velocity using the empirical model between formation physical parameters and speed by inversion method, then utilize mostly
Speed is converted reservoir physical parameter by empirical model.But understand according to inventor, this method is a kind of indirect inverting in fact
Method has error by the formation velocity that actual seismic data inversion obtains, then by there is the formation velocity parameter of error to turn
There is also certain error, the cumulative accuracy of forecast that will affect reservoir parameter of error for the reservoir physical parameter got in return.
A research hotspot of the machine learning as the field of data mining, is that mined information establishes prediction model from data
Effective ways, Nonlinear Mapping relationship can be established between input and output, and do not need determining forward model.
Machine learning is used for reservoir parameter forecast, the limitation of various assumed conditions and complicated forward modeling theory is got rid of, can be directed to
Different work area geologic settings establish prediction model respectively.Classical convolution Learning Algorithm is towards large sample, big number
According to learning method, need the learning sample of magnanimity, and due to economic cost, exploration and development degree and skill in reservoir parameter forecast
Art reason etc., the well shake data that can be used for machine learning are often limited, are a small sample problems, are unsatisfactory for machine learning institute
The big data condition needed.
Summary of the invention
The disclosure to solve the above-mentioned problems, proposes a kind of compact reservoir physical parameter intelligent predicting side of data-driven
Method, the disclosure are utilized for small-sample learning problem between empirical model and well logging-earthquake is relied in reservoir properties prediction
The machine learning algorithm of convolution and inner product hybrid operator, substitution rely on the inversion method of empirical model, realize data-driven
Compact reservoir physical parameter intelligent predicting.
According to some embodiments, the disclosure is adopted the following technical scheme that
A kind of compact reservoir physical parameter intelligent Forecasting of data-driven, comprising the following steps:
Learning sample collection is constructed using 3-D seismics Wave data and well-log information, learning sample collection is pre-processed;
The machine learning network model that building convolution is mixed with inner product operator, to pretreated 3-D seismics Wave data
Using convolution operation, inner product operation is used to additional space position, along network propagated forward;
The training that machine learning network model is carried out using learning sample collection, until the network mould for obtaining meeting error requirements
Type;
Centered on sampled point, multiple data are acquired in a plurality of directions as sampled data, using sampled data as instruction
The input variable for the machine learning network model practiced realizes the prediction to compact reservoir physical parameter.
In above-mentioned technical proposal, using the machine learning algorithm of convolution and inner product hybrid operator, substitution relies on empirical model
Inversion method, be no longer indirect inverting, reduce actual seismic data inversion and obtain formation velocity, then converted by formation velocity
Obtain accumulated error brought by physical parameter.
Meanwhile using the machine learning algorithm of convolution and inner product hybrid operator, also solves existing convolution Neural Network Science
Data can only be shaken for limited well and not applicable problem towards large sample, large data objects by practising algorithm.
As one or more embodiment optinal plans, the building process of the learning sample collection includes:
Reservoir physical parameter, including effecive porosity, permeability are extracted from well-log information, are exported and are become as machine learning
Measure Y;
Machine learning input variable X is constructed using the corresponding seismic trace near well data of well-log information, it is pre- in order to reduce single-point
The multi-solution of survey respectively takes the total 3n data of n sampled point as defeated centered on sampled point along three orthogonal directions of seismic data
Enter variable.
As one or more embodiment optinal plans, the pretreated process includes at least normalized.
Further, normalized detailed process are as follows: the seismic amplitude value after normalization are as follows: the seismic amplitude of current sample
The difference of value and earthquake amplitude minimum in current sample, the ratio with the difference of seismic amplitude value minimum and maximum in current sample
Value.
As one or more embodiment optinal plans, the machine learning network model that convolution is mixed with inner product operator is constructed
Detailed process include:
Determine machine learning network model basic parameter;
Convolution operation is carried out to normalized seismic amplitude data, and along network propagated forward;
Inner product operation is carried out to spatial position corresponding to sampled point, and along network propagated forward:
The output of two propagated forwards is connected entirely, and carries out inner product operation and obtains the output of network model.
Further, machine learning network model basic parameter specifically includes convolution operation convolution kernel size, the network number of plies
With every node layer number, the inner product operation network number of plies and every node layer number.
As one or more embodiment optinal plans, in the training process of network model, convolution and interior integrating are calculated
Error function between the machine learning network model output of son mixing and desired output learns net according to error amount computing machine
The renewal amount of network model parameter value, and optimization is updated to network model according to renewal amount.
