CN110837115B - Seismic identification method and device for lithology of land-facies mixed rock compact reservoir - Google Patents
Seismic identification method and device for lithology of land-facies mixed rock compact reservoir Download PDFInfo
- Publication number
- CN110837115B CN110837115B CN201911051312.4A CN201911051312A CN110837115B CN 110837115 B CN110837115 B CN 110837115B CN 201911051312 A CN201911051312 A CN 201911051312A CN 110837115 B CN110837115 B CN 110837115B
- Authority
- CN
- China
- Prior art keywords
- lithology
- training
- model
- data
- seismic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000011435 rock Substances 0.000 title claims abstract description 102
- 238000000034 method Methods 0.000 title claims abstract description 80
- 208000035126 Facies Diseases 0.000 title claims abstract description 69
- 238000012549 training Methods 0.000 claims description 111
- 238000004422 calculation algorithm Methods 0.000 claims description 58
- 230000006870 function Effects 0.000 claims description 49
- 238000012360 testing method Methods 0.000 claims description 28
- 238000004590 computer program Methods 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 15
- 238000003066 decision tree Methods 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 11
- 238000005553 drilling Methods 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 9
- 230000005484 gravity Effects 0.000 claims description 7
- 238000011160 research Methods 0.000 claims description 7
- 239000000654 additive Substances 0.000 claims description 6
- 230000000996 additive effect Effects 0.000 claims description 6
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000000151 deposition Methods 0.000 claims description 2
- 230000008021 deposition Effects 0.000 claims 1
- 238000010801 machine learning Methods 0.000 description 26
- 238000010586 diagram Methods 0.000 description 22
- 238000005457 optimization Methods 0.000 description 21
- 230000008901 benefit Effects 0.000 description 14
- 238000012545 processing Methods 0.000 description 14
- 238000004891 communication Methods 0.000 description 8
- 230000002349 favourable effect Effects 0.000 description 8
- 238000011161 development Methods 0.000 description 7
- 230000018109 developmental process Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 6
- 238000002790 cross-validation Methods 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 238000013138 pruning Methods 0.000 description 6
- 235000009508 confectionery Nutrition 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 235000021185 dessert Nutrition 0.000 description 3
- 230000000704 physical effect Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- BVKZGUZCCUSVTD-UHFFFAOYSA-L Carbonate Chemical compound [O-]C([O-])=O BVKZGUZCCUSVTD-UHFFFAOYSA-L 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000007500 overflow downdraw method Methods 0.000 description 2
- 239000003208 petroleum Substances 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 238000003892 spreading Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 108010001267 Protein Subunits Proteins 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000007795 chemical reaction product Substances 0.000 description 1
- 235000011850 desserts Nutrition 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 230000035699 permeability Effects 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 239000013049 sediment Substances 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- 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
- G01V1/306—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
- 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
- G01V2210/66—Subsurface modeling
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Geophysics (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention provides a method and a device for identifying lithology of a land-facies mixed rock tight reservoir, wherein the method comprises the following steps: acquiring earthquake elastic parameters of a lithologic work area to be identified; and performing lithology recognition on the continental facies mixed-deposit rock tight reservoir of the lithology work area to be recognized by using the elastic parameters and a pre-established target lithology recognition model. By using the method, the accuracy of lithology prediction of the continental facies mixed rock compact reservoir can be effectively improved under the condition of a small amount of learning samples.
Description
Technical Field
The invention relates to the field of petroleum exploration, in particular to the field of geophysical exploration, and particularly relates to a method and a device for seismic identification of lithology of a land-facies mixed rock compact reservoir.
Background
In recent years, with the increasing demand for energy by human beings, unconventional resources such as dense oil and gas have become a major area of interest for global oil exploration and development. The compact oil refers to petroleum stored in reservoirs such as compact sandstone, compact carbonate rock and the like with the overburden matrix permeability of less than or equal to 0.1 mD. The Chinese continental compact reservoir mainly comprises fine sandstone, siltstone, argillaceous siltstone, cloud rock, marl rock and the like deposited in the environments of lake facies anterior delta subphase, semi-deep lake facies gravity flow subphase and the like. The reservoir heterogeneous method is comprehensively controlled by factors such as the scale of a lake basin, ancient landforms, ancient climate and sediment sources, the facies change of a continental compact reservoir is fast, the thickness change is large, thin layers and thin interbed layers are more, the spreading scale is relatively small, lithological combinations are various, and the reservoir heterogeneous is strong. Therefore, the continental compact reservoir is mainly composed of mixed rock formed by mixing and depositing various continental clastics, carbonate rocks and other components, and the lithology identification is of great significance for compact reservoir 'sweet spot' description and favorable interval optimization. Because the difference between the mineral components of the reservoir and the adjacent non-reservoir rock is small, the lithology of the compact reservoir of the mixed rock is difficult to accurately identify by the traditional seismic rock physical template and the attribute fusion method.
Because the lithology-seismic elasticity characteristic mapping relation of the mixed rock also presents highly nonlinear characteristics, and the seismic rock physical response characteristic differentiation degree of various lithologies is very low, the traditional machine learning method (a support vector machine, a fuzzy logic algorithm and the like) is difficult to be well suitable for the problem, such as overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like.
Disclosure of Invention
Aiming at the problems in the prior art, the method and the device for identifying the lithology of the facies mixed laminated rock tight reservoir can effectively improve the accuracy of the lithology prediction of the facies mixed laminated rock tight reservoir by combining a plurality of weak classifiers into one strong classifier under the condition of a small number of learning samples.
In order to solve the technical problems, the invention provides the following technical scheme:
in a first aspect, the invention provides a method for identifying lithology of a land-facies mixed rock tight reservoir, which comprises the following steps:
acquiring earthquake elastic parameters of a lithologic work area to be identified;
and performing lithology recognition on the continental facies mixed-deposit rock tight reservoir of the lithology work area to be recognized by using the elastic parameters and a pre-established target lithology recognition model.
In one embodiment, the acquiring the seismic elasticity parameter of the lithological work area to be identified includes: acquiring seismic elasticity parameters of a lithologic work area to be identified according to seismic data, wherein the seismic elasticity parameters comprise: longitudinal wave impedance and longitudinal-transverse wave velocity ratio.
In one embodiment, the seismic identification method for lithology of the land-facies mixed rock tight reservoir further comprises the following steps: the step of establishing the target lithology identification model comprises the following steps:
generating lithology classification data of the tight reservoir according to at least one of drilling data, logging data and core data of a research area;
acquiring logging data corresponding to the lithologic classification data;
generating logging elastic parameters according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio;
and establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm.
