CN113344359A - Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest - Google Patents
Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest Download PDFInfo
- Publication number
- CN113344359A CN113344359A CN202110599248.4A CN202110599248A CN113344359A CN 113344359 A CN113344359 A CN 113344359A CN 202110599248 A CN202110599248 A CN 202110599248A CN 113344359 A CN113344359 A CN 113344359A
- Authority
- CN
- China
- Prior art keywords
- random forest
- influencing
- main control
- factors
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000003066 decision tree Methods 0.000 claims abstract description 45
- 238000011158 quantitative evaluation Methods 0.000 claims abstract description 15
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 11
- 238000005070 sampling Methods 0.000 claims abstract description 11
- 230000001419 dependent effect Effects 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 4
- 239000011148 porous material Substances 0.000 claims description 40
- 239000002245 particle Substances 0.000 claims description 17
- 238000004458 analytical method Methods 0.000 claims description 14
- 230000011218 segmentation Effects 0.000 claims description 14
- 229910052500 inorganic mineral Inorganic materials 0.000 claims description 13
- 239000011707 mineral Substances 0.000 claims description 13
- 238000012512 characterization method Methods 0.000 claims description 11
- 239000004568 cement Substances 0.000 claims description 8
- 238000005056 compaction Methods 0.000 claims description 4
- 239000012535 impurity Substances 0.000 claims description 4
- 206010044625 Trichorrhexis Diseases 0.000 claims 1
- 238000011161 development Methods 0.000 abstract description 14
- 230000000694 effects Effects 0.000 abstract description 5
- 230000006872 improvement Effects 0.000 abstract description 2
- 239000007789 gas Substances 0.000 description 22
- 238000004519 manufacturing process Methods 0.000 description 18
- 230000018109 developmental process Effects 0.000 description 13
- 230000035699 permeability Effects 0.000 description 10
- 238000011160 research Methods 0.000 description 10
- 238000012360 testing method Methods 0.000 description 10
- 238000010276 construction Methods 0.000 description 8
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000008021 deposition Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 239000010453 quartz Substances 0.000 description 4
- 239000011435 rock Substances 0.000 description 4
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N silicon dioxide Inorganic materials O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 3
- 238000005520 cutting process Methods 0.000 description 3
- ZOOODBUHSVUZEM-UHFFFAOYSA-N ethoxymethanedithioic acid Chemical group CCOC(S)=S ZOOODBUHSVUZEM-UHFFFAOYSA-N 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 239000010433 feldspar Substances 0.000 description 3
- 229910052742 iron Inorganic materials 0.000 description 3
- 238000005192 partition Methods 0.000 description 3
- 238000004062 sedimentation Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 2
- DLHONNLASJQAHX-UHFFFAOYSA-N aluminum;potassium;oxygen(2-);silicon(4+) Chemical compound [O-2].[O-2].[O-2].[O-2].[O-2].[O-2].[O-2].[O-2].[Al+3].[Si+4].[Si+4].[Si+4].[K+] DLHONNLASJQAHX-UHFFFAOYSA-N 0.000 description 2
- 239000011575 calcium Substances 0.000 description 2
- 229910052791 calcium Inorganic materials 0.000 description 2
- 238000005266 casting Methods 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000004090 dissolution Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 229910052900 illite Inorganic materials 0.000 description 2
- NLYAJNPCOHFWQQ-UHFFFAOYSA-N kaolin Chemical compound O.O.O=[Al]O[Si](=O)O[Si](=O)O[Al]=O NLYAJNPCOHFWQQ-UHFFFAOYSA-N 0.000 description 2
- 229910052622 kaolinite Inorganic materials 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- VGIBGUSAECPPNB-UHFFFAOYSA-L nonaaluminum;magnesium;tripotassium;1,3-dioxido-2,4,5-trioxa-1,3-disilabicyclo[1.1.1]pentane;iron(2+);oxygen(2-);fluoride;hydroxide Chemical compound [OH-].[O-2].[O-2].[O-2].[O-2].[O-2].[F-].[Mg+2].[Al+3].[Al+3].[Al+3].[Al+3].[Al+3].[Al+3].[Al+3].[Al+3].[Al+3].[K+].[K+].[K+].[Fe+2].O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2.O1[Si]2([O-])O[Si]1([O-])O2 VGIBGUSAECPPNB-UHFFFAOYSA-L 0.000 description 2
- 229910052655 plagioclase feldspar Inorganic materials 0.000 description 2
- 238000003908 quality control method Methods 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 239000012991 xanthate Substances 0.000 description 2
- 229910021532 Calcite Inorganic materials 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000003556 assay Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002734 clay mineral Substances 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 229910000514 dolomite Inorganic materials 0.000 description 1
- 239000010459 dolomite Substances 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001125 extrusion Methods 0.000 description 1
- 238000000227 grinding Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 229910052901 montmorillonite Inorganic materials 0.000 description 1
- 239000003345 natural gas Substances 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000012764 semi-quantitative analysis Methods 0.000 description 1
- 229910021647 smectite Inorganic materials 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V9/00—Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Software Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Geophysics (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- General Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明公开了一种基于随机森林的致密砂岩气储层质量主控因素定量评价方法，包括以下步骤：S1：收集研究区影响储层质量的相关影响因素，并根据所述影响因素的参数类型进行参数处理；S2：将处理后的参数保存为逗号分隔值文件；S3：以储层质量的表征参数作为因变量，以所述影响因素作为自变量；S4：根据所述因变量和所述自变量，利用随机森林算法，采用有放回的随机抽样方式抽取训练数据，构建决策树和随机森林；S5：针对各影响因素，计算所述决策树的袋外数据误差，根据所述袋外数据误差优选出主控影响因素。本发明能够利用袋外数据误差实现对致密砂岩气储层质量主控因素的定量评价，为下一步油田开发效果改善、气藏描述提供可参考的地质依据。
