CN116168224A - Machine learning lithology automatic identification method based on imaging gravel content - Google Patents
Machine learning lithology automatic identification method based on imaging gravel content Download PDFInfo
- Publication number
- CN116168224A CN116168224A CN202111396092.6A CN202111396092A CN116168224A CN 116168224 A CN116168224 A CN 116168224A CN 202111396092 A CN202111396092 A CN 202111396092A CN 116168224 A CN116168224 A CN 116168224A
- Authority
- CN
- China
- Prior art keywords
- lithofacies
- imaging
- logging
- gravel
- lithology
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/30—Assessment of water resources
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a machine learning lithofacies automatic identification method based on imaging gravel content, which comprises the following steps: step 1, performing core observation description on a coring well, and performing four-way relation analysis; step 2, determining conventional and imaging logging response modes of different lithologies of the gritty; step 3, finely describing gravel content and gravel diameter distribution through imaging image processing, and quantitatively dividing the lithofacies of the whole well section; step 4, screening an electrical layer with stable logging characteristics as a sample layer based on lithofacies determined by imaging logging; step 5, calibrating a conventional logging curve through imaging, and determining a sensitive curve; and 6, based on a machine learning algorithm, automatically identifying the lithology by taking the lithology of the sample layer as a supervision object. The machine learning lithology automatic identification method based on the imaging gravel content improves the accuracy and the efficiency of lithology identification of the gritty, and achieves the purposes of quantitative, fine, objective and high efficiency of lithology identification.
Description
Technical Field
The invention relates to the technical field of petroleum and natural gas exploration and development, in particular to a machine learning lithology automatic identification method based on imaging gravel content.
Background
The gravel sector with various causes of development of the steep slope zone of the land phase fracture sink lake basin has the characteristics of deep burial, large thickness, large reserve, large potential, low utilization degree and the like, and becomes an important potential growth point for increasing the storage and the production of the victory oil field at present. The conglomerate sectors are usually close-source rapid stacking, longitudinal multi-period stacking, plane zonal swing migration and complex deposition mode. The gravel rock body has stronger heterogeneity in the longitudinal direction and the transverse direction, the rock types are complex and various, and the rock phase changes rapidly; the conglomerate skeleton has great influence on well logging response and complex rock-electricity relation, and meanwhile, the conventional well logging evaluation method is low in accuracy of identifying the lithology of the conglomerate due to low resolution ratio and unobvious thin interlayer response, so that the parameter interpretation accuracy is low, the oil-water layer is difficult to accurately identify, and the conglomerate skeleton becomes one of the bottlenecks for restricting the effective utilization of the conglomerate oil reservoir. The lithofacies identification is the basis of the evaluation of the sandstone reservoir, and the accurate identification of the lithofacies has important significance for accurately solving physical parameters of the sandstone reservoir, effectively identifying the reservoir, and even grasping the deposition rule, the dominant reservoir spreading and other rules. Therefore, the invention discloses a machine learning lithofacies automatic identification method based on imaging gravel content calculation aiming at the problem of low lithofacies identification precision of a sandstone reservoir, which can effectively improve lithofacies identification precision and provides a reliable basis for the evaluation and research of the lithofacies reservoir logging.
The micro-resistivity scanning imaging logging is a response reflecting the resistivity change of the well wall, has high longitudinal resolution (0.2 in), has visual effect, and can visually reflect the internal structure of the gravel rock mass. The method can accurately analyze the lithology of the gritty, the change of the sediment grain sequence, the relative sizes of gravel particles and the like on an imaging chart, can reflect the lithology information of the well wall after the rock core is scaled, solves the problems of small information quantity and low longitudinal resolution (20-80 cm) of the conventional logging curve, and has the advantage of being more visual and reliable than the conventional logging data for carrying out reservoir characteristic description. However, at present, most of imaging logging of gritty is mainly performed by qualitative application, imaging data information is mined deeply at this time, static imaging data is used as a basis, through rock core scales, imaging images are utilized to display tone ranges to represent different lithofacies types, bright colors represent high resistance, dark colors represent low resistance, distribution areas of different components are used as lithofacies percentage content, and the gravel content is calculated to realize quantitative lithofacies division.
