CN117390970B - Oil gas storage quantification method considering multi-element driving prediction - Google Patents

Oil gas storage quantification method considering multi-element driving prediction Download PDF

Info

Publication number
CN117390970B
CN117390970B CN202311657528.1A CN202311657528A CN117390970B CN 117390970 B CN117390970 B CN 117390970B CN 202311657528 A CN202311657528 A CN 202311657528A CN 117390970 B CN117390970 B CN 117390970B
Authority
CN
China
Prior art keywords
gas storage
data
survey model
model
storage state
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
Application number
CN202311657528.1A
Other languages
Chinese (zh)
Other versions
CN117390970A (en
Inventor
马彦彦
毕彩芹
孔丽云
周惠
康海霞
罗卫锋
张云枭
李娟�
刘海浩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Oil & Gas Survey Cgs
Original Assignee
Oil & Gas Survey Cgs
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Oil & Gas Survey Cgs filed Critical Oil & Gas Survey Cgs
Priority to CN202311657528.1A priority Critical patent/CN117390970B/en
Publication of CN117390970A publication Critical patent/CN117390970A/en
Application granted granted Critical
Publication of CN117390970B publication Critical patent/CN117390970B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH OR ROCK DRILLING; MINING
    • E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B49/00—Testing the nature of borehole walls; Formation testing; Methods or apparatus for obtaining samples of soil or well fluids, specially adapted to earth drilling or wells
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/282—Application of seismic models, synthetic seismograms
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/30—Analysis
    • G01V1/301—Analysis for determining seismic cross-sections or geostructures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/30—Analysis
    • G01V1/306—Analysis for determining physical properties of the subsurface, e.g. impedance, porosity or attenuation profiles
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
    • G01V1/30—Analysis
    • G01V1/307—Analysis for determining seismic attributes, e.g. amplitude, instantaneous phase or frequency, reflection strength or polarity
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
    • G01V1/40—Seismology; Seismic or acoustic prospecting or detecting specially adapted for well-logging
    • 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
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • 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
    • G06N3/098—Distributed learning, e.g. federated learning
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/616—Data from specific type of measurement
    • G01V2210/6161—Seismic or acoustic, e.g. land or sea measurements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/61—Analysis by combining or comparing a seismic data set with other data
    • G01V2210/616—Data from specific type of measurement
    • G01V2210/6169—Data from specific type of measurement using well-logging
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/62—Physical property of subsurface
    • G01V2210/624—Reservoir parameters
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/63—Seismic attributes, e.g. amplitude, polarity, instant phase
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V2210/00—Details of seismic processing or analysis
    • G01V2210/60—Analysis
    • G01V2210/64—Geostructures, e.g. in 3D data cubes
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Remote Sensing (AREA)
  • General Physics & Mathematics (AREA)
  • Geology (AREA)
  • Environmental & Geological Engineering (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • Acoustics & Sound (AREA)
  • Geophysics (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Mining & Mineral Resources (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Biophysics (AREA)
  • Geochemistry & Mineralogy (AREA)
  • Fluid Mechanics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

The invention relates to the technical field of geological survey, in particular to an oil gas storage quantification method considering multi-element driving prediction, which comprises the following steps: obtaining a modeling driving survey model for obtaining an oil gas storage state prediction result through a seismic inversion model; obtaining a data driving survey model for obtaining a prediction result of the oil gas storage state through a neural network model; obtaining a multi-element driving survey model for obtaining an oil gas storage state prediction result through a joint learning mechanism; and predicting the oil gas storage state of the target land block through the multi-element driving survey model. According to the invention, the quantitative characteristic parameters for predicting the oil and gas storage state are obtained through geological knowledge guidance, so that the current result of surveying the oil and gas storage state accords with geological rationality, the prediction precision is improved from the data layer, a plurality of driving models are jointly learned through a joint learning mechanism, the surveying effect is improved, the limitation is reduced, and the robustness is enhanced.