The invention discloses a method for quantitative evaluation of the main control factors of tight sandstone gas reservoir quality based on random forest, comprising the following steps: S1: collecting relevant influencing factors affecting the quality of the reservoir in a study area, and determining the parameters according to the parameter types of the influencing factors. Carry out parameter processing; S2: save the processed parameters as a comma-separated value file; S3: take the parameter representing the reservoir quality as the dependent variable, and take the influencing factor as the independent variable; S4: according to the dependent variable and the For independent variables, random forest algorithm is used to extract training data by random sampling with replacement, and a decision tree and random forest are constructed; S5: For each influencing factor, calculate the out-of-bag data error of the decision tree, according to the out-of-bag data error. Data errors are selected to select the main control influencing factors. The invention can realize quantitative evaluation of the main control factors of tight sandstone gas reservoir quality by utilizing the data error outside the bag, and provide a reference geological basis for the improvement of the oilfield development effect and the gas reservoir description in the next step.
Description
Technical Field
The invention relates to the technical field of unconventional oil and gas reservoir development, in particular to a method for quantitatively evaluating main control factors of the quality of a tight sandstone gas reservoir based on a random forest.
Background
At the present stage, the world enters the conventional oil gas stable production-up and unconventional oil gas rapid development stage, and the compact sandstone gas becomes the key direction of unconventional natural gas development. A large number of development practices show that effective evaluation of the main quality control factors of the tight sandstone gas reservoir is a key basic problem for realizing large-scale and profitable development of the gas reservoir.
The current methods for evaluating the reservoir quality master control factors include the following categories: firstly, carrying out compact sandstone quality evaluation on a reservoir by using sedimentology and logging rock physics; secondly, reservoir quality is researched and distinguished by using the modes of core observation, slice identification, physical property analysis and the like; thirdly, fully combining the data of well logging, core observation, analysis and assay, oil testing and production trial, and the like, establishing a reservoir evaluation index system, and comprehensively analyzing reservoir quality influence factors; and fourthly, establishing a reservoir evaluation method and criterion by combining logging data and adopting a numerical simulation mode based on reservoir characteristics.
The method is mainly based on single-factor qualitative or semi-quantitative evaluation constrained by artificial experience, however, development proves that the quality of the tight sandstone reservoir is jointly controlled by multiple factors such as deposition, diagenesis, construction and the like. The existing method mainly focuses on qualitative or semi-quantitative analysis of single factor parameters, and the evaluation results of the method have the problems of subjectivity, non-universality and the like, so that the quality definition of a compact sandstone gas reservoir is unclear and inaccurate, the design of a development scheme and well location deployment are influenced, and particularly in a reservoir with serious heterogeneity, the permeability cannot be calculated with high precision through a pore structure and the radius of a pore throat. Therefore, it is highly desirable to provide a quantitative evaluation method of multifactorial linkage.
Disclosure of Invention
In order to solve the problems, the invention aims to provide a method for quantitatively evaluating the main control factors of the quality of the compact sandstone gas reservoir based on a random forest.
The technical scheme of the invention is as follows:
a method for quantitatively evaluating the main control factors of the quality of a tight sandstone gas reservoir based on a random forest comprises the following steps:
s1: collecting relevant influence factors influencing reservoir quality in a research area, and performing parameter processing according to the parameter types of the influence factors:
if the influence factors are data parameters, checking whether the data parameters are missing, and if a certain data parameter is missing, removing all the influence factors of the same batch; if no deletion exists, reserving;
if the influence factor is a text parameter, carrying out assignment processing on the text parameter;
s2: storing the processed parameters as comma separated value files;
s3: taking the characterization parameters of the reservoir quality as dependent variables of the reservoir quality analysis, and taking the influencing factors as independent variables of the reservoir quality analysis;
s4: extracting training data by using a random forest algorithm and a replaced random sampling mode according to the dependent variable and the independent variable to construct a decision tree and a random forest;
s5: and calculating the error of the data outside the bag of the decision tree aiming at each influence factor, and preferably selecting the main control influence factor according to the error of the data outside the bag.