Along with the rapid development of artificial intelligence, a method for rapidly analyzing data, training and learning and realizing rapid and automatic lithology identification by taking big data as a basis and taking a machine learning algorithm as a means becomes a trend. Because of the high imaging logging cost, only 1-2 wells are imaged and logged in one block, and only part of well sections of the coring well are cored, the application of the whole-area lithofacies identification work still needs to be carried out on the conventional logging curve.
In application number: in the chinese patent application CN201910078669.5, a method and a system for intelligent identification of multi-well complex lithology based on logging data are related, the method first determines a target logging data file, performs format conversion and normalization preprocessing, then performs feature screening and/or feature combination expansion on logging curve data according to the known lithology of a whole-region key coring well in a coring well section to obtain logging curve data sensitive to lithology, then performs labeling calibration on the logging curve data sensitive to lithology response to form a sample database, and simultaneously forms the logging curve data of the whole-region unlabeled logging curve data into a database to be tested, and further utilizes the data of the sample database and combines a plurality of machine learning algorithms to perform machine learning training to automatically establish a plurality of lithology identification models.
In application number: in the chinese patent application CN201911190561.1, a method, an apparatus, a computer device and a storage medium for constructing a core saturation prediction model are related to obtaining a sample logging curve and a corresponding sample core saturation thereof, where the sample logging curve is obtained after an original logging curve is calibrated, so that accuracy of obtained sample data can be improved.
In application number: in the chinese patent application CN201910440723.6, a quantitative prediction method for land hydrocarbon source rock based on seismic inversion and machine learning is used to predict the spatial distribution and organic content of land hydrocarbon source rock in a certain area, the elastic attribute sensitive to the lithology distinction and organic content of the land sedimentary stratum sand and shale is first optimized, then the machine learning network of the mapping relationship of "elastic attribute-lithology" and "elastic attribute-organic content" is characterized in the training stage, finally the trained machine learning network is combined with the pre-stack elastic parameter inversion result of the pre-stack seismic data, so as to predict the spatial distribution and organic content of the hydrocarbon source rock.
The prior art is greatly different from the invention, the technical problem which is needed to be solved by the user is not solved, and the invention provides a novel machine learning lithofacies automatic identification method based on imaging gravel content.
Disclosure of Invention
The invention aims to provide an automatic identification method of a machine learning lithology based on imaging gravel content, which is used for calculating the gravel content on the basis of imaging logging image processing, training and learning through a machine learning algorithm and determining lithology type lithology discrimination of the lithology.
The aim of the invention can be achieved by the following technical measures: the automatic machine learning lithology identification method based on the imaging gravel content comprises the following steps:
step 1, performing core observation description on a coring well, and performing four-way relation analysis;
step 2, determining conventional and imaging logging response modes of different lithologies of the gritty;
step 3, finely describing gravel content and gravel diameter distribution through imaging image processing, and quantitatively dividing the lithofacies of the whole well section;
step 4, screening an electrical layer with stable logging characteristics as a sample layer based on lithofacies determined by imaging logging;
step 5, calibrating a conventional logging curve through imaging, and determining a sensitive curve;
and 6, based on a machine learning algorithm, automatically identifying the lithology by taking the lithology of the sample layer as a supervision object.
The aim of the invention can be achieved by the following technical measures:
in step 1, core observation is performed on a coring well with imaging logging, various lithology, gravel diameter size, longitudinal change and single-layer thickness are described, and lithology characteristics corresponding to obvious-change intervals in imaging and conventional logging curves are particularly focused.
In the step 1, because the system error exists between the drilling coring depth and the logging depth, the imaging image and the core scanning gamma are utilized to assist in core homing, so that the rock electricity corresponding relation is more reasonable; four-way relation analysis is carried out by combining logging and analysis tests.