Description

Oil gas storage quantification method considering multi-element driving prediction
Technical Field
The invention relates to the technical field of geological survey, in particular to an oil gas storage quantification method considering multi-element driving prediction.
Background
With the deep, ultra-deep, deep water, ultra-deep water and other complex geological environments becoming the main matrix for the important discovery of oil and gas in China, the exploration and development of oil and gas resources face new technical challenges and problems. Among them, it is important to improve the ability of hydrocarbon reservoir prediction and description. Taking reservoir parameter (lithology, elasticity, physical property, oil-gas property and the like) prediction of a target reservoir such as an oil reservoir, a gas reservoir, a coal mine and the like as an example, the traditional geophysical exploration technology is mainly combined with the technologies of seismic attribute qualitative analysis, seismic inversion quantitative prediction and the like to develop comprehensive research of reservoir prediction.
The existing hydrocarbon reservoir prediction method lacks guidance of geological knowledge in the aspects of seismic attribute analysis and seismic inversion, namely, the known geological knowledge is rarely utilized to guide key steps such as seismic attribute analysis and seismic inversion, so that the predicted reservoir parameter result lacks geological rationality, namely, the qualitative or quantitative hydrocarbon reservoir prediction result is easy to deviate from geological reality, so that the prediction accuracy is lower and is inconsistent with the deposit evolution history or a deposit system. The existing seismic attribute analysis method does not have an explicit criterion, the attribute analysis and the importance sorting are carried out only by relying on the correlation between the seismic attribute and the target logging parameter, the selected seismic attribute generally has only mathematical significance, the actual geological rule may not be satisfied, and the corresponding geological significance needs to be interpreted by combining data such as drilling, logging and the like afterwards. The known seismic inversion method depends on the vertical information of the logging data and the vertical and horizontal information of the seismic data to represent the limited space change capacity of the reservoir, and the prediction result often does not meet the geological knowledge and deviates from the underground condition. Meanwhile, in the prior art, only the traditional inversion model is used for carrying out oil gas storage prediction, the quantification of the current oil gas storage state of a model driving mode is realized, the model comprises hypothetical conditions, noise is difficult to process, the exploration quantification of the model driving mode is caused, the error is large, the stability is low, and only the neural network is used for carrying out oil gas storage prediction, the quantification of the current oil gas storage state of a data driving mode is realized, the prior data quality requirement is high, the calculation cost is high, the exploration quantification of the data driving mode is caused, the accuracy is limited by the data quality, and the calculation cost is high.
Therefore, the surveying effect of the oil gas storage state surveying quantification in the prior art is limited by a surveying method, the limitation is strong, the robustness is insufficient, the surveying is lack of guidance of geological knowledge, the current result of oil gas storage state surveying is caused to lack of geological rationality, the current result of oil gas storage state surveying is easily deviated from geological reality, and the prediction precision is low.
Disclosure of Invention
The invention aims to provide an oil gas storage quantification method considering multi-element driving prediction, which aims to solve the technical problems that in the prior art, the surveying effect of oil gas storage state surveying quantification is limited by a surveying method, the limitation is strong, the robustness is insufficient, and the surveying is lack of guidance of geological knowledge, so that the result of surveying the current state of oil gas storage is lack of geological rationality, and the result is easy to deviate from geological reality, so that the prediction precision is lower.
In order to solve the technical problems, the invention specifically provides the following technical scheme:
a method for quantifying oil gas storage taking into account multi-element driving prediction comprises the following steps:
Selecting a group of sample plots with known oil and gas storage states, and measuring seismic data, drilling data and logging data of each sample plot;
Guiding the seismic attribute analysis of the sample plot and the reservoir sensitivity analysis of the sample plot to obtain quantitative characteristic parameters for predicting the oil gas storage state through geological knowledge based on the seismic data, the drilling data and the logging data respectively;
Carrying out modeling driving survey on the oil gas storage state based on the quantitative characteristic parameters through a seismic inversion model to obtain a modeling driving survey model for obtaining an oil gas storage state prediction result;
carrying out data driving survey on the oil gas storage state based on the quantitative characteristic parameters through a neural network model to obtain a data driving survey model for obtaining an oil gas storage state prediction result;
The modeling driving survey model and the data driving survey model are subjected to joint learning through a joint learning mechanism, so that a multi-element driving survey model for obtaining an oil gas storage state prediction result is obtained;
and predicting the oil gas storage state of the target land block through the multi-element driving survey model.
As a preferred embodiment of the present invention, the determining of the quantization parameter includes:
Determining reservoir parameters and reservoir properties of well point positions of a sample plot by using well drilling data and well logging data of the sample plot, connecting points with the same values of the reservoir parameters into a loop, directly projecting the loop onto a plane of an oil-gas reservoir of the sample plot to form a horizontal curve, and obtaining a reservoir plane characteristic change trend graph representing loop combinations of different reservoir parameter values;
Using the reservoir plane characteristic change trend graph as geological knowledge, guiding to screen out seismic attributes basically consistent with geological rules reflected by the reservoir plane characteristic change trend graph from various seismic attributes of the sample land parcels, and eliminating other seismic attributes which cannot reflect the geological rules;
drawing two-dimensional or three-dimensional intersection graphs on different logging data of a sample plot to obtain a logging curve sensitive to the oil and gas storage state of a reservoir;
Logging rock physical intersection analysis is carried out on the logging curve, and logging parameters with highest sensitivity to the oil and gas storage state of a reservoir are screened out from a sample plot;
And taking the seismic attribute and logging parameter screened by the sample land block as the quantitative characteristic parameter.
