CN111797575B - A method for optimizing process parameters of deep shale fracturing - Google Patents

A method for optimizing process parameters of deep shale fracturing Download PDF

Info

Publication number
CN111797575B
CN111797575B CN202010716805.1A CN202010716805A CN111797575B CN 111797575 B CN111797575 B CN 111797575B CN 202010716805 A CN202010716805 A CN 202010716805A CN 111797575 B CN111797575 B CN 111797575B
Authority
CN
China
Prior art keywords
fracturing
clustering
distance
deep shale
matrix
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010716805.1A
Other languages
Chinese (zh)
Other versions
CN111797575A (en
Inventor
刘彧轩
蒲麒兵
郭建春
路千里
陈天翔
何杰
谢宗财
王世彬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest Petroleum University
Original Assignee
Southwest Petroleum University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest Petroleum University filed Critical Southwest Petroleum University
Priority to CN202010716805.1A priority Critical patent/CN111797575B/en
Publication of CN111797575A publication Critical patent/CN111797575A/en
Application granted granted Critical
Publication of CN111797575B publication Critical patent/CN111797575B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • G06F30/20Design optimisation, verification or simulation
    • G06F30/27Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • EFIXED CONSTRUCTIONS
    • E21EARTH OR ROCK DRILLING; MINING
    • E21BEARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B43/00Methods or apparatus for obtaining oil, gas, water, soluble or meltable materials or a slurry of minerals from wells
    • E21B43/25Methods for stimulating production
    • E21B43/26Methods for stimulating production by forming crevices or fractures
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/16Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/23Clustering techniques
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00Details relating to CAD techniques
    • G06F2111/10Numerical modelling

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Optimization (AREA)
  • Mining & Mineral Resources (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Geology (AREA)
  • Computational Mathematics (AREA)
  • Software Systems (AREA)
  • Mathematical Analysis (AREA)
  • Pure & Applied Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Geometry (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Environmental & Geological Engineering (AREA)
  • Fluid Mechanics (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Algebra (AREA)
  • Geochemistry & Mineralogy (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Biology (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明公开了一种深层页岩压裂工艺参数优化方法,包括以下步骤:收集多个不同区块的深层页岩的基本参数;采用模糊聚类法对所述深层页岩进行储层划分;利用软件模拟水力裂缝形态,根据模拟结果确定划分后的每类储层的水力压裂目标;根据所述水力压裂目标,采用数值模拟方法模拟不同施工参数的水力压裂结果,根据所述水力压裂结果优化施工参数。本发明能够将多个不同区块的深层页岩进行聚类储层划分,然后对每一类的储层进行压裂工艺参数优化,大大减少了压裂工艺参数优化的工作量,节省成本。

Figure 202010716805

The invention discloses a method for optimizing process parameters of deep shale fracturing, comprising the following steps: collecting basic parameters of deep shale in multiple different blocks; adopting a fuzzy clustering method to divide the deep shale into reservoirs; Use software to simulate the shape of hydraulic fractures, and determine the hydraulic fracturing target of each type of reservoir after division according to the simulation results; Fracturing results optimize construction parameters. The invention can divide the deep shale of multiple different blocks into clustering reservoirs, and then optimize the fracturing process parameters for each type of reservoir, greatly reducing the workload of fracturing process parameter optimization and saving costs.

Figure 202010716805

Description

一种深层页岩压裂工艺参数优化方法A method for optimizing process parameters of deep shale fracturing

技术领域technical field

本发明涉及压裂改造技术领域,特别涉及一种深层页岩压裂工艺参数优化方法。The invention relates to the technical field of fracturing reformation, in particular to a method for optimizing process parameters of deep shale fracturing.