In step 2, the core is imaged and well-logged, the lithology observed by the core is divided into a plurality of lithofacies which can be identified on the well-logging according to the characteristics of lithology and gravel size, the lithofacies are similar in lithology, the lithofacies are the lithology combinations similar in the well-logging response characteristics, the purpose is to consider the identifiable capacity of the well-logging response and convert the problem of serious heterogeneity into the problem of relative homogeneity, and then the conventional and imaging well-logging response modes of different lithofacies of the sandstone are determined.
In step 3, the imaging logging is the response of resistivity change of the well wall, the bright color represents high resistance, the dark color represents low resistance, the core scale imaging is used for determining 4 components of gravel, sandy, silty and argillaceous based on static imaging data, wherein the gravel of the component 1 is white spots and clusters, the sandy of the component 2 is bright color, dispersed or layered, the silty of the component 3 is dark color, dispersed, and the argillaceous of the component 4 is black layered.
And 3, depicting bright colors, namely gravel edges, by adjusting a tone threshold value, converting an imaging image into an image formed by different tone areas, wherein the different tones represent different lithofacies, determining the tone ranges of the different lithofacies through core calibration, further determining the content of different lithofacies types within a certain window length by utilizing superposition of the same tone areas, taking the distribution areas of the different components as the lithofacies percentage content, calculating the gravel content, and realizing quantitative division of the lithofacies of the whole well section.
In step 4, the sample layer is screened based on lithofacies determined by imaging logs, based on the resolution of the corresponding conventional log series.
In step 4, the screening principle is as follows: the lithofacies are relatively homogeneous, the single-layer thickness is more than or equal to 2m, and the characteristics of the corresponding conventional logging curve are stable.
In step 5, by using the methods of intersection graph technology and principal component analysis, a curve with high resolution capability on different lithofacies is selected as a sensitive curve by combining the vertical resolution and the transverse detection depth of different logging series.
In step 5, 5 sensitivity curves are determined, namely Compensated Neutron Logging (CNL), density logging (DEN), natural Gamma (GR), natural potential (SP) and micro potential (RN), respectively.
In step 6, the sample layer screened in step 4 is used as a supervision object, the sensitive curve logging value determined in step 5 is used as a training object, mathematical algorithms such as self-organizing feature mapping, BP neural network and support vector machine are adopted to carry out repeated training learning and classification on the basis of normalization processing, a model with high matching degree with the rock core and imaging rock is preferably selected as a machine learning model of the well, and the model is finally popularized and applied to other wells with full blocks, so that automatic identification of the lithology of the sandstone is realized.
According to the automatic identification method of the machine learning lithofacies based on the imaging gravel content, which is disclosed by the invention, the automatic identification method of the multi-scale and multi-parameter fusion lithofacies based on imaging logging image processing, gravel content quantitative calculation, logging sensitivity analysis, machine learning and the like is realized, the lithofacies sample layer based on imaging identification is used as a supervision object for training and learning, the sensitive characteristic value of a conventional logging curve is extracted, and a machine learning lithofacies identification model is established, so that the automatic identification of the lithofacies is realized.
Drawings
FIG. 1 is a flow chart of one implementation of the machine-learned lithofacies automatic identification method based on imaged gravel content of the present invention;
FIG. 2 is a flow chart of one embodiment of a machine-learned lithofacies automatic identification method based on imaged gravel content of the present invention;
FIG. 3 is a flow chart of automatic recognition of machine-learned lithofacies in accordance with one embodiment of the present invention;
FIG. 4 is a flow chart of automatic recognition of machine-learned lithofacies in accordance with another embodiment of the present invention.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present invention. As used herein, the singular forms also are intended to include the plural forms unless the context clearly indicates otherwise, and furthermore, it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, and/or combinations thereof.