As a preferred scheme of the invention, the well-seismic joint inversion technology performs screening and determination by the similarity between the predicted oil gas storage state and the reservoir plane characteristic change trend graph, and comprises the following steps:
And carrying out inversion test on the sample plot based on the quantized characteristic parameters by utilizing a plurality of inversion methods, and selecting an inversion method corresponding to the oil gas storage state with the highest similarity between the sample plot reservoir plane characteristic change trend graphs to carry out well-seismic joint inversion technology.
As a preferred embodiment of the present invention, the modeling-driven survey model establishment includes:
Obtaining quantization characteristic parameters of a sample land block;
inversion is carried out based on the quantized characteristic parameters of the sample plots by using a well-seismic joint inversion technology, so that a modeling driving survey model is obtained;
the modeling driven survey model is:
H Reverse =Reverse_mode(S)；
Wherein H Reverse is the oil gas storage state predicted by the modeling driving survey model, S is the quantitative characteristic parameter, and reverse_mode is the well-seismic joint inversion function body;
The learning targets of the modeling-driven survey model are as follows:
Loss Reverse=L1(H Reverse,Ho)；
In the formula, loss Reverse is a learning target of the modeling-driven survey model, H Reverse is an oil gas storage state predicted by the modeling-driven survey model, ho is an oil gas storage state known by a sample plot, L1 is an L1 norm operator, and L1 (H Reverse, ho) is an L1 norm operation of H Reverse and Ho.
As a preferred embodiment of the present invention, the establishing of the data-driven survey model includes:
Obtaining quantization characteristic parameters of a sample land block;
learning and training by using a CNN neural network based on the quantized characteristic parameters of the sample plots and the known oil gas storage state of the sample plots to obtain the data driving survey model;
the databased drive survey model is:
H CNN =CNN_mode(S)；
wherein H CNN is the oil gas storage state predicted by the data driving survey model, S is the quantitative characteristic parameter, and CNN_mode is a CNN neural network structure;
the learning targets of the data-driven survey model are as follows:
Loss CNN=L1(HCNN,Ho)；
Where Loss CNN is a learning target of the data-driven survey model, H CNN is a hydrocarbon storage state predicted by the data-driven survey model, ho is a true hydrocarbon storage state of the sample plot, L1 is an L1 norm operator, and L1 (H CNN, ho) is an L1 norm operation of H CNN and Ho.
As a preferred embodiment of the present invention, the construction of the multi-element driving survey model includes:
Combining learning targets of the data driving survey model and the modeling driving survey model to obtain a shared learning target for combined learning of the data driving survey model and the modeling driving survey model, wherein the shared learning target is as follows:
；
；
；
Where Loss share is a shared learning objective, w CNN (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, loss CNN (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, w Reverse (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, loss Reverse (k) is a learning objective of the data-driven survey model at the kth joint learning stage, M CNN (k) is a model evaluation index of the data-driven survey model at the kth joint learning stage, M Reverse (k) is a model evaluation index of the data-driven survey model at the kth joint learning stage, r CNN is an external regulatory parameter of w CNN (k) at the kth joint learning stage, r Reverse is an external regulatory parameter of w Reverse (k) at the kth joint learning stage;
Sharing an input layer, an output layer and a shared learning target by utilizing a joint learning mechanism for the data driving survey model and the modeling driving survey model, and performing joint learning for the data driving survey model and the modeling driving survey model through the shared input layer, the shared output layer and the shared learning target to obtain the multi-element driving survey model;
The multi-element driving survey model is as follows:
；
Wherein H share is the oil gas storage state predicted by the multi-element driving survey model, S share is the quantitative characteristic parameter, CNN_mode is a CNN neural network structure body, and reverse_mode is a well-seismic joint inversion function body.
As a preferred aspect of the present invention, the predicting the oil gas storage state of the target land by the multi-element driving survey model includes:
Measuring seismic data, drilling data and logging data of a target land block;
guiding seismic attribute analysis of the target land block and reservoir sensitivity analysis of the target land block based on the seismic data, the drilling data and the logging data respectively through geological knowledge to obtain quantitative characteristic parameters of the target land block;
and inputting the quantitative characteristic parameters of the target land block into a multi-element driving survey model, and predicting the oil gas storage state of the target land block by the multi-element driving survey model.
As a preferred embodiment of the present invention, the model evaluation index includes a model accuracy, a model recall, an F1 Score, and an ROC curve.
As a preferred embodiment of the present invention, the oil-gas storage state includes at least one of lithology, elasticity, physical properties, and oil-gas properties of a reservoir, and lithology, elasticity, physical properties, and oil-gas properties of a coal mine reservoir.
As a preferred embodiment of the present invention, the inversion method includes a trace integral inversion, a recursive inversion, a generalized linear inversion, a wideband constraint inversion, a sparse pulse inversion, a simulated annealing inversion, a stochastic inversion, a geostatistical inversion, a sequential gaussian simulation, a sequential indication simulation, a markov chain monte carlo stochastic simulation, and a seismic waveform indication markov chain monte carlo stochastic simulation.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the quantitative characteristic parameters for predicting the oil and gas storage state are obtained through geological knowledge guidance, so that the geological rationality of the current result of surveying the oil and gas storage state is enhanced, the prediction precision is improved from the data layer, the modeling driving survey model and the data driving survey model are subjected to joint learning through a joint learning mechanism, the surveying effect is improved by joint learning of a plurality of driving models, the limitation is reduced, and the robustness is enhanced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flow chart of a method for quantifying oil and gas storage with multi-component drive prediction according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, the invention provides an oil gas storage quantification method considering multi-element driving prediction, which comprises the following steps:
Selecting a group of sample plots with known oil and gas storage states, and measuring seismic data, drilling data and logging data of each sample plot;
guiding the seismic attribute analysis of the sample plot and the reservoir sensitivity analysis of the sample plot to obtain quantitative characteristic parameters for predicting the oil gas storage state through geological knowledge based on the seismic data, the drilling data and the logging data respectively;