Preferably, in step S1, the relevant influencing factors include a sedimentation-type influencing factor, a diagenesis-type influencing factor, and a construction-type influencing factor.
Preferably, the sedimentation influencing factors comprise granularity lithology type, mineral lithology type, sorting property, roundness grinding, various granularity parameters, different types of mineral particle content, impurity base content and primary pore content; the diagenetic influence factors comprise cementing action types, different types of cementing material contents, different types of dissolving hole contents, different types of alternate mineral contents and compaction strength; the construction type influence factors comprise fracture types and different types of fracture contents.
Preferably, in step S3, the characterization parameter of the reservoir quality is a flow cell index.
Preferably, step S4 specifically includes the following sub-steps:
s41: by using a random forest algorithm, taking the processed data set as an input sample data set, randomly extracting the data set for multiple times to form subset data in a replacement mode, wherein the sampling times are consistent with the number of samples, and the subset data obtained by sampling is used for constructing a decision tree;
s42: randomly extracting part of influence factors from the sub-data set to form a candidate partition set, selecting one influence factor from the candidate partition set as a partition point of the decision tree according to a minimum node purity principle, continuing splitting by adopting the principle until all samples of the node reach leaf nodes, and finishing splitting;
s43: and repeating the step S41 and the step S42 to establish a plurality of decision trees, and forming the random forest by the decision trees.
Preferably, step S5 specifically includes the following sub-steps:
s51: putting all influence factors in the data outside the bags into the constructed random forest, and calculating the predicted value of each data outside the bags through the random forest aiming at a certain decision tree;
s52: calculating a mean square error I between a predicted value and a true value of each data outside the bag;
s53: selecting a certain influence factor in the data outside the bag, randomly adding noise into the influence factor, then placing the influence factor into the random forest, and calculating to obtain a predicted value after the noise is added; calculating a mean square error II between a predicted value and a true value of the influencing factor with noise;
s54: judging the importance of the influence factor according to the magnitude of the first mean square error and the second mean square error of the influence factor:
if the second mean square error is larger than the first mean square error after random noise is added, and the difference value between the second mean square error and the first mean square error is larger than a threshold value, the influence factor is important, otherwise, the influence factor is unimportant;
s55: repeating the step S53 and the step S54, and judging all the remaining influence factors in the data outside the bag;
s56: and (4) repeating the step S51 and the step S55 aiming at each remaining decision tree in the random forest, calculating the out-of-bag data error of each influence factor, taking the average value of the out-of-bag data errors as the importance value of each influence factor, arranging the importance values in a descending order according to the size of the average value, wherein the influence factors in the top order are the main control factors influencing the reservoir quality.
Preferably, the method further comprises the following steps:
s6: reconstructing a decision tree and a random forest according to the main control influence factors selected by the out-of-bag data errors;
s7: calculating the out-of-bag data errors of the optimal main control influence factors, and taking the average value of the out-of-bag data errors of the main control influence factors as the importance value of each main control influence factor;
s8: and calculating the percentage of the importance value of each preferred main control influence factor to all the preferred main control factor importance values, and quantitatively representing the influence degree of each main control influence factor on the reservoir quality.
The invention has the beneficial effects that:
according to the method, the random forest algorithm is used for carrying out quantitative evaluation on related influence factors on the characterization parameters reflecting the reservoir quality, so that the extraction of key factors for controlling the reservoir quality is realized, the deposition and diagenesis of the reservoir are determined, and the formation and distribution of the tight sandstone high-quality reservoir are effectively judged. The traditional method mainly depends on human experience, numerical simulation, scanning electron microscope analysis, core test analysis and other methods to determine main control factors, and cannot quantitatively evaluate the importance of each influence factor. Compared with the traditional method, the method applies the random forest algorithm to the evaluation research of the main control factors of the quality of the compact sandstone gas reservoir, not only can automatically screen out the main control factors, but also can quantitatively evaluate the main control factors. The result obtained by the method has the advantages of universality, objectivity, accuracy and the like, and can provide a referential geological basis for the next gas reservoir description and the oil field development effect improvement.
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, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic flow chart of the quantitative evaluation method for the quality control factors of the tight sandstone gas reservoir according to the invention;
FIG. 2 is a flow chart illustrating data processing according to the present invention;
FIG. 3 is a schematic flow chart of dependent variable and independent variable selection according to the present invention;
FIG. 4 is a schematic flow chart of decision tree and random forest construction according to the present invention;
FIG. 5 is a schematic flow chart of the present invention for quantitatively evaluating reservoir dominating factors.
Detailed Description
The invention is further illustrated with reference to the following figures and examples. It should be noted that, in the present application, the embodiments and the technical features of the embodiments may be combined with each other without conflict. It is noted that, unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The use of the terms "comprising" or "including" and the like in the present disclosure is intended to mean that the elements or items listed before the term cover the elements or items listed after the term and their equivalents, but not to exclude other elements or items.