In an embodiment, the establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm includes:
generating a training set and test data according to the logging elasticity parameters and the lithology classification data;
generating an initial lithology recognition model according to the training loss, the positive regularization item, the objective function parameter and the training set;
training the initial lithology recognition model by utilizing an addition training algorithm and a multi-tree learning algorithm to generate a training result;
verifying the training result according to the test data;
and changing the number and the maximum depth of the decision trees in the initial lithology recognition model according to a verification result so as to generate the target lithology recognition model.
In a second aspect, the invention provides a seismic recognition device for lithology of a land-facies mixed rock tight reservoir, comprising:
the earthquake elastic parameter acquisition unit is used for acquiring earthquake elastic parameters of the lithological work area to be identified;
and the lithology recognition unit is used for carrying out lithology recognition on the land-facies mixed rock compact reservoir of the lithology work area to be recognized by utilizing the elastic parameters and a pre-established target lithology recognition model.
In one embodiment, the seismic elastic parameter acquisition unit includes: the seismic elastic parameter acquisition module is used for acquiring seismic elastic parameters of the lithological work area to be identified according to the seismic data, and the seismic elastic parameters comprise: longitudinal wave impedance and longitudinal-to-transverse wave velocity ratio.
In one embodiment, the seismic identification device for lithology of the land-facies mixed rock tight reservoir further comprises: an object model building unit comprising:
the lithology selection module is used for generating lithology classification data of the compact reservoir according to at least one of drilling data, logging data and core data of a research area;
the logging data acquisition module is used for acquiring logging data corresponding to the lithology classification data;
the logging elastic parameter generating module is used for generating logging elastic parameters according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio;
and the target model establishing module is used for establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by utilizing an extreme gradient lifting algorithm.
In one embodiment, the object model building module comprises:
the training set generation module generates a training set and test data according to the logging elasticity parameters and the lithology classification data;
the initial model generation module is used for generating an initial lithology recognition model according to the training loss, the positive regularization item, the target function parameter and the training set;
the result generation module is used for training the initial lithology recognition model by utilizing an addition training algorithm and a multi-tree learning algorithm so as to generate a training result;
the result verification module is used for verifying the training result according to the test data;
and the target model generation module is used for changing the number and the maximum depth of the decision trees in the initial lithology recognition model according to a verification result so as to generate the target lithology recognition model.
In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for seismic identification of lithology of a continental facies mixed-stratigraphic rock tight reservoir.
In a fourth aspect, the invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of a method for seismic identification of land-mixed petroliferous compact reservoir lithology.
The invention provides a method and a device for seismic identification of lithology of a land-facies mixed-deposit rock tight reservoir, which are characterized in that an improved XGBOOST (eXtreme Gradient boosting) provided by the method is used, based on an ensemble learning algorithm, a training set and a testing set are constructed by using well logging data or rock physical experimental data of known lithology, XGBOOST models under different parameters are established, parameter optimization is carried out, a final XGBOOST model is constructed, and the spatial distribution of the lithology of mixed-deposit rock is predicted by combining elastic parameter inversion results of seismic data before stacking. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In summary, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of the XGB OST algorithm is fully utilized. The XGBOST model based on the XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and spatial distribution of mixed rock lithology prediction can be obtained, so that important technical support is provided for sweet spot identification and favorable zone development of a tight oil reservoir.
Drawings
In order to more clearly illustrate the embodiments or technical solutions of the present invention, the drawings used in the embodiments or technical solutions in the prior art are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a first schematic flow chart of a seismic identification method for lithology of a land-phase mixed rock tight reservoir in an embodiment of the invention;
FIG. 2 is a flowchart illustrating step 100 according to an embodiment of the present invention;
FIG. 3 is a second schematic flow chart of the seismic identification method for lithology of a land-facies mixed rock tight reservoir in the embodiment of the invention;
FIG. 4 is a flowchart illustrating step 300 according to an embodiment of the present invention;
FIG. 5 is a flowchart illustrating step 304 according to an embodiment of the present invention;
FIG. 6 is a schematic flow chart of a seismic identification method for lithology of a land-facies mixed rock tight reservoir in a specific application example of the invention;
FIG. 7 is a schematic diagram of the flow of the concept of the method for seismic identification of lithology of a land-facies mixed rock tight reservoir in a specific application example of the present invention;
FIG. 8 is a diagram of A1 well lithology achievement in a specific application example of the present invention;
FIG. 9 is a graph of an A1 well IP log in an example embodiment of the present invention;
FIG. 10 is a Vp/Vs logging graph for A1 well in an example embodiment of the present invention;
FIG. 11 is a diagram of the lithologic performance of a B1 well in an example of an implementation of the present invention;
FIG. 12 is a graph of a B1 well IP log in an example embodiment of the present invention;
FIG. 13 is a chart of Vp/Vs log of a B1 well in an example embodiment of the present invention;
FIG. 14 is a diagram of the actual lithologic performance of the A1 well in an example of the practice of the present invention;
FIG. 15 is a diagram of the actual lithologic performance of the B1 well in an example of the practice of the present invention;
FIG. 16 is a diagram of predicted lithologic performance of the A1 well in an example embodiment of the present invention;
FIG. 17 is a diagram of predicted lithologic performance of a B1 well in an example embodiment of the present invention;
FIG. 18 is a diagram illustrating seismic predictions of the lithology of a continental facies mixed-stratigraphic tight reservoir in an exemplary embodiment of the present invention;
FIG. 19 is a first schematic structural diagram of a seismic recognition device for lithology of a land-facies mixed rock tight reservoir in an embodiment of the invention;
FIG. 20 is a schematic structural diagram of a seismic elastic parameter acquisition unit according to an embodiment of the invention;
FIG. 21 is a structural schematic diagram of a second seismic identification device for lithology of a land-facies mixed rock tight reservoir in an embodiment of the invention;
FIG. 22 is a schematic structural diagram of a target model building unit according to an embodiment of the present invention;
FIG. 23 is a block diagram illustrating a target model building block according to an embodiment of the present invention;
fig. 24 is a schematic structural diagram of an electronic device in an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Aiming at the problems that in the prior art, continental compact reservoirs mainly comprise fine sandstone, siltstone, argillaceous siltstone, cloud rock, marlite and the like deposited in environments such as lake-facies front delta subphase, semi-deep lake facies gravity flow subphase and the like, and have the characteristics of fast phase zone change, large thickness change, multiple thin layers and thin interbeddes, relatively small spreading scale, various lithological combinations, strong reservoir heterogeneity and the like. Therefore, the lithology of the compact reservoir of the mixed rock is difficult to accurately identify by the traditional seismic rock physical template and the attribute fusion method.