As shown in fig. 1, fig. 1 is a flow chart of the machine learning lithofacies automatic identification method based on imaging gravel content of the present invention. The machine learning lithofacies automatic identification method based on imaging gravel content comprises the following steps:
step 1, performing core observation description on a coring well, and performing four-way relation analysis;
step 2, core scale imaging logging, and determining conventional and imaging logging response modes of different lithologies of the gritty;
step 3, finely describing gravel content and gravel diameter distribution through imaging image processing, and quantitatively dividing the lithofacies of the whole well section;
step 4, screening an electrical layer with stable logging characteristics as a sample layer based on lithofacies determined by imaging logging;
step 5, calibrating a conventional logging curve through imaging, and determining a sensitive curve;
and 6, based on a machine learning algorithm, automatically identifying the lithology by taking the lithology of the sample layer as a supervision object.
The following are several specific examples of the application of the present invention.
Example 1:
in the embodiment 1 to which the present invention is applied, as shown in fig. 2, fig. 2 is a flowchart of one embodiment of the machine learning lithofacies automatic identification method based on imaging gravel content of the present invention.
Step 1, carrying out core observation on a coring well with imaging logging, describing various lithology, gravel diameter size, longitudinal change, single-layer thickness and other conditions as much as possible, and particularly focusing on lithology characteristics corresponding to a layer section with obvious change in imaging and conventional logging curves; because the system error exists between the drilling coring depth and the logging depth, the imaging image and the core scanning gamma are utilized to assist in core homing, so that the rock electricity corresponding relation is more reasonable; and carrying out four-way relation analysis by combining logging, analysis and test and other data.
Step 2, performing scale imaging logging on the core, dividing the lithology observed by the core into a plurality of lithofacies which can be identified on logging according to characteristics such as lithology, gravel size and the like according to logging curve response characteristics, wherein the lithofacies are similar in lithology, and the logging response characteristics are similar in lithology combination, so that the purposes of considering the identifiable capacity of logging response and converting serious problems of heterogeneity into problems of relative homogeneity are achieved, and then determining conventional and imaging logging response modes of different lithofacies of the sandy rock are achieved.
And 3, imaging logging is response to change of resistivity of a well wall, bright color represents high resistance, dark color represents low resistance, core scale imaging is utilized to determine 4 components such as gravel, sandy, silty and argillaceous based on static imaging data, wherein the gravel of the component 1 is white spots and lumps, the sandy of the component 2 is bright color and dispersed or layered, the silty of the component 3 is dark color and dispersed, and the argillaceous of the component 4 is black layered. The method comprises the steps of depicting a bright color (gravel) edge by adjusting a tone threshold value, converting an imaging image into an image formed by different tone areas, determining the tone range of different lithofacies by core calibration, further determining the content of different lithofacies types within a certain window length by utilizing the superposition of the same tone areas, calculating the gravel content by taking the distribution areas of different components as the lithofacies percentage content, and realizing the quantitative division of the lithofacies of the whole well section.
And 4, screening a sample layer on the basis of lithofacies determined by imaging logging and on the basis of the resolution capability of a corresponding conventional logging series, wherein the screening principle is as follows: the lithofacies are relatively homogeneous, the single-layer thickness is more than or equal to 2m, and the characteristics of the corresponding conventional logging curve are stable.
And 5, selecting curves with high resolution capability on different lithofacies as sensitive curves by combining the methods of cross-plot technology, principal component analysis and the like with different logging series, wherein 5 sensitive curves are determined in the example and are respectively Compensated Neutron Logging (CNL), density logging (DEN), natural Gamma (GR), natural potential (SP) and micro potential (RN).
And 6, taking the sample layer screened in the step 4 as a supervision object, taking the sensitive curve logging value determined in the step 5 as a training object, and carrying out repeated training learning and classification by adopting mathematical algorithms such as self-organizing feature mapping, BP neural network, support vector machine and the like on the basis of normalization processing, wherein a model with high matching degree with a rock core and imaging rock is preferably selected as a machine learning model of the well, and finally, the model is popularized and applied to other wells in a whole block to realize automatic identification of the lithology of the sandstone, and 5 lithology are identified in the example: mudstone, sandstone, gravel-containing sandstone, gravel-like sandstone, and conglomerate. (see FIG. 3).