As shown in fig. 1-5, the invention provides a method for quantitatively evaluating the main control factor of the quality of a tight sandstone gas reservoir based on a random forest, which comprises the following steps:
s1: relevant influencing factors influencing reservoir quality of the research area are collected, wherein the relevant influencing factors comprise sedimentation influencing factors, diagenetic influencing factors and construction influencing factors.
In a specific embodiment, the sedimentation type influencing factors comprise granularity lithology type, mineral lithology type, sorting property, roundness, various granularity parameters, contents of different types of mineral particles, impurity base content and primary pore content; the diagenetic influence factors comprise cementing action types, different types of cementing material contents, different types of dissolving hole contents, different types of alternate mineral contents and compaction strength; the construction type influence factors comprise fracture types and different types of fracture contents. It should be noted that the relevant influencing factors may be different for different research areas, and in other research areas, the influencing factors may not be applicable, or may have other influencing factors besides the influencing factors.
S2: performing parameter processing according to the parameter types of the influence factors, and saving the processed parameters as a comma separated value file (. csv) suitable for R language reading; the specific treatment method comprises the following steps:
if the influence factors are data parameters, checking whether the data parameters are missing, if a certain data parameter is missing, removing all influence factors of the same batch (for example, X influence factors are obtained, XY influence factor data are obtained through Y batch test, if the ith batch test only obtains the data of X-j (j is more than or equal to 1) influence factors, removing all influence factor data of the ith batch, namely, finally obtaining X (Y-1) influence factor data); if no deletion exists, reserving; therefore, the integrity of the data can be ensured, and the calculation precision of the subsequent steps is improved.
If the influence factor is a text parameter, carrying out assignment processing on the text parameter; for example, the values of "poor", "medium" and "good" in the sorting property are respectively assigned to 0, 1 and 2, and the values of the support type particles and the miscellaneous base are assigned to 0 and 1, so that the model can conveniently process the text parameters, and it should be noted that the size of the assignment of the text parameters has no influence on the result.
S3: taking the characterization parameters of the reservoir quality as dependent variables of the reservoir quality analysis, and taking the influencing factors as independent variables of the reservoir quality analysis; the characterization parameter of the reservoir quality is determined by the following steps:
(1) collecting and arranging production test data, wherein the production test data comprises oil production, gas production, water production and liquid production;
(2) collecting and sorting (or calculating) parameter data reflecting reservoir quality, wherein the parameter data comprises porosity, permeability, flow unit index (FZI) and pore throat structure; the flow cell index (FZI) is calculated by the following formula:
in the formula: FZI is a flow unit index, and is dimensionless; RQI is the quality index of the reservoir, and is dimensionless; k is the permeability, D;
porosity,%;
is the ratio of pore volume to particle volume.
(3) And (3) analyzing the correlation between the production test data and porosity, permeability, flow unit index (FZI) and pore throat structure parameters by taking the production test data as a standard for evaluating the quality of the reservoir, and determining the parameter which can most reflect the reservoir quality as a characterization parameter of the reservoir quality.
In a specific embodiment, the above method is used to determine the characterization parameter of the reservoir quality, and finally, the flow cell index is preferably selected as the characterization parameter of the reservoir quality.
S4: extracting training data by using a random forest algorithm and a replaced random sampling mode according to the dependent variable and the independent variable to construct a decision tree and a random forest; the method specifically comprises the following substeps:
s41: and (3) by using a random forest algorithm, taking the processed data set as an input sample data set, randomly extracting for many times to form subset data in a replacement mode, wherein the sampling times are consistent with the number of samples, and the subset data obtained by sampling is used for constructing a decision tree.
S42: randomly extracting part of influence factors from the sub-data set to form a candidate segmentation set, selecting one influence factor from the candidate segmentation set as a segmentation point of the decision tree according to a minimum node purity principle, continuing splitting by adopting the principle until all samples of the node reach leaf nodes, and finishing splitting.
The minimum principle of node purity, namely the minimum principle of the Kearny coefficient, can be characterized by calculating the Kearny coefficient of the segmentation point, and the node with the minimum Kearny coefficient is selected as the segmentation node, so that the minimum node purity after segmentation can be realized.
In the process of forming the decision tree, each node is split according to the mode. For example, in any subset of data, if the influence factors of random extraction include primary inter-granular pores, main particle size and cast mold pores, the kini coefficient is calculated, the primary inter-granular pores are selected as the segmentation nodes, and the splitting is finished until all samples of the node reach the leaf node.
S43: and repeating the step S41 and the step S42 to establish a plurality of decision trees, and forming the random forest by the decision trees.
S5: and calculating the error of the data outside the bag of the decision tree aiming at each influence factor, and preferably selecting the main control influence factor according to the error of the data outside the bag.