As one or more embodiment optinal plans, in the training process of network model, until the value of error function
When less than setting value, training process terminates.
A kind of computer readable storage medium, wherein being stored with a plurality of instruction, described instruction is suitable for by terminal device
Reason device loads and executes a kind of compact reservoir physical parameter intelligent Forecasting of data-driven.
A kind of terminal device, including processor and computer readable storage medium, processor is for realizing each instruction;It calculates
Machine readable storage medium storing program for executing is suitable for being loaded by processor and being executed a kind of data and drives for storing a plurality of instruction, described instruction
Dynamic compact reservoir physical parameter intelligent Forecasting.
Compared with prior art, the disclosure has the beneficial effect that
The disclosure utilizes the machine learning algorithm of convolution and inner product hybrid operator, and substitution relies on the inverting side of empirical model
Method realizes the compact reservoir geophysical parameter prediction of data-driven, solves the problems, such as small-sample learning between well logging-earthquake.
The disclosure gets rid of experience by establishing mapping relations directly between seismic waveform data and reservoir physical parameter
The limitation of model is realized from empirical model inverting decision is relied on to the transformation for doing decision by data;Using convolution and inner product
The machine learning algorithm of hybrid operator adapts to the small-sample learning condition in reservoir parameter forecast.
The disclosure improves geophysical parameter prediction process by joined the spatial positional information of sample in input information
Stability and prediction result precision.
Detailed description of the invention
The Figure of description for constituting a part of this disclosure is used to provide further understanding of the disclosure, and the disclosure is shown
Meaning property embodiment and its explanation do not constitute the improper restriction to the disclosure for explaining the disclosure.
Fig. 1 is learning sample building schematic diagram.
Fig. 2 is the compact reservoir physical parameter intelligent Forecasting flow chart of data-driven.
Fig. 3 a- Fig. 3 b is machine learning convolution operation and inner product operation schematic diagram.
Fig. 4 a- Fig. 4 c is the reservoir effecive porosity parameter profile figure using the present embodiment prediction.
Specific embodiment:
The disclosure is described further with embodiment with reference to the accompanying drawing.
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the disclosure.Unless another
It indicates, all technical and scientific terms used herein has usual with disclosure person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the disclosure.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
For small-sample learning problem between empirical model and well logging-earthquake is relied in reservoir properties prediction, provide
The compact reservoir physical parameter intelligent Forecasting of data-driven.Using the machine learning algorithm of convolution and inner product hybrid operator,
Substitution relies on the inversion method of empirical model, realizes the compact reservoir physical parameter intelligent predicting of data-driven.
Detailed process includes:
(1) learning sample collection is constructed using 3-D seismics Wave data and well-log information;
(2) learning sample collection is normalized;
(3) convolution+inner product operator mixing machine learning network is constructed;
(4) training that machine learning network is carried out using learning sample collection, obtains the network model for meeting error requirements;
(5) the complete machine learning network of training is used for three-dimensional reservoir parameter prediction, completes compact reservoir physical parameter intelligence
It can prediction.
It is specific:
In step (1), learning sample collection is constructed using 3-D seismics Wave data and well data.
Specific step is as follows:
1. extracting reservoir physical parameter from well data as machine learning output variable Y;
2. machine learning input variable X is constructed using the corresponding seismic trace near well data of well data, centered on sampled point,
N sampled point is respectively taken along three orthogonal directions of seismic data, total 3n data are as input variable.
In step (2), input variable is concentrated to be normalized by formula (1) learning sample:
In formula, x is a seismic amplitude value of current sample, xmaxFor earthquake amplitude maximum in current sample, xminFor
Minimum earthquake amplitude in current sample,To normalize later seismic amplitude value corresponding to amplitude x.
In step (3), building convolution+inner product operator mixing machine learning network, to normalized seismic waveform data
Using convolution operation, inner product operation is used to spatial position corresponding to sampled point.Specific step is as follows:
1. inputting network parameter: convolution operation convolution kernel size C, network number of plies Hc, every node layer numberInner product operation net
Network layers number HI, every node layer number
2. carrying out convolution operation according to formula (2) to normalized seismic amplitude data, and along network propagated forward:
In formula,For the output valve of i-th of node in h layers, whIt (t) is h-1 layers to h layers convolution kernel.