Carrying out modeling driving survey on the oil gas storage state based on the quantitative characteristic parameters through the seismic inversion model to obtain a modeling driving survey model for obtaining the oil gas storage state prediction result;
Carrying out data driving survey on the oil gas storage state based on the quantitative characteristic parameters through a neural network model to obtain a data driving survey model for obtaining the oil gas storage state prediction result;
The modeling driving survey model and the data driving survey model are subjected to joint learning through a joint learning mechanism, so that a multi-element driving survey model for obtaining an oil gas storage state prediction result is obtained;
and predicting the oil gas storage state of the target land block through the multi-element driving survey model.
In order to enable the oil gas storage state quantized by survey to conform to geological reality, in survey quantization, the quantized feature parameters for predicting the oil gas storage state are determined by using geological knowledge guidance, so that the seismic attribute in the quantized feature parameters has mathematical significance, and the geological knowledge guidance is changed to meet the actual geological rule and corresponds to the interpretable geological significance, and the interpretation of data such as drilling, logging and the like is not required to be combined afterwards.
In order to improve the surveying effect, reduce the limitation and enhance the robustness, the modeling driving surveying model for obtaining the oil gas storage state prediction result and the data driving surveying model for obtaining the oil gas storage state prediction result are subjected to joint learning, so that the jointly learned multi-element driving surveying model has the advantages of two prediction modes of data driving and model driving, namely, the modeling driving model has the advantages of strong interpretability, data efficiency and generalization capability, the modeling driving model has the advantages of strong adaptability and strong complex data processing capability, the modeling driving model and the model driving model are combined, the respective defects can be complemented, the modeling driving model can compensate the calculation error defects caused by the supposition condition and noise of the modeling driving mode, and the modeling driving mode can compensate the overfitting and dependence data quality defects of the data driving mode.
Therefore, the invention establishes a new technology and a new paradigm of oil and gas reservoir prediction of geological knowledge guidance, model driving and data driving, and seismic, logging and geological information are effectively fused into an oil and gas storage state calculation model (a multi-element driving survey model) to carry out compatible expression and information complementation. The new paradigm can reduce the multi-solution property and uncertainty of reservoir prediction, and the obtained oil and gas storage state prediction result has multiple meanings of mathematics, physics and geology.
According to the invention, geological information is permeated into the links of well-seismic joint inversion and seismic attribute analysis, so that the seismic attribute with geological and seismic characteristics is found, the prediction result of the oil-gas storage state with geological significance is obtained, and the multi-element information depth fusion of earthquake, logging and geological information is realized to a certain extent, and the method comprises the following specific steps:
the determination of the quantization characteristic parameters comprises the following steps:
Determining reservoir parameters and reservoir properties of well point positions of the sample land by using well drilling data and well logging data of the sample land, connecting points with the same values of the reservoir parameters into a loop, directly projecting the loop onto a plane where an oil-gas reservoir of the sample land is positioned to form a horizontal curve, and obtaining a reservoir plane characteristic change trend graph representing loop combinations of different reservoir parameter values;
Using the reservoir plane characteristic change trend graph as geological knowledge, guiding to screen out seismic attributes basically consistent with geological rules reflected by the reservoir plane characteristic change trend graph from various seismic attributes of the sample land parcels, and eliminating other seismic attributes which cannot reflect the geological rules;
drawing two-dimensional or three-dimensional intersection graphs on different logging data of a sample plot to obtain a logging curve sensitive to the oil and gas storage state of a reservoir;
logging rock physical intersection analysis is carried out on logging curves, and logging parameters with highest sensitivity to the oil and gas storage state of a reservoir are screened out from a sample plot;
And taking the seismic attribute and the logging parameter screened by the sample land parcels as quantitative characteristic parameters.
The invention provides an earthquake attribute analysis example, which takes coal series reservoir thickness prediction as an example, is limited by the limitation of earthquake data resolution and strong multi-solution property and uncertainty of an inversion method, and reservoir parameter results obtained by earthquake attribute and earthquake inversion can not completely reflect the space change condition of the thickness of a thin coal seam, but the space spreading rule and the change on a plane represent the change trend of the thickness of the coal seam to a certain extent. Therefore, there is a need to predict reservoir thickness using a combination of inversion results and attribute screening results. The invention provides an intelligent oil and gas reservoir prediction method and device guided by geological knowledge, and aims to improve the accuracy of oil and gas reservoir prediction links such as attribute screening, seismic inversion, reservoir thickness prediction and the like. Firstly, drawing a coal seam thickness change trend chart by using coal seam thickness data counted by drilling coring. Similar to the drawing method of contour map in geography, the invention connects the same points of the coal layer thickness into a loop line based on the coal layer thickness data of the well point position, directly projects the loop line onto the plane of the target layer to form a horizontal curve, and finally, the loop line combination representing different coal layer thicknesses forms a coal layer thickness change trend graph. The coal seam thickness change trend graph provides a new way for converting abstract geological knowledge into specific geological data, and can reflect the spatial distribution of the coal seam near a target layer, so that the method approximately accords with actual geological conditions.