In the construction process of the decision tree, as a replaced random sampling mode is adopted, part of sample data is always not extracted and is called as data outside a bag; and the error between the predicted value of the data outside the bag in the decision tree and the real value thereof is the error of the data outside the bag. Step S5 specifically includes the following substeps:
s51: putting all influence factors in the data outside the bags into the constructed random forest, and calculating the predicted value of each data outside the bags through the random forest aiming at a certain decision tree;
s52: calculating a mean square error I between a predicted value and a true value of each data outside the bag;
s53: selecting a certain influence factor in the data outside the bag, randomly adding noise into the influence factor, then placing the influence factor into the random forest, and calculating to obtain a predicted value after the noise is added; calculating a mean square error II between a predicted value and a true value of the influencing factor with noise;
s54: judging the importance of the influence factor according to the magnitude of the first mean square error and the second mean square error of the influence factor:
if the second mean square error is larger than the first mean square error after random noise is added, and the difference value between the second mean square error and the first mean square error is larger than a threshold value, the influence factor is important, otherwise, the influence factor is unimportant;
s55: repeating the step S53 and the step S54, and judging all the remaining influence factors in the data outside the bag;
s56: and (4) repeating the step S51 and the step S55 aiming at each remaining decision tree in the random forest, calculating the out-of-bag data error of each influence factor, taking the average value of the out-of-bag data errors as the importance value of each influence factor, arranging the importance values in a descending order according to the size of the average value, wherein the influence factors in the top order are the main control factors influencing the reservoir quality.
In a specific embodiment, the method for quantitatively evaluating the reservoir quality main control factor further comprises the following steps:
s6: reconstructing a decision tree and a random forest according to the main control influence factors selected by the out-of-bag data errors;
s7: calculating the out-of-bag data errors of the optimal main control influence factors, and taking the average value of the out-of-bag data errors of the main control influence factors as the importance value of each main control influence factor;
s8: and calculating the percentage of the importance value of each preferred main control influence factor to all the preferred main control factor importance values, and quantitatively representing the influence degree of each main control influence factor on the reservoir quality.
In a specific embodiment, taking a research area of a tight sandstone gas reservoir as an example, the research area is located in a sunken xanthate structure in the west lake of the east-sea basin, a target layer is the lower section of a gradual-new Huagang group, the structure is located in the middle and south of a sunken central inversion structure zone in the west lake of the east-sea land frame basin, the structure is an NE-SW anticline structure, the stratum is relatively flat, the xanthate 1-1 mainly develops anticline encirclement, and the xanthate 2-2 mainly develops a low-amplitude anticline and broken anticline structure group on a secondary extrusion zone. The hong Kong group is a product deposited and filled in the initial new fracture-depression transition stage, a shallow lake-delta sedimentary system under the main development continental ground background has the sedimentary thickness of between 1000 and 2000m, the total thickness of the lower flower section is less than that of the upper flower section, and the main lithology is as follows: the porosity of the secondary feldspar sandstone, the secondary cuttings sandstone, the feldspar cuttings sandstone and the rock cuttings feldspar sandstone is between 2.1% and 12.5%, the porosity is mainly concentrated on 8% to 10%, the permeability is between 0.02 and 22.72mD, the permeability is higher except one fracture position, the permeability is lower than 2mD, the permeability is mainly concentrated on 0.1mD to 0.4mD, and the low-porosity and low-permeability sandstone reservoir belongs to a typical low-porosity and low-permeability compact sandstone reservoir. Early development practices show that although sand bodies in the research area are large in vertical thickness and continuously distributed in the transverse direction, the reservoir heterogeneity is extremely high, the productivity difference between different intervals of the same development well and between adjacent different development wells is extremely large, the gas production rate of some intervals reaches dozens of thousands of squares, and even some intervals have no capacity, so that the key of the production problem is the reservoir quality difference. Therefore, the main control factor for knowing the reservoir quality is the core basic problem for breaking the efficient development of the tight sandstone gas reservoir in the research area.
The reservoir quality main control factor quantitative evaluation method of the research area comprises the following steps:
the first step is as follows: and collecting sedimentary, diagenetic and tectonic parameters for controlling the quality of the geological reservoir.
(1) Collecting rock core description and experimental analysis data, wherein the rock core description and experimental analysis data comprise 636 pieces of deposition parameter data such as granularity lithology type, mineral lithology type, sorting property, roundness, different types of mineral particle content, impurity-based content, various granularity parameters (average value, standard deviation, skewness and kurtosis), different types of mineral percentage content and primary pore content, 345 pieces of formation parameter data such as cementing action type, different types of cement content (siliceous cement content, calcium cement content, argillaceous cement content and iron ore content), different types of soluble pore content (including inter-granular soluble pore content, intra-granular soluble pore content and cast film pore content), different types of clay mineral content, different types of cross-substituted mineral content and compaction strength, and 640 pieces of construction parameter data such as crack type and different types of crack content;
(2) for text type parameter data, classifying and assigning values to the text data, for example, assigning values of 'poor', 'medium' and 'good' in the sorting property to 0, 1 and 2 respectively, and assigning values of particles and miscellaneous bases of the support type to 0 and 1;
(3) by adopting the steps, samples with more missing parameters are removed, 340 pieces of sample data are obtained, the data set is represented by Q, and the data set is stored as a comma separated value file (. csv) and is suitable for reading in R language.