In addition, due to the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like, the lithology of the land-facies mixed laminated rock tight reservoir is difficult to accurately identify by the traditional machine learning method (a support vector machine, a fuzzy logic algorithm and the like), and based on the problems, the embodiment of the invention provides a specific implementation mode of the seismic identification method of the lithology of the land-facies mixed laminated rock tight reservoir, and the method specifically comprises the following contents:
step 100: and acquiring the seismic elasticity parameters of the lithological work area to be identified.
The seismic elastic parameters in step 100 may include: longitudinal wave impedance and longitudinal and transverse wave speeds, on one hand, the two elastic properties are relatively sensitive to lithology of a dense reservoir of the land-phase mixed rock; on the other hand, the two attributes respectively represent AVO intercept and gradient response characteristics of the pre-stack seismic data and can be accurately obtained from the seismic data.
Step 200: and performing lithology recognition on the continental facies mixed-deposit rock tight reservoir of the lithology work area to be recognized by using the elastic parameters and a pre-established target lithology recognition model.
It is understood that the XGBOOST (eXtreme Gradient boost) algorithm is integrated by many CART regression trees. Unlike the bagging integration of random forests, it is a boosting ensemble learning (decision is combined by multiple associated decision trees, and the next decision tree input sample is related to the training and prediction of the previous decision tree). The objective is to establish K regression trees, so that the predicted values of the tree group are as close to the true values (accuracy) as possible and have as large generalization capability (seeking more essential things) as possible. The XGB OST algorithm has obvious advantages in reservoir physical property parameter prediction by virtue of the characteristic of ensemble learning.
The invention provides a seismic identification method for lithology of a land-facies mixed-deposit tight reservoir, which is characterized in that an improved XGBOOST provided by the method is based on an integrated learning algorithm, a training set and a testing set are constructed by utilizing well logging data or rock physical experiment data of known lithology, XGBOOST models under different parameters are established, parameter optimization is carried out, a final XGBOOST model is constructed, and the space distribution of mixed-deposit lithology is predicted by combining an elastic parameter inversion result of pre-stack seismic data. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved.
In one embodiment, referring to fig. 2, step 100 comprises:
step 101: and acquiring the seismic elasticity parameters of the lithological work area to be identified according to the seismic data. The seismic elastic parameters include: longitudinal wave impedance and longitudinal-to-transverse wave velocity ratio.
It can be understood that the longitudinal wave impedance and the longitudinal wave velocity ratio can be obtained through seismic data or through calculation of logging data, but the logging data is often limited due to one bite, and cannot be widely applied to a target work area, and the seismic data can be widely applied to the target work area.
In one embodiment, referring to fig. 3, the method for seismic identification of lithology of a land-phase mixed rock tight reservoir further comprises:
step 300: and establishing the target lithology identification model.
Referring to fig. 4, step 300 specifically includes:
step 301: and generating lithology classification data of the tight reservoir according to at least one of drilling data, logging data and core data of the research area.
Specifically, calibrated lithology classification data are obtained from drilling data, logging or core geological description, and the aim is to construct a hybrid rock tight reservoir lithology and elasticity feature learning sample with a label. It can be understood that the well drilling data can reflect certain lithology data, such as a drilling time curve and the like, and the logging and core data can visually acquire and classify the lithology data.
Step 302: and acquiring logging data corresponding to the lithology classification data.
Step 303: generating logging elasticity parameters according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio.
Step 304: and establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm.
In steps 303 to 304, it can be understood that the log data has a disadvantage of "between holes", but has high resolution and accuracy, so that the accuracy of identifying the lithology of the geospatial mixed rock tight reservoir can be improved by selecting the longitudinal wave impedance and the longitudinal-transverse wave velocity ratio generated by the log data to generate the target lithology identification model.
In one embodiment, referring to FIG. 5, step 304 includes:
Specifically, a training set and a test set are constructed according to a certain proportion by selecting appropriate elastic parameters (such as longitudinal wave impedance, longitudinal wave velocity ratio and transverse wave velocity ratio).
It will be appreciated that the training loss parameter may be used to measure how well the model matches the training data. The positive regularization term is used to measure the complexity of the model, and others encourage the model to be simplified and more stable.
The model is trained and optimized using additive training (Boosting), starting from constant prediction, adding a new function each time. The targets for each round are: finding the appropriate parameters minimizes the objective function.
And (4) inputting the elastic parameters of the data of the test set by utilizing the model created in the last step, and predicting the reservoir physical property parameters of the test set. And comparing the prediction result with the real situation to obtain the cross correlation coefficient of the prediction result and the real situation under the current parameter.
In steps 304a to 394e, specifically, the recognition result is compared with the real lithology, so as to obtain the lithology prediction accuracy under the current parameters. The optimization objects mainly comprise the number of decision trees and the maximum depth of the trees, different models are established by changing the parameters, other conditions are constant, and the steps are repeated to obtain results under different parameters. And observing the influence of the parameters on the result to obtain the optimal parameters. And finally, combining the target lithology identification model established by the XGB learning network method with the elastic parameters to obtain the lithology prediction result of the target work area.
The invention provides a seismic identification method for lithology of a land-facies mixed-deposit tight reservoir, which is characterized in that an improved XGBOOST provided by the method is based on an integrated learning algorithm, a training set and a testing set are constructed by utilizing well logging data or rock physical experiment data of known lithology, XGBOOST models under different parameters are established, parameter optimization is carried out, a final XGBOOST model is constructed, and the space distribution of mixed-deposit lithology is predicted by combining an elastic parameter inversion result of pre-stack seismic data. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In summary, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of XGB OST is fully utilized. The XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and spatial distribution of mixed rock lithology prediction can be obtained, so that important technical support is provided for sweet spot identification and favorable zone development of a tight oil reservoir.
To further illustrate the scheme, the invention provides a specific application example of the seismic identification method for the lithology of the land-facies mixed rock compact reservoir by taking a well A1 and a well B1 of a certain oil field as an example, and the specific application example specifically comprises the following contents, and refer to fig. 6 and 7.