3. carrying out inner product operation according to formula (3) to spatial position corresponding to sampled point, and along network propagated forward:
In formula,For the output valve of the i-th node in h layers,For i-th of node of h-1 layer to h j-th of node of layer
Connection weight,For i-th of node threshold values of h layer, f is excitation function f (x)=max (0, x).
4. the output by step 2., 3. is connected entirely, and is carried out inner product operation using formula (3) and obtained the output of network
In step (4), the training of network is carried out using learning sample collection, interative computation process is as follows:
1. calculating reality output in step (3) by formula (4)With the error function value between desired output y:
In formula, E is error amount, | | indicate vector field homoemorphism;
2. according to the renewal amount of error value E computing machine learning network parameter value:
In formulaFor the renewal amount of connection weight,For threshold value renewal amount, η is learning rate, ghFor inner product
The error propagation item of h layers of layer:
ξhError propagation item for h layers of convolutional layer:
3. the parameter renewal amount being 2. calculated using step, carries out the optimization of network parameter:
2. 3. 4. step is repeated, until error size meets termination condition: E < E0, complete network training process.
In step (5), the machine learning network for completing training process is used for the prediction of three-dimensional reservoir, completes compact reservoir
Geophysical parameter prediction.
Specific step is as follows:
1. constructing input sample collection according to method shown in Fig. 1 to entire target work area;
2. obtaining reservoir parameter forecast result according to step (3) using the machine learning network after training.
Fig. 1 is learning sample building schematic diagram, the corresponding seismic trace near well Wave data of log data as input variable,
Reservoir parameter is output variable in well.In order to reduce the multi-solution of single-point prediction, when constructing input variable, selection is with sampled point
Centered on, respectively taking the total 3n data of n sampled point along three orthogonal directions of seismic data is input variable.
Fig. 2 is the compact reservoir physical parameter intelligent Forecasting flow chart of data-driven.
Learning sample collection will be constructed using 3-D seismics Wave data and well data first, and to data in each sample set
Do normalized.Convolution operation is used to normalized seismic waveform data, spatial position corresponding to sampled point is used
Then two operation results are carried out inner product operation by inner product operation again, obtain network output result.By reality output and expectation
Output comparison modifies connection weight and threshold value if error is unsatisfactory for training termination condition, recalculates network output.It presses
The above process is iterated, and until meeting error termination condition, completes training process.Trained machine learning will finally be utilized
Network model is used for the reservoir parameter forecast of destination layer, obtains reservoir parameter forecast result.
Fig. 3 a is machine learning convolution operation schematic diagram, and previous node layer is connected by convolution operator with latter node layer
It connects, is hereby connection type in Tobe;Fig. 3 b is machine learning inner product operation schematic diagram, previous node layer and each node of later layer
It all interconnects, for a kind of connection type connected entirely.Extraction office is more advantageous to using convolution operation processing sample seismic data
The weight of portion's feature, realization is shared and local experiences, can effectively reduce parameter amount to be learned.Using inner product operator to space bit
It sets coordinate and carries out operation, the seismic data local feature then extracted with convolution operation is connect entirely, and it is special to be more advantageous to progress
The synthesis of sign.Using the machine learning algorithm of convolution and inner product hybrid operator, the small-sample learning in reservoir parameter forecast is adapted to
Condition.
Fig. 4 a is original seismic profile;Fig. 4 b is the prestack inversion predicting reservoir effecive porosity based on empirical model
Section;Fig. 4 c this method predicting reservoir effecive porosity section.Comparison diagram 4b and Fig. 4 c can be seen that the method for the present invention prediction knot
Fruit profile morphology is stronger to the resolution capability of thin layer naturally, longitudinal resolution is higher, while cross directional variations rule and earthquake change
Feature matches.After resolution ratio raising, physical property pinching point understands that thin reservoir is obviously embodied in result section, can be effectively anti-
Reflect the porosity geological information of underground.
As other embodiments, a kind of computer readable storage medium is also provided, wherein it is stored with a plurality of instruction, the finger
Enable the compact reservoir physical parameter for being suitable for being loaded by the processor of terminal device and executed a kind of data-driven intelligently pre-
Survey method.
A kind of terminal device, including processor and computer readable storage medium, processor is for realizing each instruction;It calculates
Machine readable storage medium storing program for executing is suitable for being loaded by processor and being executed a kind of data and drives for storing a plurality of instruction, described instruction
Dynamic compact reservoir physical parameter intelligent Forecasting.