背景技术Background technique

我国页岩气资源十分丰富,根据2012年国土资源部油气中心最新研究成果表明,我国页岩气可采资源量为25万亿方,仅四川省地区寒武系筇竹寺组和志留系龙马溪组的页岩气资源就可以与四川盆地的常规天然气资源总量相媲美,开发利用潜力巨大。为了降低我国天然气资源的对外依存度,改变我国目前以煤炭为主的能源消费结构,发展绿色环保的页岩气资源,对国民经济长期快速发展和国家能源安全具有重要的战略意义。my country is very rich in shale gas resources. According to the latest research results of the Oil and Gas Center of the Ministry of Land and Resources in 2012, the recoverable resources of shale gas in my country are 25 trillion cubic meters. Only the Cambrian Qiongzhusi Formation and the Silurian Longma in Sichuan Province The shale gas resources in the Xi Formation can be compared with the total conventional natural gas resources in the Sichuan Basin, and the development and utilization potential is huge. In order to reduce the external dependence of my country's natural gas resources, change my country's current coal-based energy consumption structure, and develop green and environmentally friendly shale gas resources, it has important strategic significance for the long-term rapid development of the national economy and national energy security.

页岩气的成藏机理和赋存机制与常规油气有本质区别。页岩由于自身致密性,既是生烃母岩,也是储层,为原位饱和早成藏。页岩储层基质渗透率极低,尤其是页岩气储层,为纳达西级。要想实现页岩气的经济开发需要采用大规模水力压裂,尽最大限度将储层压碎,形成复杂裂缝网络,从而有效增加页岩气渗流面积,降低渗流阻力。The accumulation mechanism and occurrence mechanism of shale gas are fundamentally different from those of conventional oil and gas. Due to its compactness, shale is not only a hydrocarbon-generating parent rock, but also a reservoir, which is in-situ saturation and early accumulation. The matrix permeability of shale reservoirs is extremely low, especially shale gas reservoirs, which are Nadaxi grade. In order to realize the economic development of shale gas, large-scale hydraulic fracturing is required to crush the reservoir as much as possible to form a complex fracture network, thereby effectively increasing the seepage area of shale gas and reducing the seepage resistance.

由于地层岩石在平面上和纵向上的非均质性,使得各区块储层岩石矿物成分、力学性能、天然裂缝以及地应力状态存在差异较大,因此各区块页岩储层在水力压裂过程中形成复杂缝网网络的能力也存在一定差异。Due to the heterogeneity of the formation rock in plane and longitudinal direction, the mineral composition, mechanical properties, natural fractures and in-situ stress state of the reservoir rocks in each block are quite different. There are also certain differences in the ability to form complex network of seams.

发明内容SUMMARY OF THE INVENTION

针对上述问题,本发明旨在提供一种深层页岩压裂工艺参数优化方法,针对不同区块页岩储层特征,采用聚类法进行储层划分,采用数值模拟方法,以实现“有效改造体积最大化,改造体积内裂缝复杂程度最大化”的目标和思路对施工参数进行精细化设计,形成了一套针对不同页岩储层特征的压裂参数优化方法。In view of the above problems, the present invention aims to provide a method for optimizing the process parameters of deep shale fracturing. According to the characteristics of shale reservoirs in different blocks, the clustering method is used to divide the reservoirs, and the numerical simulation method is used to realize "effective fracturing". The goal and idea of maximizing the volume and maximizing the complexity of the fractures in the reformed volume” refined the design of the construction parameters, and formed a set of fracturing parameter optimization methods for different shale reservoir characteristics.

本发明的技术方案如下:The technical scheme of the present invention is as follows:

一种深层页岩压裂工艺参数优化方法,包括以下步骤:收集多个不同区块的深层页岩的基本参数;采用模糊聚类法对所述深层页岩进行储层划分;利用软件模拟水力裂缝形态,根据模拟结果确定划分后的每类储层的水力压裂目标;根据所述水力压裂目标,采用数值模拟方法模拟不同施工参数的水力压裂结果,根据所述水力压裂结果优化施工参数。A method for optimizing process parameters of deep shale fracturing, comprising the following steps: collecting basic parameters of deep shale in multiple different blocks; using fuzzy clustering method to divide the deep shale into reservoirs; using software to simulate hydraulic pressure Fracture shape, according to the simulation results to determine the hydraulic fracturing target of each type of reservoir; according to the hydraulic fracturing target, use the numerical simulation method to simulate the hydraulic fracturing results of different construction parameters, and optimize according to the hydraulic fracturing results construction parameters.