S0: and acquiring the longitudinal wave impedance and the longitudinal-transverse wave velocity ratio of the lithological work area to be identified according to the seismic data.
It can be understood that the longitudinal wave impedance and the longitudinal and transverse wave velocity ratio are sensitive to lithology; on the other hand, the longitudinal wave impedance and the longitudinal-transverse wave velocity ratio respectively represent the AVO intercept and the gradient response characteristics of the pre-stack seismic data and can be accurately obtained from the seismic data.
S1: and generating lithology classification data of the tight reservoir according to the core data of the research area.
In this specific application example, the lithology classification data is: cloud mudstone, cloud siltstone, muddy siltstone, yun Yan and non-reservoir rock.
S2: and acquiring logging data corresponding to the lithology classification data.
S3: and generating longitudinal wave impedance and a longitudinal-to-transverse wave velocity ratio according to the logging data.
FIGS. 8-10 are well log data (IP) of lithology, compressional impedance, and compressional-compressional velocity ratio (Vp/Vs) of a known commingled rock tight reservoir for the A1 well, respectively; fig. 11 to 13 are well log data (IP) of lithology, longitudinal wave impedance and longitudinal-to-transverse wave velocity ratio (Vp/Vs) of a known tight reservoir of a mixed rock in a B1 well, respectively. As can be seen from fig. 8 to 13, the IP data and the Vp/Vs data can well represent lithology of the facies-contaminated compact reservoir (cloud mudstone, cloud siltstone, argillaceous siltstone, yun Yan, and non-reservoir rock).
S4: and generating a training set and test data according to the logging elasticity parameters and the lithology classification data.
Selecting proper elastic parameters (such as longitudinal wave impedance, longitudinal wave velocity ratio and transverse wave velocity ratio) and constructing a training set and a test set according to a certain proportion.
S5: and generating an initial lithology recognition model according to the training loss, the positive regularization item, the objective function parameters and the training set.
First, a training loss L (θ) is defined: for a given certain sample x i ∈R d One predicted value can be obtained:
wherein,which represents the result of the prediction to be made,represents a function space containing all regression trees, and f k It is the model tree that needs to be obtained.
The training loss of the model can be used to measure the matching degree of the model and the training data, and is represented as:
wherein, y i Represents the true value of the sample, andrepresenting the predicted value of the sample. The training loss of the model is often expressed using a squared loss. The square penalty can be expressed as:
next, a regularization term Ω (θ) is defined, which can be used to control the complexity of the model:
the number and depth of nodes in the common tree; she Bichong L2 norm; the L1 norm of leaf specific gravity defines Ω.
Defining an objective function: 0bj (θ) = L (θ) + Ω (θ)
Wherein, L (theta) represents training loss and is used for measuring the matching degree of the model and the training data. It encourages the model to predict, making the results better approximate the true distribution. Ω (θ) is a regularization term used to measure the complexity of the model. It encourages model simplification and makes it more stable.
S6: and performing gradient training on the initial lithology recognition model.
The model is trained and optimized using additive training (Boosting), starting with constant prediction, adding a new function each time.
Wherein,training the model for t round, f t (x i ) The new function added for each round is also the term that needs to be decided at t rounds.
For the t-th round of prediction:
the goals of the wheel are: finding a suitable f t The objective function is minimized, and the following shows an example when the square loss is considered. When considering the square loss, the objective function can be expressed as:
according to the Taylor expansion:
wherein,
the tree is often defined by a fractional vector in a cotyledon and a cotyledon index mapping function that maps an instance to a cotyledon:
f t (x)=ω q(x) ,ω∈R T ,q:R d →{1,2,...,T}
while the complexity of a tree is generally defined as:
wherein T is the number of cotyledons;l2 norm representing cotyledon score. The values of γ and λ can be set artificially, and a larger γ indicates that a tree with a simpler structure is more desirable because the penalty for a tree with more sub leaf nodes is larger. Similarly, a larger λ is more desirable to obtain a structurally simple tree.
Define the set of instances in leaf j as: I.C. A j ={i|q(x i ) = j, recombining targets by leaves:
the objective function can be simplified to:
wherein,
assuming that the structure of the tree (q (x)) is fixed, the optimal weight for each cotyledon:
the value of the objective function is:
the smaller its score, the better the structure.
Next, a large number of trees are planted starting with a tree of depth 0 using a multi-tree learning algorithm. For each sub-leaf node of the tree, an attempt is made to add one split. The target changes after adding segmentation are:
wherein,a score representing the left branch;a score representing the right branch;represents the score when not split; gamma represents the complexity cost of introducing additional leaves.
And scanning each node, determining whether to cut after the scanning is finished, and if the two cut nodes are cut, recursively calling the cutting process to obtain a relatively good tree structure.
S7: and verifying the training result according to the test data.
And (4) inputting the elastic parameters of the data of the test set by utilizing the model created in the last step, and predicting the reservoir physical property parameters of the test set. And comparing the prediction result with the real situation to obtain the cross correlation coefficient of the prediction result and the real situation under the current parameter. And optimizing the model according to the coefficient, wherein the optimized objects are mainly the number of decision trees and the maximum depth of the trees, changing the parameter to establish different models, and repeating the steps S6-S7 under certain other conditions to obtain results under different parameters. And observing the influence of the parameters on the result to obtain optimal parameters, thereby generating a target lithology identification model.
S8: and predicting lithology of the compact reservoir of the continental facies mixed rock.
Specifically, the longitudinal wave impedance and the longitudinal wave velocity ratio of the lithologic work area to be identified, which are obtained in the step S0, are input into the target lithologic identification model generated in the step S11, and the lithologic property of the facies mixed rock tight reservoir of the lithologic work area to be identified is generated.
Fig. 14 and 15 show the actual lithology of the A1 well and the B1 well, respectively, and fig. 16 and 17 show the lithology prediction results obtained by the method described in the present embodiment. The number of the decision trees is 500, 50% of the decision trees are randomly selected as training data, all samples are input for testing, and the lithological prediction accuracy of the two wells is 90.47% and 85.80% respectively. In addition, the lithology prediction result also shows that the XBBOOST algorithm improved by the specific application entity has high data expression capacity and generalization performance.