The second step is that: characterizing parameters for evaluating reservoir quality are determined.
(1) Collecting and arranging production test data including oil production, gas production, water production and liquid production;
(2) collecting and sorting (or calculating) parameter data reflecting reservoir quality, wherein the parameter data comprises porosity, permeability, flow unit index (FZI) and pore throat structure;
(3) analyzing the correlation between the production data and parameters such as porosity, permeability, flow unit index (FZI), pore throat structure and the like by taking the production data as a reservoir quality standard, preferably selecting the FZI as the parameter which can best reflect the reservoir quality, and determining the FZI as a reservoir quality characterization parameter;
(4) the flow unit index (FZI) is used as a dependent variable for controlling the quality of the compact sandstone gas reservoir, and other related influence factors of the reservoir quality, such as the content of inter-granular dissolved pores, the main grain diameter, the content of cast-die pores, the content of primary inter-granular pores and other parameters are used as independent variables.
The third step: and (5) constructing a decision tree and a random forest.
(1) And (3) by utilizing a random forest algorithm, taking the processed 340 sample data sets Q as input sample data sets, and randomly extracting 340 times from the input sample data sets in a replacement mode to form subset data M, wherein the subset data is used for constructing a decision tree.
(2) Each sample in M contains 34 characteristic values, i.e. 34 influencing factors, such as primary intergranular pore content, argillaceous foreign base content, quartz content and the like. Randomly extracting 1/3 influence factors from M to form a candidate division point set of the decision tree, and using C1And (4) showing. For example, the influence factors of random extraction are 11 influence factors such as potassium feldspar content, plagioclase feldspar content, quartz content, roundness, sorting property, support type, primary intergranular pore content and intergranular pore content, and the like, which constitute C1. Calculating C1Selecting the particle size soluble pore with the smallest influence factor as a division point of the decision tree, and dividing M into left and right sets respectively represented as MLAnd MR；
The calculation method of the kini coefficient is as follows: the random forest adopts CART decision tree, in CART algorithm, because of binary tree classification, if the sample subset M is divided into M according to whether the characteristic A takes a certain possible value b or notLAnd MRThen, under the condition of the feature a, the kini coefficient of the set M is:
in the formula: gini (M) represents the uncertainty of the set M; gini (M, a) represents the uncertainty of the set M after a ═ b segmentation; the larger the kini coefficient, the greater the uncertainty in representing the sample, and the greater the node purity after segmentation.
(3) For set MLBy adopting the method, 11 influencing factors are randomly extracted to form a decisionSet of candidate segmentation points of tree C2For example, the influence factors such as the argillaceous content, the iron ore content, the roundness, the quartz content and the like are randomly extracted, and C is calculated2Selecting iron ore with the smallest influence factor as a segmentation node, and further using M as a node for segmentationLDividing the classification into a left classification set and a right classification set; this process is repeated until MLThe splitting is ended when each sample in the set reaches a leaf node.
(4)MRBy reaction of a compound with MLSame way of treatment up to MRWhen all the samples in the group of the leaf node reach the leaf node, the splitting is finished; at this point, the decision tree construction of the subset data M is completed.
(5) And (4) repeating the steps (1) to (4) for multiple times, establishing a plurality of decision trees, and forming a random forest by the plurality of decision trees.
The fourth step: determining main control factors influencing reservoir quality, and carrying out quantitative expression.
(1) Setting decision tree T1The corresponding data outside the bag is O1(ii) a Mixing O with1Putting the obtained solution into a constructed random forest for calculation to obtain a predicted value e of the random forest1(ii) a Calculating the predicted value e of the data outside the bag1Mean square error from the true value E, denoted error1：
error1＝mean(E-e1)2 (5)
(2) Data outside bag O1In (2), for each sample, an intergranular pore (with x) was selected1Express) this feature adds random noise, other feature values remain unchanged, and the out-of-bag data after adding noise is recorded as O2Substituting the prediction value into a random forest for calculation, and recording the obtained prediction value as e2Calculating the true value E and the predicted value E2Mean square error between, noted error2：
error2＝mean(E-e2)2 (6)
(3) For inter-granular pore (x)1) Calculating the difference of the two mean square errors, and recording as Sx1：
Sx1＝error2-error1 (7)
(4) Repeating the steps, respectively calculating the mean square errors of the remaining 33 characteristics in the data outside the bag, such as the argillaceous content, the primary intergranular pore content, the casting die pore content and the like, and respectively recording the mean square errors as Sxi(i＝2,……,34)；
(5) Repeating the steps (1) to (4) aiming at each decision tree in the random forest, calculating the error of the data outside the bag of each influence factor, taking the average value of the errors as the importance value of each influence factor, and recording the importance value as Wxi(i＝1,2,……,34)：
In the formula: k represents the number of decision trees in the random forest.