Example 2:
in the embodiment 2 to which the present invention is applied, as shown in fig. 2, fig. 2 is a flowchart of one embodiment of the machine learning lithofacies automatic identification method based on imaging gravel content of the present invention.
Step 1, carrying out core observation on a coring well with imaging logging, describing various lithology, gravel diameter size, longitudinal change, single-layer thickness and other conditions as much as possible, and particularly focusing on lithology characteristics corresponding to a layer section with obvious change in imaging and conventional logging curves; because the system error exists between the drilling coring depth and the logging depth, the imaging image and the core scanning gamma are utilized to assist in core homing, so that the rock electricity corresponding relation is more reasonable; and carrying out four-way relation analysis by combining logging, analysis and test and other data.
Step 2, performing scale imaging logging on the core, dividing the lithology observed by the core into a plurality of lithofacies which can be identified on logging according to characteristics such as lithology, gravel size and the like according to logging curve response characteristics, wherein the lithofacies are similar in lithology, and the logging response characteristics are similar in lithology combination, so that the purposes of considering the identifiable capacity of logging response and converting serious problems of heterogeneity into problems of relative homogeneity are achieved, and then determining conventional and imaging logging response modes of different lithofacies of the sandy rock are achieved.
And 3, imaging logging is response to change of resistivity of a well wall, bright color represents high resistance, dark color represents low resistance, core scale imaging is utilized to determine 4 components such as gravel, sandy, silty and argillaceous based on static imaging data, wherein the gravel of the component 1 is white spots and lumps, the sandy of the component 2 is bright color and dispersed or layered, the silty of the component 3 is dark color and dispersed, and the argillaceous of the component 4 is black layered. The method comprises the steps of depicting a bright color (gravel) edge by adjusting a tone threshold value, converting an imaging image into an image formed by different tone areas, determining the tone range of different lithofacies by core calibration, further determining the content of different lithofacies types within a certain window length by utilizing the superposition of the same tone areas, calculating the gravel content by taking the distribution areas of different components as the lithofacies percentage content, and realizing the quantitative division of the lithofacies of the whole well section.
And 4, screening a sample layer on the basis of lithofacies determined by imaging logging and on the basis of the resolution capability of a corresponding conventional logging series, wherein the screening principle is as follows: the lithofacies are relatively homogeneous, the single-layer thickness is more than or equal to 2m, and the characteristics of the corresponding conventional logging curve are stable.
And 5, selecting curves with high resolution capability on different lithofacies as sensitive curves by combining the methods of cross-plot technology, principal component analysis and the like with different logging series, wherein 4 sensitive curves are determined in the example and are respectively Compensated Neutron Logging (CNL), density logging (DEN), natural Gamma (GR) and micro-potential (RN).
And 6, taking the sample layer screened in the step 4 as a supervision object, taking the sensitive curve logging value determined in the step 5 as a training object, and carrying out repeated training learning and classification by adopting mathematical algorithms such as self-organizing feature mapping, BP neural network, support vector machine and the like on the basis of normalization processing, preferably taking a model with high matching degree with a rock core and imaging rock as a machine learning model of the well, and finally popularizing and applying the model to other wells with all blocks to realize automatic identification of the lithology of the sandstone. (see FIG. 4).
Example 3:
in embodiment 3 to which the present invention is applied, similarly to embodiments 1 and 2, the implementation flow is a flow chart (fig. 2) of one embodiment of a machine learning lithofacies automatic identification method based on imaging gravel content.
Step 1, performing core observation description on a coring well, and performing four-way relation analysis;
step 2, core scale imaging logging, and determining conventional and imaging logging response modes of different lithologies of the gritty;
step 3, finely describing gravel content and gravel diameter distribution through imaging image processing, and quantitatively dividing the lithofacies of the whole well section;
step 4, screening an electrical layer with stable logging characteristics as a sample layer based on lithofacies determined by imaging logging;
step 5, calibrating a conventional logging curve through imaging to determine a sensitivity curve, wherein the acoustic time difference curve in this example 3 is more sensitive to lithology recognition, so that 5 sensitivity curves are determined in this example, namely, compensated Neutron Logging (CNL), density logging (DEN), natural Gamma (GR), acoustic time difference (AC) and micro-potential (RN), respectively.