Table 1 compares the lithology prediction effects of various machine learning algorithms for the same set of well log data. It can be seen that the result of the XGBOOST algorithm improved by the present specific application example is significantly better than other currently used machine learning algorithms for lithology prediction, and the essence is to integrate the learning idea and integrate a plurality of weak classifiers (decision trees) to form a strong classifier. The XGB OST algorithm is distinguished from the lithology prediction problem of the compact oil reservoir of the mixed rock, and the relationship between the lithology and the elastic parameter of the mixed rock is a complex and nonlinear mapping relationship, so that the problems of overfitting, difficult parameter optimization, less dependence on experience, insufficient sample amount and the like are brought to other machine learning algorithms. The XGBOOST greatly reduces the error rate of lithology prediction by virtue of the integration, can process input samples with high-dimensional features, can evaluate the importance of each feature on the classification problem, and does not need to carry out a large amount of parameter debugging work.
TABLE 1 comparison table of results of lithology prediction by multiple machine learning algorithms
FIG. 18 is a lithology seismic prediction result of two sets of desserts in compact oil storage of a mixed-reservoir rock obtained by applying an XGBOOST network established by using learning samples of two wells (A1 and B1) to an prestack elastic parameter inversion result. It can be seen that the lithology of the seismic identification is well matched with the lithology calibrated by well logging at the positions of the two wells. Meanwhile, the lithology prediction of the tight reservoir of the mixed rock can effectively describe the spatial distribution of lithology, and provides technical support for identifying the high-quality reservoir of the tight oil, modeling geology and selecting favorable sections for well drilling and fracturing.
The invention provides a seismic identification method for lithology of a continental facies mixed rock tight reservoir, which is characterized in that an improved XGBOST provided by the method is based on an integrated learning algorithm, a training set and a testing set are constructed by utilizing well logging data or rock physical experiment data of known lithology, XGBOST models under different parameters are established, parameter optimization is carried out, a final XGBOST model is constructed, and the elastic parameter inversion result of pre-stack seismic data is combined to predict the spatial distribution of the lithology of mixed rock. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In conclusion, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of XGB OST is fully utilized. The XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and spatial distribution of mixed rock lithology prediction can be obtained, so that important technical support is provided for sweet spot identification and favorable zone development of a tight oil reservoir.
Based on the same inventive concept, the embodiment of the application also provides a seismic recognition device for lithology of a land-facies mixed rock tight reservoir, which can be used for realizing the method described in the embodiment, such as the following embodiment. The principle of solving the problems of the seismic recognition device for the lithology of the land-facies mixed laminated rock tight reservoir is similar to that of the seismic recognition method for the lithology of the land-facies mixed laminated rock tight reservoir, so the implementation of the seismic recognition device for the lithology of the land-facies mixed laminated rock tight reservoir can be referred to the implementation of the seismic recognition method for the lithology of the land-facies mixed laminated rock tight reservoir, and repeated parts are not repeated. As used hereinafter, the term "unit" or "module" may be a combination of software and/or hardware that implements a predetermined function. While the system described in the embodiments below is preferably implemented in software, implementations in hardware, or a combination of software and hardware are also possible and contemplated.
The embodiment of the invention provides a specific implementation mode of a seismic recognition device for the lithology of a land-facies mixed laminated rock tight reservoir, which can realize the seismic recognition method for the lithology of the land-facies mixed laminated rock tight reservoir, and referring to fig. 19, the seismic recognition device for the lithology of the land-facies mixed laminated rock tight reservoir specifically comprises the following contents:
the earthquake elastic parameter acquisition unit 10 is used for acquiring earthquake elastic parameters of a lithologic work area to be identified;
and the lithology identification unit 20 is configured to perform lithology identification on the land-facies mixed rock tight reservoir of the lithology work area to be identified by using the elastic parameters and a pre-established target lithology identification model.
In one embodiment, referring to fig. 20, the seismic elastic parameter acquisition unit 10 includes: the seismic elastic parameter obtaining module 101 is configured to obtain a seismic elastic parameter of a lithologic work area to be identified according to seismic data, where the seismic elastic parameter includes: longitudinal wave impedance and longitudinal-transverse wave velocity ratio.
In one embodiment, referring to fig. 21, the seismic identification device for lithology of a land-phase mixed rock tight reservoir further includes: the object model building unit 30, see fig. 22, the object model building unit 30 comprises:
the lithology selection module 301 is configured to generate lithology classification data of the tight reservoir according to at least one of drilling data, logging data and core data of a research area;
a logging data obtaining module 302, configured to obtain logging data corresponding to the lithology classification data;
a logging elastic parameter generating module 303, configured to generate a logging elastic parameter according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio;
and the target model establishing module 304 is configured to establish the target lithology identification model according to the well logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm.
In one embodiment, referring to fig. 23, the object model building module 304 comprises:
the training set generation module 304a generates a training set and test data according to the logging elasticity parameters and the lithology classification data;
an initial model generation module 304b, configured to generate an initial lithology recognition model according to the training loss, the positive regularization term, the objective function parameter, and the training set;
a result generating module 304c, configured to train the initial lithology recognition model by using an addition training algorithm and a multi-tree learning algorithm to generate a training result;
a result verification module 304d, configured to verify the training result according to the test data;
and the target model generation module 304e is configured to change the number and the maximum depth of the decision trees in the initial lithology recognition model according to the verification result to generate the target lithology recognition model.
The invention provides a seismic recognition device for lithology of a land-facies mixed-deposit tight reservoir, which is characterized in that an improved XGBOOST provided by the method is based on an integrated learning algorithm, a training set and a testing set are constructed by utilizing well logging data or rock physical experiment data of known lithology, XGBOOST models under different parameters are established, parameter optimization is carried out, a final XGBOOST model is constructed, and the space distribution of mixed-deposit lithology is predicted by combining an elastic parameter inversion result of pre-stack seismic data. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In summary, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of XGB OST is fully utilized. The XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and the spatial distribution of the lithology prediction of the mixed-deposit rock can be obtained, so that important technical support is provided for dessert identification and favorable zone development optimization of the compact oil reservoir.
The embodiment of the present application further provides a specific implementation manner of an electronic device capable of implementing all steps in the method for identifying lithology of a land-facies mixed deposit tight reservoir in the foregoing embodiment, and referring to fig. 24, the electronic device specifically includes the following contents:
a processor (processor) 1201, a memory (memory) 1202, a communication Interface (Communications Interface) 1203, and a bus 1204;
the processor 1201, the memory 1202 and the communication interface 1203 complete communication with each other through the bus 1204; the communication interface 1203 is configured to implement information transmission between related devices such as a server-side device, an acquisition device, and a client device.