(6) To WxiAnd (3) sorting in a descending order, eliminating influence factors behind the sorting, such as removing 20 influence factors including plagioclase feldspar content, siliceous cement content, potassium feldspar content, sorting property, quartz content, intra-granular soluble pore content and the like, and preliminarily selecting 14 factors including inter-granular soluble pore content, cast mold pore content, a graphical method granularity average value, primary inter-granular pore content, main grain size and the like as main control factors influencing the reservoir quality.
(7) Based on the preliminarily selected main control influence factors, the decision tree and the random forest are reconstructed by adopting the steps, the error of data outside the bag is calculated, the experiment is repeated for a plurality of times, the average value of the error is taken as the importance value of each influence factor, the importance value is converted into a percentage form, and the result is shown in table 1:
TABLE 1 quantitative evaluation results of reservoir quality master control factors
| Serial number | Influencing factor | Mean square error | Percentage of |
| 1 | Content of intergranular pores | 23.78726498 | 44.56％ |
| 2 | Content of die holes | 5.978337387 | 11.20％ |
| 3 | Average particle size by graphical method | 5.916006387 | 11.08％ |
| 4 | Primary intergranular pore content | 3.495146425 | 6.55％ |
| 5 | Kaolinite content | 3.256391689 | 6.10％ |
| 6 | Major particle size | 2.980834705 | 5.58％ |
| 7 | Calcium cement content (calcite)Dolomite) | 2.246414229 | 4.21％ |
| 8 | Content of illite | 1.420160649 | 2.66％ |
| 9 | Mud content | 1.340817622 | 2.51％ |
| 10 | Maximum particle size | 1.33884933 | 2.51％ |
| 11 | X.S | 0.616936967 | 1.16％ |
| 12 | Content of illite-montmorillonite mixed layer | 0.508265171 | 0.95％ |
| 13 | Lithology of granularity | 0.406021903 | 0.76％ |
| 14 | Of the cement type | 0.097017516 | 0.18％ |
The percentage of importance of each influence factor in table 1 can be used for describing the influence degree of each influence factor on the reservoir quality, so as to realize quantitative evaluation of the main control factor of the tight sandstone gas reservoir quality. In the example, the influence of the inter-granular dissolved pore content on the reservoir of the compact sandstone is the most important, the importance of the reservoir can reach 44.56% after quantification, and then the influence of the cementing type, the granular lithology and the content of the illite-smectite mixed layer is the least and is less than 1% after quantification. Therefore, compared with the primary deposition effect (average particle size, primary inter-particle pores, main particle size and the like) and the cementation effect (kaolinite cementation, calcareous cementation, illite cementation and the like), the accumulated influence degree of inter-particle dissolution pores and casting mold pores on the reservoir quality reaches 55.8 percent, namely the dissolution and erosion effect is a key factor for controlling the reservoir quality and determines the formation and distribution of the high-quality reservoir of the compact sandstone.
The method can quantitatively evaluate the main control factors influencing the reservoir quality, and has remarkable progress compared with the prior art which relies on methods such as human experience, numerical simulation, scanning electron microscope analysis, core test analysis and the like.