And 6, taking the sample layer screened in the step 4 as a supervision object, taking the sensitive curve logging value determined in the step 5 as a training object, and carrying out repeated training learning and classification by adopting mathematical algorithms such as self-organizing feature mapping, BP neural network, support vector machine and the like on the basis of normalization processing, preferably taking a model with high matching degree with a rock core and imaging rock as a machine learning model of the well, and finally popularizing and applying the model to other wells with all blocks to realize automatic identification of the lithology of the sandstone.
The invention discloses a machine learning lithology automatic identification method based on imaging gravel content calculation, which is applied to the technical field of oilfield exploration and development, and particularly relates to a multi-scale and multi-parameter fusion-based automatic continuous lithology identification method based on imaging logging image processing, gravel content quantitative calculation, logging sensitivity analysis, machine learning and the like. The method comprises six steps of four-dimensional relation analysis, imaging logging response mode determination, gravel content calculation, sample layer screening, sensitivity curve analysis, and automatic lithology recognition by machine learning. The invention fully utilizes the new technology and method of logging, deeply excavates the rock core, images and logs the well curve information, constructs a machine learning lithology recognition method of multi-scale and multi-series logging data fusion, improves the accuracy and efficiency of the lithology recognition of the conglomerate, and achieves the purposes of quantitative, fine, objective and high efficiency of lithology recognition.
Finally, it should be noted that: the foregoing description is only a preferred embodiment of the present invention, and is not intended to limit the present invention, but although the present invention has been described in detail with reference to the foregoing embodiment, it will be apparent to those skilled in the art that modifications may be made to the technical solution described in the foregoing embodiment, or equivalents may be substituted for some of the technical features thereof. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Other than the technical features described in the specification, all are known to those skilled in the art.
Claims (11)
1. The automatic recognition method of the machine learning lithofacies based on the imaging gravel content is characterized by comprising the following steps of:
step 1, performing core observation description on a coring well, and performing four-way relation analysis;
step 2, determining conventional and imaging logging response modes of different lithologies of the gritty;
step 3, finely describing gravel content and gravel diameter distribution through imaging image processing, and quantitatively dividing the lithofacies of the whole well section;
step 4, screening an electrical layer with stable logging characteristics as a sample layer based on lithofacies determined by imaging logging;
step 5, calibrating a conventional logging curve through imaging, and determining a sensitive curve;
and 6, based on a machine learning algorithm, automatically identifying the lithology by taking the lithology of the sample layer as a supervision object.
2. The automatic identification method of lithofacies based on machine learning of imaged gravel content according to claim 1, wherein in step 1, core observation is performed on a cored well with imaged logging, and various lithologies, gravel diameter sizes, longitudinal changes, single-layer thicknesses are described, and lithologic features corresponding to obvious intervals of changes in imaging and conventional logging curves are particularly focused.
3. The automatic recognition method of the machine learning lithofacies based on the imaging gravel content is characterized in that in the step 1, because systematic errors exist between the drilling coring depth and the logging depth, rock core homing is performed by using imaging images and core scanning gamma assistance, so that the rock electricity corresponding relation is more reasonable; four-way relation analysis is carried out by combining logging and analysis tests.
4. The automatic identification method of lithofacies based on machine learning of imaging gravel content according to claim 1, wherein in step 2, the lithology observed by the core is divided into several lithofacies which can be identified on the well according to the characteristics of lithology and gravel size according to the response characteristics of the well logging curve, the lithofacies are similar in lithology, the similar lithology combination of the response characteristics of the well logging is aimed at considering the identifiable capacity of the well logging response and converting the serious problem of non-uniformity into the problem of relative homogeneity, and then the conventional and imaging well logging response modes of different lithofacies of the sand are determined.