And then carrying out well-seismic joint reservoir parameter inversion guided by geological knowledge by utilizing the coal seam thickness change trend graph. And drawing a two-dimensional or three-dimensional intersection diagram among different logging curves through logging rock physical intersection analysis, and analyzing to determine that the resistivity curve and the artificial gamma curve are sensitive to the coal seam, wherein the logging response is high resistivity and high artificial gamma. Because the artificial gamma is less accurate than the resistivity of the coal seam top-bottom interface and the coal seam thickness, the sensitive reservoir parameter is finally determined to be the resistivity. The characteristic curve can be further finely constructed, and parameters for describing the thickness of the coal seam to be more sensitive are obtained. After the parameters to be inverted are determined, testing is attempted by using different well-seismic joint reservoir parameter inversion methods. And evaluating an inversion method and the quality of the inversion result according to the similarity between the predicted reservoir parameter result and the coal seam thickness change trend graph, namely determining the inversion method and screening the reservoir parameter prediction result under the guidance of geological knowledge. The quality control of the inversion result is participated through the coal seam thickness change trend graph, the seismic waveform indication inversion is determined as an optimal inversion algorithm, and the reservoir parameter result predicted by the algorithm is most matched with the geological rule reflected by the coal seam thickness change trend graph.
The seismic waveform indication inversion analyzes the correlation and the difference between known wells according to the seismic waveform characteristics, screens well samples with high correlation degree and closer distance to the seismic channel waveform to be distinguished, establishes an initial model, and counts reservoir parameters as prior information. The seismic waveforms with dense distribution can accurately represent the low-frequency trend of the stratum space structure, and known well samples with similar low-frequency structures are screened to serve as space estimation samples according to two indexes of waveform similarity and space distance. Compared with the traditional variational function, the method is not easy to be influenced by well position distribution, and can more accurately represent the heterogeneity of the reservoir. And then carrying out matched filtering on the initial model and reservoir parameters in the seismic frequency band, and calculating to obtain a likelihood function. And combining likelihood function distribution and prior distribution under a Bayesian framework to obtain posterior probability distribution, taking the posterior probability distribution as an objective function, continuously perturbing model parameters, enabling a solution when the posterior probability distribution function is maximum to be effectively and randomly realized, and taking the average value of multiple effective realizations as an expected value of reservoir parameters to be output.
Further, the earthquake attribute analysis is guided by using the coal seam thickness change trend graph. The seismic data obtains various seismic attributes through different attribute calculation formulas, and different seismic attributes and calculation formulas thereof can be combined to form a seismic attribute comprehensive learner. A plurality of sub learners in the seismic attribute comprehensive learner correspond to a calculation formula of an attribute, and the calculation formula is a mapping function of seismic data to a certain attribute. Thus, the seismic attribute synthesis learner may be viewed as a combination of complex functions that may extract a variety of desired seismic attributes from the seismic data. The seismic attributes obtained by the seismic attribute comprehensive learner are further compared with a coal seam thickness change trend chart, the seismic attributes are ranked according to the similarity, and two seismic attributes capable of reflecting the spatial distribution characteristics of the coal seam thickness are screened out, namely the average amplitude and the instantaneous frequency.
The well-seismic joint inversion technology performs screening and determination by the similarity between the predicted oil gas storage state and the reservoir plane characteristic change trend graph, and comprises the following steps:
And carrying out inversion test on the sample plot based on the quantized characteristic parameters by utilizing a plurality of inversion methods, and selecting an inversion method corresponding to the oil gas storage state with the highest similarity between the sample plot reservoir plane characteristic change trend graphs to carry out well-seismic joint inversion technology.
In order to give consideration to data driving and model driving in quantitative prediction of an oil gas storage state, prediction models are arranged on the data driving layer and the model driving layer, namely, the seismic inversion model is used for carrying out modeling driving survey of the oil gas storage state based on quantitative characteristic parameters, and the neural network model is used for carrying out data driving survey of the oil gas storage state based on quantitative characteristic parameters, and the two driving prediction models are used for carrying out oil gas storage state prediction respectively, so that the effect of jointly learning a plurality of driving models to improve survey is realized, the limitation is reduced, and the robustness is enhanced.
In order to give consideration to data driving and model driving in quantitative prediction of oil and gas storage states, the invention sets prediction models on the data driving and model driving layers, and specifically comprises the following steps:
Modeling-driven survey model creation, comprising:
Obtaining quantization characteristic parameters of a sample land block;
Inversion is carried out based on the quantized characteristic parameters of the sample plots by using a well-seismic joint inversion technology, so that a modeling driving survey model is obtained;
the modeled driving survey model is:
H Reverse =Reverse_mode(S)；
Wherein H Reverse is the oil gas storage state predicted by the modeling driving survey model, S is the quantitative characteristic parameter, and reverse_mode is the well-seismic joint inversion function body;
the learning targets of the modeling driven survey model are:
Loss Reverse=L1(H Reverse,Ho)；
In the formula, loss Reverse is a learning target of the modeling-driven survey model, H Reverse is an oil gas storage state predicted by the modeling-driven survey model, ho is an oil gas storage state known by a sample plot, L1 is an L1 norm operator, and L1 (H Reverse, ho) is an L1 norm operation of H Reverse and Ho.
The establishment of the data-driven survey model comprises the following steps:
Obtaining quantization characteristic parameters of a sample land block;
Learning and training by using a CNN neural network based on the quantized characteristic parameters of the sample plots and the known oil gas storage state of the sample plots to obtain a data driving survey model;
the data driven survey model is:
H CNN =CNN_mode(S)；
wherein H CNN is the oil gas storage state predicted by the data driving survey model, S is the quantitative characteristic parameter, and CNN_mode is a CNN neural network structure;
the learning objectives of the data driven survey model are:
Loss CNN=L1(HCNN,Ho)；
Where Loss CNN is a learning target of the data-driven survey model, H CNN is a hydrocarbon storage state predicted by the data-driven survey model, ho is a true hydrocarbon storage state of the sample plot, L1 is an L1 norm operator, and L1 (H CNN, ho) is an L1 norm operation of H CNN and Ho.
The invention utilizes a joint learning mechanism to realize the fusion of two driving prediction models, and the joint learning carries out simultaneous learning on the two driving prediction models, and can improve the performance of each driving prediction model through sharing a network structure and a learning target, and the method comprises the following specific steps:
the construction of a multi-element drive survey model comprises the following steps:
combining learning targets of the data driving survey model and the modeling driving survey model to obtain a shared learning target for combined learning of the data driving survey model and the modeling driving survey model, wherein the shared learning target is as follows:
；
；
；
Wherein, loss share is a shared learning target, w CNN (k) is a learning target weight of the data driving survey model at the kth joint learning stage, loss CNN (k) is a learning target weight of the data driving survey model at the kth joint learning stage, w Reverse (k) is a learning target weight of the data driving survey model at the kth joint learning stage, loss Reverse (k) is a learning target of the data driving survey model at the kth joint learning stage, M CNN (k) is a model evaluation index of the data driving survey model at the kth joint learning stage, M Reverse (k) is a model evaluation index of the data driving survey model at the kth joint learning stage, r CNN is an external regulation parameter of w CNN (k) at the kth joint learning stage, r Reverse is an external regulation parameter of w Reverse (k) at the kth joint learning stage, and the external regulation parameter is used for manually intervening the model driving survey model and the data driving survey model according to actual requirements, that is, i.e. manually controlling the external regulation and control parameters are manually controlled, the external regulation and control parameters are automatically performed under the condition that the joint learning is not normally, and self-adaptive to the condition is set up under the condition that the joint learning is realized;
Sharing an input layer, an output layer and a shared learning target by utilizing a joint learning mechanism for the data driving survey model and the modeling driving survey model, and performing joint learning for the data driving survey model and the modeling driving survey model through the shared input layer, the shared output layer and the shared learning target to obtain a multi-element driving survey model;
the multi-element driving survey model is:
；
Wherein H share is the oil gas storage state predicted by the multi-element driving survey model, S share is the quantitative characteristic parameter, CNN_mode is a CNN neural network structure body, and reverse_mode is a well-seismic joint inversion function body.
Specifically, the invention utilizes a joint learning mechanism to realize the fusion of two driving prediction models, and the joint learning carries out simultaneous learning on the two driving prediction models, so that the performance of each driving prediction model can be improved through sharing a network structure and sharing a learning target. Meanwhile, the two driving prediction models can be combined into one prediction model with a multi-element driving mode through joint learning, so that training and reasoning burden of a plurality of independent models is reduced, respective defects can be complemented, the data driving mode can compensate calculation error defects caused by suppositional conditions and noise of the model driving mode, the model driving mode can compensate overfitting and dependence data quality defects of the data driving mode, calculation efficiency is improved, and generalization capability and efficiency of driving prediction can be improved finally.
Furthermore, the invention combines two driving prediction modes in a dynamic task priority mode, so that when two driving prediction models are simultaneously studied and trained, a harder task has higher study priority, the study self-adaption capability of the two driving prediction models is enhanced, the learning difficulty and even the learning effect are adjusted according to the study stages of different tasks, and finally, the learning effect of the two driving prediction models is simultaneously improved, so that the prediction quantization effect of the multi-element driving survey model on the oil gas storage state is optimal.
The invention utilizes the model evaluation index to measure the task learning target, and the higher the model evaluation index is, the lower the learning target is. The invention reflects the learning priority of the driving mode by using the weight of the learning target, and the higher the weight of the learning target is, the higher the priority is. Therefore, the model evaluation index is mapped to the weight calculation of the learning target, the task of high model evaluation index is realized, the learning is simpler, and the weight can be reduced; on the contrary, the weight of the difficult task becomes larger, namely, the task which is more difficult to learn is enabled to have higher learning target weight, the task which is more difficult to learn is enabled to have higher priority, namely, the model with higher learning target is enabled to learn more preferentially. The learning self-adaptive capacity of two driving models for adjusting according to the learning stages of different models and the learning difficulty level and even the learning effect is achieved.
Predicting an oil gas storage state of a target land block through a multi-element driving survey model, comprising:
Measuring seismic data, drilling data and logging data of a target land block;
The method comprises the steps of respectively guiding seismic attribute analysis of a target land block and reservoir sensitivity analysis of the target land block based on seismic data, drilling data and logging data through geological knowledge to obtain quantitative characteristic parameters of the target land block;
and inputting the quantitative characteristic parameters of the target land block into a multi-element driving survey model, and predicting the oil gas storage state of the target land block by the multi-element driving survey model.
The model evaluation index comprises a model accuracy rate, a model recall rate, an F1 Score and an ROC curve.
The hydrocarbon storage state includes at least one of lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir, lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir, and lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir in a coal mine.
The inversion method comprises the steps of channel integral inversion, recursion inversion, generalized linear inversion, broadband constraint inversion, sparse pulse inversion, simulated annealing inversion, random inversion, geostatistical inversion, sequential Gaussian simulation, sequential indication simulation, markov chain Monte Carlo random simulation and seismic waveform indication Markov chain Monte Carlo random simulation.
According to the invention, the quantitative characteristic parameters for predicting the oil and gas storage state are obtained through geological knowledge guidance, so that the current result of surveying the oil and gas storage state accords with geological rationality, the prediction precision is improved from the data layer, the modeling driving survey model and the data driving survey model are subjected to joint learning through a joint learning mechanism, the surveying effect is improved by joint learning of a plurality of driving models, the limitation is reduced, and the robustness is enhanced.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements of this application will occur to those skilled in the art, and are intended to be within the spirit and scope of the application.