作为优选,所述基本参数包括脆性矿物、泊松比、弹性模量、抗张强度、抗剪强度、断裂韧性、天然裂缝、力学脆性指数、最大水平主应力、最小水平主应力、水平地应力差、垂向地应力差。Preferably, the basic parameters include brittle minerals, Poisson's ratio, elastic modulus, tensile strength, shear strength, fracture toughness, natural fractures, mechanical brittleness index, maximum horizontal principal stress, minimum horizontal principal stress, and horizontal in-situ stress difference, vertical stress difference.

作为优选,所述模糊聚类法采用系统聚类法,具体包括以下步骤:Preferably, the fuzzy clustering method adopts a systematic clustering method, which specifically includes the following steps:

设置所述基本参数的数据矩阵,并对矩阵中的数据进行数据标准化,将数据压缩到[0,1]区间上;Setting the data matrix of the basic parameters, standardizing the data in the matrix, and compressing the data to the [0,1] interval;

建立模糊相似矩阵,计算每个区块的相似程度;Establish a fuzzy similarity matrix and calculate the similarity of each block;

根据所述相似程度的计算结果,对每个区块进行聚类。According to the calculation result of the similarity degree, each block is clustered.

作为优选,所述数据矩阵建立方法如下:设论域U={x1,x2,……,xn}为被分类对象,每个对象有m个指标表示其性状,即Preferably, the method for establishing the data matrix is as follows: set the universe of discourse U={x 1 , x 2 , .

xi={xi1,xi2,…,xim}(i=1,2,…,n) (1)x i ={x i1 ,x i2 ,...,x im }(i=1,2,...,n) (1)

于是,得到所述数据矩阵为:Then, the data matrix is obtained as:

Figure BDA0002598495350000021
Figure BDA0002598495350000021

其中xnm表示第n个分类对象的第m个指标的原始数据。where x nm represents the raw data of the mth index of the nth classification object.

作为优选,所述数据标准化采用平移、极差变换法实现,该方法具体如下:Preferably, the data standardization is realized by translation and range transformation, and the method is as follows:

Figure BDA0002598495350000022
Figure BDA0002598495350000022

Figure BDA0002598495350000023
Figure BDA0002598495350000023

Figure BDA0002598495350000024
Figure BDA0002598495350000024

Figure BDA0002598495350000025
Figure BDA0002598495350000025

式中:0≤x″ik≤1,且消除了量纲的影响。In the formula: 0≤x″ ik ≤1, and the influence of dimension is eliminated.

作为优选,所述相似程度通过相似系数法或距离法进行计算,所述相似系数法包括夹角余弦法、最大最小法、算数平均最小法、几何平均最小法、数量积法、相关系数法、指数相似系数法,所述距离法包括直接距离法、海明距离法、欧几里得距离法、切比雪夫距离法、倒数距离法、指数距离法。Preferably, the similarity degree is calculated by the similarity coefficient method or the distance method, and the similarity coefficient method includes the angle cosine method, the maximum and minimum method, the arithmetic mean minimum method, the geometric mean minimum method, the quantity product method, the correlation coefficient method, Exponential similarity coefficient method, the distance method includes direct distance method, Hamming distance method, Euclidean distance method, Chebyshev distance method, reciprocal distance method, and exponential distance method.

作为优选,采用基于模糊等价矩阵聚类方法或直接聚类法进行聚类,聚类时需确定阈值,所述阈值通过F统计量进行确定。Preferably, a clustering method based on a fuzzy equivalent matrix or a direct clustering method is used for clustering, and a threshold value needs to be determined during clustering, and the threshold value is determined by F statistic.