5. The automatic recognition method of machine learning lithofacies based on imaging gravel content according to claim 1, wherein in step 3, imaging logging is response to change of resistivity of a borehole wall, bright color represents high resistance, dark color represents low resistance, and based on static imaging data, 4 components of gravel, sandiness, silty and argillaceous are determined by using core scale imaging, wherein the component 1 gravel is white spots, clusters, the component 2 sandiness is bright color, dispersed or layered, the component 3 silty is dark color, dispersed, and the component 4 argillaceous is black layered.
6. The automatic recognition method of machine learning lithofacies based on imaging gravel content according to claim 5, wherein in step 3, bright colors, namely gravel edges, are depicted by adjusting tone threshold values, imaging images are converted into images composed of different tone areas, the different tones represent different lithofacies, tone ranges of the different lithofacies are determined by core calibration, further the content of different lithofacies types is determined within a certain window length by superposition of the same tone areas, the distribution areas of the different components serve as lithofacies percentage content, and the gravel content is calculated to realize quantitative division of lithofacies of all well sections.
7. The method of automatic identification of lithofacies based on machine learning of imaged gravel content of claim 1, wherein in step 4, sample layers are screened based on lithofacies determined by imaging logs based on the resolution of the corresponding conventional log series.
8. The method for automatically identifying machine-learned lithofacies based on imaged gravel content according to claim 7, wherein in step 4, the screening principle is: the lithofacies are relatively homogeneous, the single-layer thickness is more than or equal to 2m, and the characteristics of the corresponding conventional logging curve are stable.
9. The automatic recognition method of machine-learned lithofacies based on imaged gravel content according to claim 1, wherein in step 5, the methods are analyzed by cross-plot technique and principal component analysis, and the curves with high resolution to different lithofacies are selected as sensitive curves in combination with vertical resolution and lateral detection depth of different logging series.
10. The method for automatically identifying machine-learned lithofacies based on imaged gravel content according to claim 9, wherein in step 5, 5 sensitivity curves are determined, namely compensated neutron log CNL, density log DEN, natural gamma GR, natural potential SP and micro potential RN, respectively.
11. The automatic recognition method of machine learning lithofacies based on imaging gravel content according to claim 1, wherein in step 6, the sample layer screened in step 4 is used as a supervision object, the sensitive curve logging value determined in step 5 is used as a training object, and based on normalization treatment, mathematical algorithms such as self-organizing feature mapping, BP neural network and support vector machine are adopted to carry out repeated training learning and classification, preferably, a model with high matching degree with a rock core and imaging lithofacies is used as a machine learning model of the well, and finally popularized and applied to other wells of a whole block to realize automatic recognition of lithofacies of the lithofacies.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111396092.6A CN116168224A (en) | 2021-11-23 | 2021-11-23 | Machine learning lithology automatic identification method based on imaging gravel content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111396092.6A CN116168224A (en) | 2021-11-23 | 2021-11-23 | Machine learning lithology automatic identification method based on imaging gravel content |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116168224A true CN116168224A (en) | 2023-05-26 |
Family
ID=86416830
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111396092.6A Pending CN116168224A (en) | 2021-11-23 | 2021-11-23 | Machine learning lithology automatic identification method based on imaging gravel content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116168224A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116822971A (en) * | 2023-08-30 | 2023-09-29 | 长江大学武汉校区 | Well wall risk level prediction method |
| CN117609741A (en) * | 2024-01-23 | 2024-02-27 | 中国地质大学（武汉） | Shale oil reservoir thin interlayer logging identification method based on envelope curve algorithm |