The processor 1201 is configured to call the computer program in the memory 1202, and the processor executes the computer program to implement all the steps of the method for seismic identification of lithology of a continental facies mixed-stratigraphic tight reservoir in the above-mentioned embodiment, for example, the processor executes the computer program to implement the following steps:
step 100: and acquiring the seismic elasticity parameters of the lithologic work area to be identified.
Step 200: and performing lithology recognition on the continental facies mixed-deposit rock tight reservoir of the lithology work area to be recognized by using the elastic parameters and a pre-established target lithology recognition model.
According to the electronic equipment in the embodiment of the application, the improved XGB OST provided by the method is used for constructing a training set and a testing set by using well logging data or rock physical experiment data of known lithology based on an ensemble learning algorithm, XGB OST models under different parameters are established, parameter optimization is carried out, a final XGB OST model is established, and the elastic parameter inversion result of pre-stack seismic data is combined to predict the spatial distribution of the mixed-accumulation lithology. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic recognition of the land-facies mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In summary, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of XGB OST is fully utilized. The XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and the spatial distribution of the lithology prediction of the mixed-deposit rock can be obtained, so that important technical support is provided for dessert identification and favorable zone development optimization of the compact oil reservoir.
Embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps in the seismic identification method of the lithology of the continental facies mixed laminated rock tight reservoir in the above embodiments, where the computer-readable storage medium has stored thereon a computer program, and the computer program, when executed by a processor, implements all the steps of the seismic identification method of the lithology of the continental facies mixed laminated rock tight reservoir in the above embodiments, for example, when the processor executes the computer program, implements the following steps:
step 100: and acquiring the seismic elasticity parameters of the lithological work area to be identified.
Step 200: and performing lithology identification on the continental facies mixed rock tight reservoir of the lithology work area to be identified by using the elastic parameters and a pre-established target lithology identification model.
The computer-readable storage medium in the embodiment of the application builds a training set and a testing set by using the improved XGBOST provided by the method based on an ensemble learning algorithm and using well logging data or rock physical experiment data of known lithology, builds XGBOST models under different parameters, optimizes the parameters, builds a final XGBOST model, and predicts the spatial distribution of the lithology of the mixed rock by combining the elastic parameter inversion result of the seismic data before stacking. Compared with other machine learning algorithms, the prediction accuracy is greatly improved, effective support can be provided for lithologic seismic identification of the land-phase mixed rock compact oil reservoir, and the problems of overfitting, difficult parameter optimization, less experience dependence, insufficient sample size and the like of other machine learning algorithms are solved. In summary, the method has the following advantages:
1) Aiming at the characteristic that the lithology-seismic elasticity characteristic mapping relation of the continental facies mixed rock tight oil reservoir is fuzzy, the idea of XGB OST is fully utilized. The XGBOST model based on the XGBOST model has the advantages of regularization, high flexibility, missing value processing, pruning after splitting, built-in cross validation, continuous training on the basis of the existing model and the like, is different from other learning methods, and is more suitable for training and simulating geophysical reservoir parameters with complex data.
2) The prestack seismic inversion result is combined with a machine learning network based on logging data, and the spatial distribution of the lithology prediction of the mixed-deposit rock can be obtained, so that important technical support is provided for dessert identification and favorable zone development optimization of the compact oil reservoir.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the hardware + program class embodiment, since it is substantially similar to the method embodiment, the description is simple, and the relevant points can be referred to the partial description of the method embodiment.
The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
Although the present application provides method steps as described in an embodiment or flowchart, additional or fewer steps may be included based on conventional or non-inventive efforts. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When an actual apparatus or client product executes, it may execute sequentially or in parallel (e.g., in the context of parallel processors or multi-threaded processing) according to the embodiments or methods shown in the figures.
Although embodiments of the present description provide method steps as described in embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When an actual apparatus or end product executes, it may execute sequentially or in parallel (e.g., parallel processors or multi-threaded environments, or even distributed data processing environments) according to the method shown in the embodiment or the figures. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in a process, method, article, or apparatus that comprises the recited elements is not excluded.
For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, in implementing the embodiments of the present description, the functions of each module may be implemented in one or more software and/or hardware, or a module implementing the same function may be implemented by a combination of multiple sub-modules or sub-units, and the like. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of logical functional division, and other divisions may be realized in practice, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may therefore be considered as a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be conceived to be both a software module implementing the method and a structure within a hardware component.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein.
The embodiments of this specification may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The described embodiments may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment. In the description of the specification, reference to the description of "one embodiment," "some embodiments," "an example," "a specific example," or "some examples" or the like means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the embodiments of the specification. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
The above description is only an example of the embodiments of the present disclosure, and is not intended to limit the embodiments of the present disclosure. Various modifications and variations to the embodiments described herein will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the present specification should be included in the scope of the claims of the embodiments of the present specification.