作为优选,所述施工参数包括簇数、排量、用液强度。Preferably, the construction parameters include the number of clusters, displacement, and liquid intensity.

与现有技术相比,本发明具有如下优点:Compared with the prior art, the present invention has the following advantages:

本发明首先将采用模糊聚类法对多个不同区块的深层页岩进行聚类储层划分,然后对每一类的储层采用数值模拟方法进行压裂工艺参数优化,大大减少了压裂工艺参数优化的工作量,节省成本。In the present invention, the fuzzy clustering method is used to firstly divide the deep shale in different blocks into clusters and reservoirs, and then the numerical simulation method is used for each type of reservoir to optimize the fracturing process parameters, which greatly reduces the number of fracturing. Process parameter optimization workload and cost savings.

附图说明Description of drawings

为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained from these drawings without any creative effort.

图1为本发明实施例储层聚类划分结果示意图;FIG. 1 is a schematic diagram of the result of clustering division of reservoirs according to an embodiment of the present invention;

图2为本发明实施例簇数数值模拟缝宽随簇数变化结果示意图;FIG. 2 is a schematic diagram of the results of numerical simulation of the number of clusters in the embodiment of the present invention, and the change of the slit width with the number of clusters;

图3为本发明实施例簇数数值模拟半缝长随簇数变化结果示意图;FIG. 3 is a schematic diagram of the result of numerical simulation of the number of clusters according to the embodiment of the present invention;

图4为本发明实施例簇数数值模拟改造体积随簇数变化结果示意图;FIG. 4 is a schematic diagram of the results of the numerical simulation of the number of clusters in the embodiment of the present invention, and the transformation volume changes with the number of clusters;

图5为本发明实施例排量数值模拟缝宽随排量变化结果示意图;Fig. 5 is a schematic diagram of the results of the numerical simulation of displacement according to the embodiment of the present invention;

图6为本发明实施例排量数值模拟半缝长随排量变化结果示意图;FIG. 6 is a schematic diagram of the results of the numerical simulation of displacement in accordance with the embodiment of the present invention;

图7为本发明实施例排量数值模拟改造体积随排量变化结果示意图;FIG. 7 is a schematic diagram of the results of the numerical simulation of displacement according to the embodiment of the present invention;

图8为本发明实施例用液强度数值模拟缝宽随用液强度变化结果示意图;FIG. 8 is a schematic diagram of the results of numerical simulation of the liquid intensity variation of the crack width with the liquid intensity according to the embodiment of the present invention;

图9为本发明实施例用液强度数值模拟半缝长随用液强度变化结果示意图;FIG. 9 is a schematic diagram of the result of numerical simulation of the liquid strength of the embodiment of the present invention, and the change of the half-slit length with the liquid strength;

图10为本发明实施例用液强度数值模拟改造体积随用液强度变化结果示意图。Fig. 10 is a schematic diagram showing the result of numerical simulation of the strength of the liquid used in the modification volume of the liquid according to the change of the strength of the liquid used in the embodiment of the present invention.

具体实施方式Detailed ways

下面结合附图和实施例对本发明进一步说明。需要说明的是,在不冲突的情况下,本申请中的实施例及实施例中的技术特征可以相互结合。除非另外定义,本发明公开使用的技术术语或者科学术语应当为本公开所属领域内具有一般技能的人士所理解的通常意义。本发明公开使用的“包括”或者“包含”等类似的词语意指出现该词前面的元件或者物件涵盖出现在该词后面列举的元件或者物件及其等同,而不排除其他元件或者物件。The present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be noted that the embodiments in the present application and the technical features in the embodiments may be combined with each other under the condition of no conflict. Unless otherwise defined, technical or scientific terms used in the present disclosure shall have the ordinary meaning as understood by one of ordinary skill in the art to which this disclosure belongs. The use of "comprising" or "comprising" and similar words in the present disclosure means that the elements or items appearing before the word encompass the elements or items listed after the word and their equivalents, but do not exclude other elements or items.