It should be understood by those skilled in the art that, embodiment of the disclosure can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the disclosure
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the disclosure, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The disclosure is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present disclosure
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The foregoing is merely preferred embodiment of the present disclosure, are not limited to the disclosure, for the skill of this field
For art personnel, the disclosure can have various modifications and variations.It is all within the spirit and principle of the disclosure, it is made any to repair
Change, equivalent replacement, improvement etc., should be included within the protection scope of the disclosure.
Although above-mentioned be described in conjunction with specific embodiment of the attached drawing to the disclosure, model not is protected to the disclosure
The limitation enclosed, those skilled in the art should understand that, on the basis of the technical solution of the disclosure, those skilled in the art are not
Need to make the creative labor the various modifications or changes that can be made still within the protection scope of the disclosure.
Claims (7)
1. a kind of compact reservoir physical parameter intelligent Forecasting of data-driven, it is characterized in that: the following steps are included:
Learning sample collection is constructed using 3-D seismics Wave data and well-log information, learning sample collection is pre-processed;It is described
Pretreated process includes at least normalized, normalized detailed process are as follows: the seismic amplitude value after normalization are as follows: current
The difference of minimum earthquake amplitude, shakes with earthquake minimum and maximum in current sample in the seismic amplitude value of sample and current sample
The ratio of the difference of amplitude;
The machine learning network model that building convolution is mixed with inner product operator, uses pretreated 3-D seismics Wave data
Convolution operation uses inner product operation to additional space position, along network propagated forward;
The detailed process of machine learning network model that building convolution is mixed with inner product operator includes:
Determine machine learning network model basic parameter;Convolution operation is carried out to normalized seismic amplitude data, and along network
Propagated forward;Inner product operation is carried out to spatial position corresponding to sampled point, and along network propagated forward: by two propagated forwards
Output connected entirely, and carry out inner product operation and obtain the output of network model;
The training that machine learning network model is carried out using learning sample collection, until the network model for obtaining meeting error requirements;
Centered on sampled point, multiple data are acquired in a plurality of directions as sampled data, using sampled data as having trained
Machine learning network model input variable, realization compact reservoir physical parameter is predicted.
2. a kind of compact reservoir physical parameter intelligent Forecasting of data-driven as described in claim 1, it is characterized in that: institute
The building process for stating learning sample collection includes:
Reservoir physical parameter is extracted from well-log information as machine learning output variable Y;
Machine learning input variable X is constructed using the corresponding seismic trace near well data of well-log information, centered on sampled point, along ground
Shake three orthogonal directions of data respectively take the total 3n data of n sampled point as input variable.
3. a kind of compact reservoir physical parameter intelligent Forecasting of data-driven as described in claim 1, it is characterized in that: machine
Device learning network model basic parameter specifically includes convolution operation convolution kernel size, the network number of plies and every node layer number, inner product fortune
Calculate the network number of plies and every node layer number.
4. a kind of compact reservoir physical parameter intelligent Forecasting of data-driven as described in claim 1, it is characterized in that: In
In the training process of network model, calculate the machine learning network model that mix with inner product operator of convolution export and desired output it
Between error function, according to the renewal amount of error amount computing machine learning network model parameter value, and according to renewal amount to network
Model is updated optimization.
5. a kind of compact reservoir physical parameter intelligent Forecasting of data-driven as described in claim 1, it is characterized in that: In
In the training process of network model, when the value of error function is less than setting value, training process terminates.
6. a kind of computer readable storage medium, it is characterized in that: being wherein stored with a plurality of instruction, described instruction is suitable for being set by terminal
Standby processor load and perform claim requires a kind of compact reservoir physical parameter intelligence of data-driven described in any one of 1-5
It can prediction technique.