Claims (6)
1. A seismic identification method for lithology of a land-facies mixed rock tight reservoir is characterized by comprising the following steps:
acquiring earthquake elastic parameters of a lithologic work area to be identified;
performing lithology recognition on the continental facies mixed rock tight reservoir of the lithology work area to be recognized by using the elastic parameters and a pre-established target lithology recognition model;
the step of establishing the target lithology identification model comprises the following steps:
generating lithology classification data of the tight reservoir according to at least one of drilling data, logging data and core data of a research area;
acquiring logging data corresponding to the lithologic classification data;
generating logging elastic parameters according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio;
establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm;
establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm, wherein the method comprises the following steps:
generating a training set and test data according to the logging elasticity parameters and the lithology classification data;
generating an initial lithology recognition model according to the training loss, the positive regularization item, the objective function parameter and the training set;
training the initial lithology recognition model by utilizing an addition training algorithm and a multi-tree learning algorithm to generate a training result;
verifying the training result according to the test data;
changing the number and the maximum depth of decision trees in the initial lithology recognition model according to a verification result to generate the target lithology recognition model;
the generating an initial lithology recognition model according to the training loss, the positive regularization item, the objective function parameter and the training set comprises:
first, the training loss L (Θ) is defined: for a given certain sample x i ∈R d One predicted value can be obtained:
wherein,which represents the result of the prediction to be made,represents a function space containing all regression trees, and f k It is the model tree that needs to be obtained, K represents the number of the model tree, R d Represents a d-dimensional real vector space;
the training loss of the model can be used to measure the matching degree of the model and the training data, and is represented as:
wherein, y i Represents the true value of the ith sample, andrepresenting the predicted value of the ith sample, wherein n is the number of the samples; expressing the training loss of the model by using the square loss; the square penalty can be expressed as:
next, a regularization term Ω (Θ) is defined, which can be used to control the complexity of the model:
defining omega by using the number of nodes in the tree, the depth, the L2 norm of the leaf specific gravity and the L1 norm of the leaf specific gravity;
defining an objective function: obj (Θ) = L (Θ) + Ω (Θ)
Wherein, L (theta) represents training loss and is used for measuring the matching degree of the model and the training data; the model can be encouraged to predict, so that the result is better close to the real distribution; Ω (Θ) is a regularization term used to measure the complexity of the model; it encourages the model to be simplified, making it more stable;
the training the initial lithology recognition model by using an additive training algorithm and a multi-tree learning algorithm to generate a training result, comprising:
training and optimizing the initial lithology recognition model by using additive training, and starting from constant prediction, adding a new function each time:
……
wherein,training the model for t round, f t (x i ) For each oneNew functions added in turns are also terms needed to be determined in the t turns; wherein:
for the prediction of the t-th round:
the goals of the wheel are: find suitable f t The objective function is minimized, and when the squared loss is considered, the objective function can be expressed as:
according to the Taylor expansion:
wherein,
the tree is defined by a fractional vector in the cotyledon and the cotyledon index mapping function, which maps an instance to a cotyledon:
f t (x)=ω q(x) ,ω∈R T ,q:R d →{1,2,…,T}
the complexity of a tree is generally defined as:
wherein T is the number of cotyledons;an L2 norm representing a cotyledon score; λ represents the punishment strength of pair She QuanDegree;
define the set of instances in leaf j as: i is j ={i|q(x i ) = j, recombining the objective functions by leaves:
the objective function is simplified to:
wherein,
assuming that the structure q (x) of the tree is fixed, the optimal weight of each cotyledon:
the value of the objective function is:
then, starting from the tree with depth 0, for each sub-leaf node of the tree, add a split, and add the target changes after splitting:
wherein,a score representing the left branch;a score representing the right branch;represents the score when not split; γ represents the complexity cost of introducing additional leaves;
scanning each node, determining whether to cut after the scanning is finished, and if so, recursively calling the cutting process to the two cut nodes to generate the training result;
the lithology includes: dolomites, cloudy sandstones, siliceous sandstones, carbonaceous mudstones, and cloudy mudstones.
2. The seismic identification method of claim 1, wherein the obtaining of the seismic elasticity parameters of the lithological work area to be identified comprises: acquiring seismic elasticity parameters of a lithologic work area to be identified according to seismic data, wherein the seismic elasticity parameters comprise: longitudinal wave impedance and longitudinal-to-transverse wave velocity ratio.
3. The utility model provides a seismic recognition device of land facies mixed deposition rock tight reservoir lithology which characterized in that includes:
the earthquake elastic parameter acquisition unit is used for acquiring earthquake elastic parameters of the lithological work area to be identified;
the lithology recognition unit is used for carrying out lithology recognition on the land-facies mixed rock compact reservoir of the lithology work area to be recognized by utilizing the elastic parameters and a pre-established target lithology recognition model;
an object model building unit comprising:
the lithology selection module is used for generating lithology classification data of the tight reservoir according to at least one of drilling data, logging data and core data of a research area;
the logging data acquisition module is used for acquiring logging data corresponding to the lithology classification data;
the logging elastic parameter generating module is used for generating logging elastic parameters according to the logging data; the well logging elasticity parameters include: longitudinal wave impedance and longitudinal-transverse wave velocity ratio;
the target model establishing module is used for establishing the target lithology identification model according to the logging elastic parameters and the lithology classification data by using an extreme gradient lifting algorithm;
the object model building module comprises:
the training set generation module generates a training set and test data according to the logging elasticity parameters and the lithology classification data;
the initial model generation module is used for generating an initial lithology recognition model according to the training loss, the positive regularization item, the target function parameter and the training set;
the result generation module is used for training the initial lithology recognition model by utilizing an addition training algorithm and a multi-tree learning algorithm so as to generate a training result;
the result verification module is used for verifying the training result according to the test data;
the target model generation module is used for changing the number and the maximum depth of the decision trees in the initial lithology recognition model according to a verification result so as to generate the target lithology recognition model;
the initial model generation module is specifically configured to:
first, the training loss L (Θ) is defined: for a given certain sample x i ∈R d One predicted value can be obtained:
wherein,which represents the result of the prediction to be made,represents a function space containing all regression trees, and f k It is the model tree that needs to be obtained, K represents the number of the model tree, R d Represents a d-dimensional real vector space;
the training loss of the model can be used to measure the matching degree of the model and the training data, and is represented as:
wherein, y i Represents the true value of the ith sample, andrepresenting the predicted value of the ith sample, wherein n is the number of the samples; expressing the training loss of the model by using the square loss; the square penalty can be expressed as:
next, a regularization term Ω (Θ) is defined, which can be used to control the complexity of the model:
defining omega by using the node number, the depth, the L2 norm of the leaf specific gravity and the L1 norm of the leaf specific gravity in the tree;
defining an objective function: obj (Θ) = L (Θ) + Ω (Θ)
Wherein, L (theta) represents training loss and is used for measuring the matching degree of the model and the training data; the model can be encouraged to predict, so that the result is better close to the real distribution; Ω (Θ) is a regularization term used to measure the complexity of the model; it encourages the model to simplify and make it more stable;
the training the initial lithology recognition model by using an additive training algorithm and a multi-tree learning algorithm to generate a training result, comprising:
training and optimizing the initial lithology recognition model by using additive training, and starting from constant prediction, adding a new function each time:
……
wherein,training the model for t round, f t (x i ) A new function added for each round is also a term needing to be determined in the t round; wherein:
for the prediction of the t-th round:
the goals of the wheel are: finding a suitable f t The objective function is minimized, and when the square loss is considered, the objective function can be expressed as:
according to the Taylor expansion:
wherein,
the tree is defined by a fractional vector in a cotyledon and a cotyledon index mapping function that maps an instance to a cotyledon:
f t (x)=ω q(x) ,ω∈R T ,q:R d →{1,2,…,T}
the complexity of a tree is generally defined as:
define the set of instances in leaf j as: i is j ={i|q(x i ) = j, recombining the objective functions by leaves:
the objective function is simplified to:
wherein,
assuming that the structure q (x) of the tree is fixed, the optimal weight of each cotyledon:
the value of the objective function is:
then, starting from the tree with depth 0, for each sub-leaf node of the tree, add a split, and add the target changes after splitting:
wherein,a score representing the left branch;a score representing the right branch;represents the score when not split; gamma represents the complexity cost of introducing additional leaves;
Scanning each node, determining whether to cut after the scanning is finished, and if so, recursively calling the cutting process to the two cut nodes to generate the training result;
the lithology includes: dolomites, cloudy sandstones, siliceous sandstones, carbonaceous mudstones, and cloudy mudstones.