一种深层页岩压裂工艺参数优化方法,包括以下步骤:A method for optimizing process parameters of deep shale fracturing, comprising the following steps:

S1:收集多个不同区块的深层页岩的基本参数,所述基本参数包括脆性矿物、泊松比、弹性模量、抗张强度、抗剪强度、断裂韧性、天然裂缝、力学脆性指数、最大水平主应力、最小水平主应力、水平地应力差、垂向地应力差。S1: Collect basic parameters of deep shale in multiple different blocks, the basic parameters include brittle minerals, Poisson's ratio, elastic modulus, tensile strength, shear strength, fracture toughness, natural fractures, mechanical brittleness index, Maximum horizontal principal stress, minimum horizontal principal stress, horizontal in-situ stress difference, and vertical in-situ stress difference.

S2:采用模糊聚类法对所述深层页岩进行储层划分,所述模糊聚类法采用系统聚类法,具体包括以下步骤:S2: The deep shale is divided by a fuzzy clustering method, and the fuzzy clustering method adopts a systematic clustering method, which specifically includes the following steps:

S21:设置所述基本参数的数据矩阵,并对矩阵中的数据进行数据标准化,将数据压缩到[0,1]区间上,具体地:S21: Set the data matrix of the basic parameters, standardize the data in the matrix, and compress the data into the [0,1] interval, specifically:

设论域U={x1,x2,……,xn}为被分类对象,每个对象有m个指标表示其性状,即Suppose the universe of discourse U = {x 1 , x 2 ,..., x n } is the object to be classified, and each object has m indicators to represent its characteristics, namely

xi={xi1,xi2,…,xim}(i=1,2,…,n) (1)x i ={x i1 ,x i2 ,...,x im }(i=1,2,...,n) (1)

于是,得到所述数据矩阵为:Therefore, the data matrix is obtained as:

Figure BDA0002598495350000041
Figure BDA0002598495350000041

其中xnm表示第n个分类对象的第m个指标的原始数据。where x nm represents the raw data of the mth index of the nth classification object.

采用下述任意一种方法进行数据标准化:Data normalization is performed using any of the following methods:

1)平移、极差变换1) Translation, range transformation

Figure BDA0002598495350000042
Figure BDA0002598495350000042

Figure BDA0002598495350000043
Figure BDA0002598495350000043

Figure BDA0002598495350000044
Figure BDA0002598495350000044

Figure BDA0002598495350000045
Figure BDA0002598495350000045

式中:0≤x″ik≤1,且消除了量纲的影响。In the formula: 0≤x″ ik ≤1, and the influence of dimension is eliminated.

2)平移、标准差变换2) Translation, standard deviation transformation

直接通过式(4)-(6)进行数据标准化,经过变换后,每个变量的均值为0,标准差为1,且消除了量纲的影响,但是,再用得到的x′ik不一定在区间[0,1]上。Data standardization is carried out directly by formulas (4)-(6). After transformation, the mean value of each variable is 0, the standard deviation is 1, and the influence of dimension is eliminated. However, the obtained x'ik is not necessarily on the interval [0,1].

3)对数变换3) Logarithmic transformation

x′″ik=lg xik (7)x′″ ik = lg x ik (7)

取对数能缩小变量间的数量级。Taking the logarithm narrows the order of magnitude between the variables.

S22:建立模糊相似矩阵,计算每个区块的相似程度,具体地:xi和xj的相似程度rij=R(xi,xj),其值通过以下方法进行计算:S22: establish a fuzzy similarity matrix, and calculate the similarity degree of each block, specifically: the similarity degree of x i and x j r ij =R( xi , x j ), and its value is calculated by the following method:

(1)相似系数法(1) Similarity coefficient method

1)夹角余弦法1) Angle cosine method

Figure BDA0002598495350000051
Figure BDA0002598495350000051

2)最大最小法2) Maximum and minimum method

Figure BDA0002598495350000052
Figure BDA0002598495350000052

3)算数平均最小法3) Arithmetic average minimum method

Figure BDA0002598495350000053
Figure BDA0002598495350000053

4)几何平均最小法4) Geometric mean minimum method

Figure BDA0002598495350000054
Figure BDA0002598495350000054

5)数量积法5) Quantity product method

Figure BDA0002598495350000055
Figure BDA0002598495350000055

Figure BDA0002598495350000056
Figure BDA0002598495350000056

6)相关系数法6) Correlation coefficient method

Figure BDA0002598495350000057
Figure BDA0002598495350000057

Figure BDA0002598495350000058
Figure BDA0002598495350000058

7)指数相似系数法7) Exponential similarity coefficient method

Figure BDA0002598495350000061
Figure BDA0002598495350000061

(2)距离法(2) Distance method

1)直接距离法1) Direct distance method

rij=1-cd(xi,xj) (17)r ij =1-cd(x i ,x j ) (17)

其中c为适当选取的参数,使得0≤rij≤1,d(xi,xj)表示他们之间的距离,其可通过以下方法进行计算:where c is an appropriately selected parameter such that 0≤r ij ≤1, and d(x i ,x j ) represents the distance between them, which can be calculated by the following method:

海明距离Hamming distance

Figure BDA0002598495350000062
Figure BDA0002598495350000062

欧几里得距离Euclidean distance

Figure BDA0002598495350000063
Figure BDA0002598495350000063

切比雪夫距离Chebyshev distance

Figure BDA0002598495350000064
Figure BDA0002598495350000064

2)倒数距离法2) Reciprocal distance method

Figure BDA0002598495350000065
Figure BDA0002598495350000065

其中M为适当选取的参数,使得0≤rij≤1。where M is an appropriately selected parameter such that 0≤r ij ≤1.

3)指数距离法3) Exponential distance method

rij=exp[-d(xi,xj)] (22)r ij =exp[-d(x i ,x j )] (22)

S23:根据所述相似程度的计算结果,,采用基于模糊等价矩阵聚类方法或直接聚类法对每个区块进行聚类,聚类时需确定阈值,所述阈值通过F统计量进行确定。S23: According to the calculation result of the similarity degree, use the fuzzy equivalent matrix-based clustering method or the direct clustering method to cluster each block, and a threshold value needs to be determined during clustering, and the threshold value is determined by the F statistic Sure.

(1)基于模糊等价矩阵聚类方法(1) Clustering method based on fuzzy equivalent matrix

1)传递闭包法1) transitive closure method

根据标定所得的模糊矩阵R还要将其改造称模糊等价矩阵R*。用二次方法求R的传递闭包,即t(R)=R*。再让λ由大变小,就可形成动态聚类图。According to the fuzzy matrix R obtained by calibration, it should be transformed into a fuzzy equivalent matrix R * . Use the quadratic method to find the transitive closure of R, that is, t(R)=R * . Then let λ change from large to small, and a dynamic clustering graph can be formed.

2)布尔矩阵法2) Boolean matrix method

布尔矩阵法的理论依据是下面的定理:设R为U的一个子集,是R上的一个相似的布尔矩阵,则R具有传递性(当是等价布尔矩阵时)

Figure BDA0002598495350000071
矩阵R2=R*R≤R。The theoretical basis of the Boolean matrix method is the following theorem: Let R be a subset of U, a similar Boolean matrix on R, then R is transitive (when it is an equivalent Boolean matrix)
Figure BDA0002598495350000071
The matrix R 2 =R * R≤R.