Claims (5)

1. The oil gas storage quantification method taking into account multi-element driving prediction is characterized by comprising the following steps of:
Selecting a group of sample plots with known oil and gas storage states, and measuring seismic data, drilling data and logging data of each sample plot;
Guiding the seismic attribute analysis of the sample plot and the reservoir sensitivity analysis of the sample plot to obtain quantitative characteristic parameters for predicting the oil gas storage state through geological knowledge based on the seismic data, the drilling data and the logging data respectively;
Carrying out modeling driving survey on the oil gas storage state based on the quantitative characteristic parameters through a seismic inversion model to obtain a modeling driving survey model for obtaining an oil gas storage state prediction result;
carrying out data driving survey on the oil gas storage state based on the quantitative characteristic parameters through a neural network model to obtain a data driving survey model for obtaining an oil gas storage state prediction result;
The modeling driving survey model and the data driving survey model are subjected to joint learning through a joint learning mechanism, so that a multi-element driving survey model for obtaining an oil gas storage state prediction result is obtained;
predicting the oil gas storage state of a target land block through a multi-element driving survey model;
the determining of the quantization characteristic parameter comprises the following steps:
Determining reservoir parameters and reservoir properties of well point positions of a sample plot by using well drilling data and well logging data of the sample plot, connecting points with the same values of the reservoir parameters into a loop, directly projecting the loop onto a plane of an oil-gas reservoir of the sample plot to form a horizontal curve, and obtaining a reservoir plane characteristic change trend graph representing loop combinations of different reservoir parameter values;
Using the reservoir plane characteristic change trend graph as geological knowledge, guiding to screen out seismic attributes basically consistent with geological rules reflected by the reservoir plane characteristic change trend graph from various seismic attributes of the sample land parcels, and eliminating other seismic attributes which cannot reflect the geological rules;
drawing two-dimensional or three-dimensional intersection graphs on different logging data of a sample plot to obtain a logging curve sensitive to the oil and gas storage state of a reservoir;
Logging rock physical intersection analysis is carried out on the logging curve, and logging parameters with highest sensitivity to the oil and gas storage state of a reservoir are screened out from a sample plot;
Taking the seismic attribute and logging parameter screened by the sample land block as the quantitative characteristic parameter;
The well-seismic joint inversion technology performs screening and determination by the similarity between the predicted oil gas storage state and the reservoir plane characteristic change trend graph, and comprises the following steps:
Performing inversion test on the sample plot based on the quantized characteristic parameters by utilizing a plurality of inversion methods, and selecting an inversion method corresponding to the oil gas storage state with the highest similarity between the sample plot reservoir plane characteristic change trend graphs to perform well-seismic joint inversion technology;
The modeling-driven survey model establishment includes:
Obtaining quantization characteristic parameters of a sample land block;
inversion is carried out based on the quantized characteristic parameters of the sample plots by using a well-seismic joint inversion technology, so that a modeling driving survey model is obtained;
the modeling driven survey model is:
H Reverse =Reverse_mode(S)；
Wherein H Reverse is the oil gas storage state predicted by the modeling driving survey model, S is the quantitative characteristic parameter, and reverse_mode is the well-seismic joint inversion function body;
The learning targets of the modeling-driven survey model are as follows:
Loss Reverse=L1(H Reverse,Ho)；
In the formula, loss Reverse is a learning target of the modeling driving survey model, H Reverse is an oil gas storage state predicted by the modeling driving survey model, ho is an oil gas storage state known by a sample plot, L1 is an L1 norm operator, and L1 (H Reverse, ho) is an L1 norm operation of H Reverse and Ho;
The establishing of the data-driven survey model comprises the following steps:
Obtaining quantization characteristic parameters of a sample land block;
learning and training by using a CNN neural network based on the quantized characteristic parameters of the sample plots and the known oil gas storage state of the sample plots to obtain the data driving survey model;
the databased drive survey model is:
H CNN =CNN_mode(S)；
wherein H CNN is the oil gas storage state predicted by the data driving survey model, S is the quantitative characteristic parameter, and CNN_mode is a CNN neural network structure;
the learning targets of the data-driven survey model are as follows:
Loss CNN=L1(HCNN,Ho)；
In the formula, loss CNN is a learning target of a data driving survey model, H CNN is an oil gas storage state predicted by the data driving survey model, ho is a real oil gas storage state of a sample land block, L1 is an L1 norm operator, and L1 (H CNN, ho) is an L1 norm operation of H CNN and Ho;
the construction of the multi-element driving survey model comprises the following steps:
Combining learning targets of the data driving survey model and the modeling driving survey model to obtain a shared learning target for combined learning of the data driving survey model and the modeling driving survey model, wherein the shared learning target is as follows:
；
；
；
Where Loss share is a shared learning objective, w CNN (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, loss CNN (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, w Reverse (k) is a learning objective weight of the data-driven survey model at the kth joint learning stage, loss Reverse (k) is a learning objective of the data-driven survey model at the kth joint learning stage, M CNN (k) is a model evaluation index of the data-driven survey model at the kth joint learning stage, M Reverse (k) is a model evaluation index of the data-driven survey model at the kth joint learning stage, r CNN is an external regulatory parameter of w CNN (k) at the kth joint learning stage, r Reverse is an external regulatory parameter of w Reverse (k) at the kth joint learning stage;
Sharing an input layer, an output layer and a shared learning target by utilizing a joint learning mechanism for the data driving survey model and the modeling driving survey model, and performing joint learning for the data driving survey model and the modeling driving survey model through the shared input layer, the shared output layer and the shared learning target to obtain the multi-element driving survey model;
The multi-element driving survey model is as follows:
；
Wherein H share is the oil gas storage state predicted by the multi-element driving survey model, S share is the quantitative characteristic parameter, CNN_mode is a CNN neural network structure body, and reverse_mode is a well-seismic joint inversion function body.
2. The method for quantifying oil and gas storage with multi-component drive prediction according to claim 1, wherein the method is characterized in that: the predicting the oil gas storage state of the target land block through the multi-element driving survey model comprises the following steps:
Measuring seismic data, drilling data and logging data of a target land block;
guiding seismic attribute analysis of the target land block and reservoir sensitivity analysis of the target land block based on the seismic data, the drilling data and the logging data respectively through geological knowledge to obtain quantitative characteristic parameters of the target land block;
and inputting the quantitative characteristic parameters of the target land block into a multi-element driving survey model, and predicting the oil gas storage state of the target land block by the multi-element driving survey model.
3. The method for quantifying oil and gas storage with multi-component drive prediction according to claim 1, wherein the method is characterized in that: the model evaluation index comprises a model accuracy rate, a model recall rate, an F1 Score and an ROC curve.
4. The method for quantifying oil and gas storage with multi-component drive prediction according to claim 1, wherein the method is characterized in that: the hydrocarbon storage state includes at least one of lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir, lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir, and lithology, elasticity, physical properties, hydrocarbon-bearing properties of the reservoir in a coal mine.
5. The method for quantifying oil and gas storage with multi-component drive prediction according to claim 1, wherein the method is characterized in that: the inversion method comprises the steps of channel integral inversion, recursion inversion, generalized linear inversion, broadband constraint inversion, sparse pulse inversion, simulated annealing inversion, random inversion, geostatistical inversion, sequential Gaussian simulation, sequential indication simulation, markov chain Monte Carlo random simulation and Markov chain Monte Carlo random simulation.
CN202311657528.1A 2023-12-06 2023-12-06 Oil gas storage quantification method considering multi-element driving prediction Active CN117390970B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311657528.1A CN117390970B (en) 2023-12-06 2023-12-06 Oil gas storage quantification method considering multi-element driving prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311657528.1A CN117390970B (en) 2023-12-06 2023-12-06 Oil gas storage quantification method considering multi-element driving prediction