4. The seismic recognition device of claim 3, wherein the seismic elastic parameter acquisition unit comprises: the seismic elastic parameter acquisition module is used for acquiring seismic elastic parameters of the lithological work area to be identified according to the seismic data, and the seismic elastic parameters comprise: longitudinal wave impedance and longitudinal-transverse wave velocity ratio.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program performs the steps of the method for seismic identification of tight reservoir lithology of a continental-phase mixed rock according to any of claims 1 to 2.
6. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method for seismic identification of lithology of a land-mixed compact reservoir as defined in any one of claims 1 to 2.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911051312.4A CN110837115B (en) | 2019-10-31 | 2019-10-31 | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911051312.4A CN110837115B (en) | 2019-10-31 | 2019-10-31 | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110837115A CN110837115A (en) | 2020-02-25 |
CN110837115B true CN110837115B (en) | 2023-02-28 |
Family
ID=69575924
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911051312.4A Active CN110837115B (en) | 2019-10-31 | 2019-10-31 | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110837115B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111753871B (en) * | 2020-05-11 | 2022-12-16 | 同济大学 | Lithology and fluid type identification method based on XGB OST algorithm and feature engineering |
CN113960656B (en) * | 2020-07-20 | 2023-07-25 | 中国石油天然气股份有限公司 | Method and device for identifying morphological characteristics of seismic data target |
CN112507615B (en) * | 2020-12-01 | 2022-04-22 | 西南石油大学 | Intelligent identification and visualization method for lithofacies of continental tight reservoir |
CN117250658B (en) * | 2023-11-17 | 2024-02-09 | 核工业北京地质研究院 | Method for creating seismic dataset of investigation region |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106168676A (en) * | 2015-05-22 | 2016-11-30 | 中国石油化工股份有限公司 | Formation lithology based on seismic data and Fluid Identification Method and device |
CN109345007A (en) * | 2018-09-13 | 2019-02-15 | 中国石油大学(华东) | A kind of Favorable Reservoir development area prediction technique based on XGBoost feature selecting |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9447681B2 (en) * | 2011-09-26 | 2016-09-20 | Saudi Arabian Oil Company | Apparatus, program product, and methods of evaluating rock properties while drilling using downhole acoustic sensors and a downhole broadband transmitting system |
CN108830421B (en) * | 2018-06-21 | 2022-05-06 | 中国石油大学(北京) | Method and device for predicting gas content of tight sandstone reservoir |
-
2019
- 2019-10-31 CN CN201911051312.4A patent/CN110837115B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106168676A (en) * | 2015-05-22 | 2016-11-30 | 中国石油化工股份有限公司 | Formation lithology based on seismic data and Fluid Identification Method and device |
CN109345007A (en) * | 2018-09-13 | 2019-02-15 | 中国石油大学(华东) | A kind of Favorable Reservoir development area prediction technique based on XGBoost feature selecting |
Non-Patent Citations (3)
Title |
---|
XGBoost算法在致密砂岩气储层测井解释中的应用;闫星宇等;《石油地球物理勘探》;20190430;第54卷(第2期);第448-452页 * |
晏信飞等.致密砂岩储层贝叶斯岩性判别与孔隙流体检测.《石油地球物理勘探》.2012,第47卷(第6期),第946-949页. * |
致密砂岩储层贝叶斯岩性判别与孔隙流体检测;晏信飞等;《石油地球物理勘探》;20121231;第47卷(第6期);第946-949页 * |
Also Published As
Publication number | Publication date |
---|---|
CN110837115A (en) | 2020-02-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110837115B (en) | Seismic identification method and device for lithology of land-facies mixed rock compact reservoir | |
Roy et al. | Generative topographic mapping for seismic facies estimation of a carbonate wash, Veracruz Basin, southern Mexico | |
CN102713683B (en) | Method and system reservoir characteristics being moved towards to modeling is simulated for utilizing multi-point statistic | |
CN104101902B (en) | seismic attribute clustering method and device | |
CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
CN103775075A (en) | Method for identifying lithology of whole well section | |
Bagheripour et al. | Fuzzy ruling between core porosity and petrophysical logs: Subtractive clustering vs. genetic algorithm–pattern search | |
CN110927793B (en) | Reservoir prediction method and system based on sequential random fuzzy simulation | |
CN109113729A (en) | Lithology Identification Methods and device based on log | |
CN108680954A (en) | Frequency domain multi-data body time varying window waveform clustering method and device | |
CN109613623A (en) | A kind of lithology prediction method based on residual error network | |
Korjani et al. | Reservoir characterization using fuzzy kriging and deep learning neural networks | |
Al-Mudhafar et al. | Tree-based ensemble algorithms for lithofacies classification and permeability prediction in heterogeneous carbonate reservoirs | |
Alameedy et al. | Evaluating machine learning techniques for carbonate formation permeability prediction using well log data | |
CN113033637B (en) | Integrated learning lithofacies prediction method based on rock physical sample enhancement | |
You et al. | Explainable convolutional neural networks driven knowledge mining for seismic facies classification | |
CN111894563A (en) | Method and system for determining classification of fractured reservoir in buried hill section | |
Narayan et al. | Accuracy assessment of various supervised machine learning algorithms in litho-facies classification from seismic data in the Penobscot field, Scotian Basin | |
Nguyen et al. | Application of random forest algorithm to predict lithofacies from well and seismic data in Balder field, Norwegian North Sea | |
Bubnova et al. | Automatic determination of sedimentary units from well data | |
Malvić et al. | Neural networks in petroleum geology as interpretation tools | |
Mollajan et al. | Fuzzy classifier fusion: an application to reservoir facies identification | |
Abadpour et al. | Characterization and uncertainty reduction on Facies distribution and probability cubes with ensemble Kalman filter history matching | |
Chen et al. | A geospatial case‐based reasoning model for oil–gas reservoir evaluation | |
Gao et al. | Seismic predictions of fluids via supervised deep learning: Incorporating various class-rebalance strategies |
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 |