布尔矩阵法的具体步骤如下:The specific steps of the Boolean matrix method are as follows:

求模糊相似矩阵的λ-截矩阵Rλ,若Rλ按定理判定为等价的,则由Rλ可得U在λ水平上的分类;若Rλ判定为不等价,则Rλ在某一排列下有上述形式的特殊子矩阵,此时只要将其中特殊子矩阵的0一律改成1直到不再产生上述形式的子矩阵即可。如此得到的

Figure BDA0002598495350000072
为等价矩阵。因此,由
Figure BDA0002598495350000073
可得λ水平上的分类。Find the λ-truncated matrix R λ of the fuzzy similarity matrix, if R λ is determined to be equivalent according to the theorem, then the classification of U at the λ level can be obtained from R λ ; if R λ is determined to be unequal, then R λ is in There is a special sub-matrix of the above form in a certain arrangement, at this time, it is only necessary to change the 0 of the special sub-matrix to 1 until the sub-matrix of the above form is no longer generated. so obtained
Figure BDA0002598495350000072
is an equivalent matrix. Therefore, by
Figure BDA0002598495350000073
Classification at the λ level is available.

(2)直接聚类法(2) Direct clustering method

所谓直接聚类法,是指在建立模糊相似矩阵之后,不去求传递闭包t(R),也不用布尔矩阵法,而是直接从模糊相似矩阵出发求得聚类图。其步骤如下:The so-called direct clustering method means that after establishing the fuzzy similarity matrix, instead of seeking the transitive closure t(R), nor using the Boolean matrix method, the clustering graph is obtained directly from the fuzzy similarity matrix. The steps are as follows:

1)取λ1=1(最大值),对每个xi作相似类[xi]R,且1) Take λ 1 =1 (maximum value), make a similarity class [ xi ] R for each x i , and

[xi]R={xj|rij=1} (23)[x i ] R = {x j |r ij =1} (23)

即将满足rij=1的xi与xj放在一类,构成相似类。相似类与等价类的不同之处是,不同的相似类可能有公共元素,即可出现That is, x i and x j satisfying r ij =1 are placed in one class to form a similar class. The difference between similar classes and equivalence classes is that different similar classes may have common elements, which can appear

Figure BDA0002598495350000074
Figure BDA0002598495350000074

此时只要将有公共元素的相似类合并,即可得λ1=1水平上的等价分类。At this time, as long as the similar classes with common elements are merged, the equivalent classification at the level of λ 1 =1 can be obtained.

2)取λ2为次大值,从R中直接找出相似度为λ2的元素对(xi,xj)(即rij=λ2),将对应于λ1=1的等价分类中xi所在的类与xj所在的类合并,将所有的这些情况合并后,即得到对应于λ2的等价分类。2) Take λ 2 as the second largest value, and directly find the element pair (x i , x j ) (ie r ij2 ) with a similarity of λ 2 from R, which will correspond to the equivalent of λ 1 =1 In the classification, the class where x i is located is merged with the class where x j is located, and after combining all these cases, the equivalent classification corresponding to λ 2 is obtained.

3)取λ3为第三大值,从R中直接找出相似度为λ3的元素对(xi,xj)(即rij=λ3),将对应于λ2的等价分类中xi所在的类与xj所在的类合并,将所有的这些情况合并后,即得到对应于λ3的等价分类。3) Take λ 3 as the third largest value, directly find the element pair (x i , x j ) (ie r ij3 ) with a similarity of λ 3 from R, and classify the equivalent corresponding to λ 2 The class where x i is located is merged with the class where x j is located. After combining all these cases, the equivalent classification corresponding to λ 3 is obtained.

4)以此类推,直到合并到U成为一类为止。4) And so on, until the merging to U becomes a class.

(3)阈值的确定(3) Determination of the threshold value

在模糊聚类分析中对于各个不同的λ∈[0,1],可得到不同的分类,许多实际问题需要选择某个阈值λ,确定样本的一个具体分类,这就提出了如何确定阈值的问题。一般有以下两个方法:In fuzzy clustering analysis, for each different λ∈[0,1], different classifications can be obtained. Many practical problems need to select a certain threshold λ to determine a specific classification of the sample, which raises the question of how to determine the threshold . There are generally two methods:

1)按实际需要,在动态聚类图中,调整λ的值以得到适当的分类,而不需要事先准确地估计好样本应分成几类。当然,也可由具有丰富经验的专家结合专业知识确定阈值λ,从而得出在λ水平上的等价分类。1) According to the actual needs, in the dynamic clustering diagram, adjust the value of λ to get the appropriate classification, without the need to accurately estimate how many classes the samples should be divided into in advance. Of course, the threshold λ can also be determined by experts with rich experience combined with professional knowledge, so as to obtain an equivalent classification at the λ level.