Although the present invention has been described with reference to a preferred embodiment, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (7)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110599248.4A CN113344359A (en) | 2021-05-31 | 2021-05-31 | Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110599248.4A CN113344359A (en) | 2021-05-31 | 2021-05-31 | Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113344359A true CN113344359A (en) | 2021-09-03 |
Family
ID=77472448
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110599248.4A Pending CN113344359A (en) | 2021-05-31 | 2021-05-31 | Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113344359A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115128701A (en) * | 2022-07-13 | 2022-09-30 | 西南石油大学 | A seismically intelligent prediction method for sand bodies based on RF-GAM |
| CN116304921A (en) * | 2023-02-08 | 2023-06-23 | 西南石油大学 | An Intelligent Interpretation Method for Reservoir Diagenetic Facies Logging |
| CN119151071A (en) * | 2024-11-11 | 2024-12-17 | 西安石油大学 | Ultra-deep gas well sand production degree prediction method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106405050A (en) * | 2016-09-28 | 2017-02-15 | 西安石油大学 | Method for quantitatively evaluating ultra-deep reservoir diagenesis and pore evolution |
| CN106778836A (en) * | 2016-11-29 | 2017-05-31 | 天津大学 | A kind of random forest proposed algorithm based on constraints |
| CN106841001A (en) * | 2017-01-17 | 2017-06-13 | 西南石油大学 | A kind of tight sand porosity based on reservoir quality Analysis The Main Control Factor, Permeability Prediction method |
| CN110992200A (en) * | 2019-12-11 | 2020-04-10 | 长江大学 | Shale gas well staged fracturing effect evaluation and yield prediction method based on random forest |
| CN112329804A (en) * | 2020-06-30 | 2021-02-05 | 中国石油大学(北京) | Naive Bayes lithofacies classification integrated learning method and device based on feature randomness |
-
2021
- 2021-05-31 CN CN202110599248.4A patent/CN113344359A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106405050A (en) * | 2016-09-28 | 2017-02-15 | 西安石油大学 | Method for quantitatively evaluating ultra-deep reservoir diagenesis and pore evolution |
| CN106778836A (en) * | 2016-11-29 | 2017-05-31 | 天津大学 | A kind of random forest proposed algorithm based on constraints |
| CN106841001A (en) * | 2017-01-17 | 2017-06-13 | 西南石油大学 | A kind of tight sand porosity based on reservoir quality Analysis The Main Control Factor, Permeability Prediction method |
| CN110992200A (en) * | 2019-12-11 | 2020-04-10 | 长江大学 | Shale gas well staged fracturing effect evaluation and yield prediction method based on random forest |
| CN112329804A (en) * | 2020-06-30 | 2021-02-05 | 中国石油大学(北京) | Naive Bayes lithofacies classification integrated learning method and device based on feature randomness |
Non-Patent Citations (1)
| Title |
|---|
| 高强山: "基于随机回归森林的储层预测方法研究", 《中国优秀博硕士学位论文全文数据库（硕士）-基础科学辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115128701A (en) * | 2022-07-13 | 2022-09-30 | 西南石油大学 | A seismically intelligent prediction method for sand bodies based on RF-GAM |
| CN116304921A (en) * | 2023-02-08 | 2023-06-23 | 西南石油大学 | An Intelligent Interpretation Method for Reservoir Diagenetic Facies Logging |
| CN116304921B (en) * | 2023-02-08 | 2025-05-09 | 西南石油大学 | An intelligent interpretation method for reservoir diagenetic phase logging |
| CN119151071A (en) * | 2024-11-11 | 2024-12-17 | 西安石油大学 | Ultra-deep gas well sand production degree prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113344359A (en) | Method for quantitatively evaluating quality master control factors of tight sandstone gas reservoir based on random forest | |
| CN108446797B (en) | Method for predicting capacity of tight oil reservoir horizontal well at initial volume fracturing stage | |
| CN110276827B (en) | A method for evaluating the effectiveness of mud shale reservoirs | |
| CN110644980A (en) | Comprehensive classification evaluation method for ultra-low permeability oil reservoir | |
| CN108009716A (en) | A kind of horizontal well volume fracturing influential effect factor mutiple-stage model method | |
| CN111749686A (en) | Drill bit rapid optimization method based on stratum drilling resistance parameters | |
| CN114215513A (en) | Quantitative determination method, device, medium and equipment for buried hill oil reservoir mode | |
| CN103308433A (en) | Method for analyzing and evaluating tight sandstone reservoir diagenetic facies based on porosity evolution | |
| CN106405050A (en) | Method for quantitatively evaluating ultra-deep reservoir diagenesis and pore evolution | |
| CN110318745A (en) | A kind of lower partial size lithologic log evaluation method of sedimentary micro constraint | |
| CN109034647A (en) | A kind of method that densification oil-gas reservoir volume fracturing horizontal well refracturing selects well | |
| CN107346455A (en) | A kind of method for identifying shale gas production capacity | |
| CN109031424B (en) | A method for identifying diagenetic facies of low-permeability reservoirs based on logging multi-parameters | |
| CN116127675A (en) | Prediction method for maximum recoverable reserve of shale oil horizontal well volume fracturing | |
| CN116537773B (en) | Shale reservoir compressibility confidence evaluation method considering parameter uncertainty | |
| CN107301483A (en) | The rapid integrated method for evaluating non-producing reserves economic producing feasibility | |
| CN105425309B (en) | The method for determining gamma sandstone tufaceous content high | |
| CN110347720B (en) | Fracturing well selection and layer selection method based on flow process | |
| CN112196513A (en) | Longmaxi group shale gas well productivity prediction method based on horizontal well trajectory evaluation | |
| US10954766B2 (en) | Methods, systems, and computer-readable media for evaluating service companies, identifying candidate wells and designing hydraulic refracturing | |
| CN109403960B (en) | Method for judging reservoir fluid properties by logging gas peak-logging state | |
| CN111101914A (en) | Horizontal well fracturing segment cluster optimization method and equipment | |
| CN113820754A (en) | Deep tight sandstone reservoir evaluation method based on artificial intelligence recognition of reservoir lithogenesis | |
| CN117890279B (en) | Front land basin dense sandstone ancient permeability recovery method based on pore structure analysis | |
| CN114638147A (en) | Method for determining fracturing process parameters of oil and gas reservoir |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210903 |