| CN117951476A (en) * | 2024-01-15 | 2024-04-30 | 长江大学 | Construction method of lithology recognition model of shale oil reservoir and lithology recognition method |
-
2021
- 2021-11-23 CN CN202111396092.6A patent/CN116168224A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116822971A (en) * | 2023-08-30 | 2023-09-29 | 长江大学武汉校区 | Well wall risk level prediction method |
| CN116822971B (en) * | 2023-08-30 | 2023-11-14 | 长江大学武汉校区 | Well wall risk level prediction method |
| CN117951476A (en) * | 2024-01-15 | 2024-04-30 | 长江大学 | Construction method of lithology recognition model of shale oil reservoir and lithology recognition method |
| CN117609741A (en) * | 2024-01-23 | 2024-02-27 | 中国地质大学（武汉） | Shale oil reservoir thin interlayer logging identification method based on envelope curve algorithm |
| CN117609741B (en) * | 2024-01-23 | 2024-04-02 | 中国地质大学（武汉） | Shale oil reservoir thin interlayer logging identification method based on envelope curve algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Raeesi et al. | Classification and identification of hydrocarbon reservoir lithofacies and their heterogeneity using seismic attributes, logs data and artificial neural networks | |
| Alizadeh et al. | Artificial neural network modeling and cluster analysis for organic facies and burial history estimation using well log data: A case study of the South Pars Gas Field, Persian Gulf, Iran | |
| Alnahwi et al. | Mineralogical composition and total organic carbon quantification using x-ray fluorescence data from the Upper Cretaceous Eagle Ford Group in southern Texas | |
| CN116168224A (en) | Machine learning lithology automatic identification method based on imaging gravel content | |
| CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
| Yasin et al. | An integrated fracture parameter prediction and characterization method in deeply-buried carbonate reservoirs based on deep neural network | |
| Abdideh et al. | Cluster analysis of petrophysical and geological parameters for separating the electrofacies of a gas carbonate reservoir sequence | |
| Ali et al. | A novel machine learning approach for detecting outliers, rebuilding well logs, and enhancing reservoir characterization | |
| Zaitouny et al. | Fast automatic detection of geological boundaries from multivariate log data using recurrence | |
| Masroor et al. | A multiple-input deep residual convolutional neural network for reservoir permeability prediction | |
| Wang et al. | Improved permeability prediction based on the feature engineering of petrophysics and fuzzy logic analysis in low porosity–permeability reservoir | |
| Akinyokun et al. | Well log interpretation model for the determination of lithology and fluid | |
| Eftekhari et al. | Electrical facies of the Asmari Formation in the Mansouri oilfield, an application of multi-resolution graph-based and artificial neural network clustering methods | |
| CN109826623A (en) | Knowledge method is sentenced in a kind of geophysical log of tight sandstone reservoir stratification seam | |
| Eftekhari et al. | Employing Statistical Algorithms and Clustering Techniques to Assess Lithological Facies for Identifying Optimal Reservoir Rocks: A Case Study of the Mansouri Oilfields, SW Iran | |
| Sarkheil et al. | The fracture network modeling in naturally fractured reservoirs using artificial neural network based on image loges and core measurements | |
| CN110118994A (en) | A kind of nonmarine source rock quantitative forecasting technique based on seismic inversion and machine learning | |
| CN113484907B (en) | Method for predicting distribution on different types of reservoir planes | |
| Dong et al. | Fracture identification in shale reservoir using a deep learning method: Chang 7 reservoirs, Triassic Yanchang formation | |
| Misra et al. | Shallow neural networks and classification methods for approximating the subsurface in situ fluid-filled pore size distribution | |
| CN114482995A (en) | Fine determination method for argillaceous content of fine-grain sediment | |
| Katterbauer et al. | A Deep Learning Wag Injection Method for Co2 Recovery Optimization | |
| Hosseinzadeh et al. | Rock type based-estimation of pore throat size distribution in carbonate reservoirs using integrated analysis of well logs and seismic attributes | |
| Garcia et al. | Geological Insights from Porosity Analysis for Sustainable Development of Santos Basin’s Pre-Salt Carbonate Reservoir | |
| Zhou et al. | Predicting mineralogy of the salt-gypsum layer by drilling cuttings and conventional well logging data using a multilayer perceptron neural network (MLPNN): A case study in Kuqa Depression, Tarim Basin |
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 |