7. a kind of terminal device, it is characterized in that: including processor and computer readable storage medium, processor is for realizing each finger
It enables;Computer readable storage medium is suitable for for storing a plurality of instruction, described instruction by processor load and perform claim requirement
A kind of compact reservoir physical parameter intelligent Forecasting of data-driven described in any one of 1-5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910261627.5A CN110009150B (en) | 2019-04-02 | 2019-04-02 | The compact reservoir physical parameter intelligent Forecasting of data-driven |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910261627.5A CN110009150B (en) | 2019-04-02 | 2019-04-02 | The compact reservoir physical parameter intelligent Forecasting of data-driven |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110009150A CN110009150A (en) | 2019-07-12 |
CN110009150B true CN110009150B (en) | 2019-11-19 |
Family
ID=67169411
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910261627.5A Active CN110009150B (en) | 2019-04-02 | 2019-04-02 | The compact reservoir physical parameter intelligent Forecasting of data-driven |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110009150B (en) |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110515123B (en) * | 2019-08-22 | 2021-03-30 | 中国海洋大学 | Convolution neural network seismic logging joint inversion method based on small sample learning |
CN111210877B (en) * | 2020-01-03 | 2023-11-24 | 山东天岳先进科技股份有限公司 | Method and device for deducing physical parameters |
CN111274689B (en) * | 2020-01-16 | 2021-03-02 | 中国地质大学(北京) | Method and apparatus for finding non-unique solutions for reservoir property combinations in history fitting |
US11326450B2 (en) | 2020-06-11 | 2022-05-10 | China University Of Petroleum (Beijing) | Intelligent prediction method and apparatus for reservoir sensitivity |
CN111694855B (en) * | 2020-06-11 | 2021-03-16 | 中国石油大学(北京) | Intelligent prediction data processing method and device for reservoir sensitivity |
CN111950193B (en) * | 2020-07-15 | 2023-03-14 | 中海油田服务股份有限公司 | Modeling method and device of pore network model based on reservoir |
CN113011639A (en) * | 2021-03-04 | 2021-06-22 | 中国石油大学(华东) | Perforation well productivity prediction method and system based on machine learning |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106886023B (en) * | 2017-02-27 | 2019-04-02 | 中国人民解放军理工大学 | A kind of Radar Echo Extrapolation method based on dynamic convolutional neural networks |
CN108375808A (en) * | 2018-03-12 | 2018-08-07 | 南京恩瑞特实业有限公司 | Dense fog forecasting procedures of the NRIET based on machine learning |
CN108645994B (en) * | 2018-04-25 | 2020-01-17 | 中国石油大学(北京) | Geological random inversion method and device based on multipoint geostatistics |
-
2019
- 2019-04-02 CN CN201910261627.5A patent/CN110009150B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110009150A (en) | 2019-07-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110009150B (en) | The compact reservoir physical parameter intelligent Forecasting of data-driven | |
CN107122861A (en) | A kind of Forecast of Gas Emission method based on PCA PSO ELM | |
CN109800863A (en) | A kind of well-log facies recognition method based on fuzzy theory and neural network | |
CN113610945B (en) | Ground stress curve prediction method based on hybrid neural network | |
WO2017198556A1 (en) | Computer implemented method for generating a field development plan (fdp) for the exploitation of oil and gas reservoirs | |
CN103454677B (en) | Based on the earthquake data inversion method that population is combined with linear adder device | |
US11846175B2 (en) | Estimating reservoir production rates using machine learning models for wellbore operation control | |
CN108645994A (en) | A kind of geology stochastic inversion methods and device based on Multiple-Point Geostatistics | |
EP3987478B1 (en) | Field development planning based on deep reinforcement learning | |
US10380793B2 (en) | Geobody surface reconstruction | |
CN103984788A (en) | Automatic intelligent design and optimization system for anchor bolt support of coal tunnel | |
CN104422969A (en) | Method for reducing non-uniqueness of electromagnetic sounding inversion result | |
CN107783185A (en) | A kind of processing method and processing device of tomographic statics | |
CN107843920A (en) | A kind of cluster analysis optimizes population reservoir attribute Forecasting Methodology | |
CN106249290B (en) | A method of shallow velocity structure model is established using multi-stage data fusion | |
CN110837115A (en) | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir | |
CN111766635A (en) | Sand body communication degree analysis method and system | |
KR102155101B1 (en) | System and method for predicting ground layer information, and a recording medium having computer readable program for executing the method | |
CN105103196A (en) | Gridless simulation of fluvio-deltaic environment | |
CN107861149A (en) | Based on the prestack P-S wave velocity ratio analogy method under drive waveform | |
Alimoradi et al. | Comparison between the performance of four metaheuristic algorithms in training a multilayer perceptron machine for gold grade estimation | |
CN107102379A (en) | A kind of method that seat earth watery prediction is carried out based on many attribution inversions | |
CN111625901B (en) | Intelligent pressure coefficient curve generation method for wing profile | |
Chen et al. | A prediction method of ground motion for regions without available observation data (LGB-FS) and its application to both Yangbi and Maduo earthquakes in 2021 | |
Chaki et al. | Prediction of porosity and sand fraction from well log data using ANN and ANFIS: a comparative study |
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 |