Publications (2)

Publication Number Publication Date
CN117390970A CN117390970A (en) 2024-01-12
CN117390970B true CN117390970B (en) 2024-09-17

Family

ID=89466869

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311657528.1A Active CN117390970B (en) 2023-12-06 2023-12-06 Oil gas storage quantification method considering multi-element driving prediction

Country Status (1)

Country Link
CN (1) CN117390970B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104569344A (en) * 2014-12-31 2015-04-29 长江大学 Shale reservoir brittle mineral earthquake quantitative characterization method
CN115267935A (en) * 2022-07-01 2022-11-01 中国石油大学(北京) Logging reservoir evaluation method and device driven by data and physical model in combined mode

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8332348B1 (en) * 2009-07-10 2012-12-11 United Services Automobile Association (Usaa) Knowledge extraction and prediction
CN112017289B (en) * 2020-08-31 2023-03-24 电子科技大学 Well-seismic combined initial lithology model construction method based on deep learning
CN115755215A (en) * 2022-12-06 2023-03-07 中国地质调查局油气资源调查中心 Geological knowledge guided intelligent oil and gas reservoir prediction method
CN115879647B (en) * 2023-02-15 2023-06-30 深圳市峰和数智科技有限公司 Stratum lithology prediction method based on combination of domain knowledge and deep learning

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104569344A (en) * 2014-12-31 2015-04-29 长江大学 Shale reservoir brittle mineral earthquake quantitative characterization method
CN115267935A (en) * 2022-07-01 2022-11-01 中国石油大学(北京) Logging reservoir evaluation method and device driven by data and physical model in combined mode

Also Published As

Publication number Publication date
CN117390970A (en) 2024-01-12

Similar Documents

Publication Publication Date Title
RU2462755C2 (en) Predicting properties of underground formation
Ismail et al. Gas channels and chimneys prediction using artificial neural networks and multi-seismic attributes, offshore West Nile Delta, Egypt
EP4133311A1 (en) Identifying hydrocarbon reserves of a subterranean region using a reservoir earth model that models characteristics of the region
CN111596978A (en) Web page display method, module and system for lithofacies classification by artificial intelligence
CN108875122A (en) The artificial intelligence approach and system of geologic parameter are calculated using well logging data
CN113553780A (en) Stratum pore pressure prediction method based on machine learning
Tian et al. Quantitative evaluation of key geological controls on regional Eagle Ford shale production using spatial statistics
CN116168224A (en) Machine learning lithology automatic identification method based on imaging gravel content
CN114529110A (en) Lithofacies inversion method and system based on deep neural network model
Wang et al. Improved permeability prediction based on the feature engineering of petrophysics and fuzzy logic analysis in low porosity–permeability reservoir
CN113033648A (en) Method for realizing logging interpretation by using machine learning algorithm
CN112505754B (en) Method for collaborative partitioning sedimentary microfacies by well-seismic based on high-precision sequence grid model
CN110118994B (en) Continental facies hydrocarbon source rock quantitative prediction method based on seismic inversion and machine learning
CN117390970B (en) Oil gas storage quantification method considering multi-element driving prediction
Sarkheil et al. The fracture network modeling in naturally fractured reservoirs using artificial neural network based on image loges and core measurements
CN115755215A (en) Geological knowledge guided intelligent oil and gas reservoir prediction method
Nath et al. Prediction and analysis of geomechanical properties using deep learning: A Permian Basin case study
Bataee et al. A New Approach in Reservoir Characterization Using Artificial Intelligence
Le Nir et al. Cross-section construction from automated well log correlation: a dynamic programming approach using multiple well logs
CN113409460B (en) Machine learning type three-dimensional quantitative characterization method for clastic rock reservoir interlayer
CN115327643B (en) Machine learning training sample expansion and evaluation method for intelligent oil gas detection
Damasceno et al. Source rock evaluation from rock to seismic: Integrated machine learning based workflow
CN117759235B (en) Method for realizing mineral exploration based on limited drilling
Otmane Integrated reservoir simulation and machine learning for enhanced reservoir characterization and performance prediction
Mohseni et al. Application of artificial neural networks for prediction of Sarvak Formation lithofacies based on well log data, Marun oil field, SW Iran

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