2)用F统计量确定最佳值:设论域U={x1,x2,……,xn}为样本空间(样本总数为n),而每个样本xi有m个特征,即公式(1),于是得到原始数据矩阵,如下表所示:2) Use F statistic to determine the best value: let the universe of discourse U={x 1 , x 2 ,..., x n } be the sample space (the total number of samples is n), and each sample x i has m features, That is, formula (1), so the original data matrix is obtained, as shown in the following table:

表1指标关系表Table 1 Index relationship table

Figure BDA0002598495350000081
Figure BDA0002598495350000081

其中,

Figure BDA0002598495350000082
Figure BDA0002598495350000083
称为总体样本的中心向量。in,
Figure BDA0002598495350000082
Figure BDA0002598495350000083
is called the center vector of the population sample.

设对应于λ值的分类数为r,第j类的样本数为nj,第j类的样本记为:

Figure BDA00025984953500000813
第j类的聚类中心为向量
Figure BDA0002598495350000084
其中
Figure BDA0002598495350000085
为第k个特征的平均值,即Let the number of classifications corresponding to the λ value be r, the number of samples of the jth class be n j , and the samples of the jth class are recorded as:
Figure BDA00025984953500000813
The cluster center of the jth class is a vector
Figure BDA0002598495350000084
in
Figure BDA0002598495350000085
is the average of the k-th feature, that is

Figure BDA0002598495350000086
Figure BDA0002598495350000086

作F统计量as F statistic

Figure BDA0002598495350000087
Figure BDA0002598495350000087

Figure BDA0002598495350000088
Figure BDA0002598495350000088

Figure BDA0002598495350000089
Figure BDA00025984953500000810
间的距离,
Figure BDA00025984953500000811
为第j类中第i个样本x(j)与其中心
Figure BDA00025984953500000812
间的距离。称为F统计量,它是遵从自由度为r-1,n-r的分布。它的分子表征类与类之间的距离,分母表征类内样本间的距离。因此,F值越大,说明类与类之间的距离越大;类与类间的差异越大,分类就越好。for
Figure BDA0002598495350000089
and
Figure BDA00025984953500000810
distance between,
Figure BDA00025984953500000811
is the i-th sample x (j) in the j-th class and its center
Figure BDA00025984953500000812
distance between. Called the F statistic, it follows a distribution with degrees of freedom r-1, nr. Its numerator represents the distance between classes and the denominator represents the distance between samples within a class. Therefore, the larger the F value, the greater the distance between classes; the greater the difference between classes, the better the classification.

S3:利用软件模拟水力裂缝形态,根据模拟结果确定划分后的每类储层的水力压裂目标。S3: Use software to simulate the shape of hydraulic fractures, and determine the hydraulic fracturing targets for each type of reservoir after division according to the simulation results.

S4:根据所述水力压裂目标,采用数值模拟方法模拟不同施工参数的水力压裂结果,根据所述水力压裂结果优化施工参数,所述施工参数包括簇数、排量、用液强度。S4: According to the hydraulic fracturing target, numerical simulation method is used to simulate hydraulic fracturing results of different construction parameters, and construction parameters are optimized according to the hydraulic fracturing results, and the construction parameters include cluster number, displacement, and liquid strength.

实施例1Example 1

S1:收集多个不同区块的深层页岩的基本参数,具体如表2所示:S1: Collect basic parameters of deep shale in multiple different blocks, as shown in Table 2:

表2储层基本参数Table 2 Basic parameters of reservoir

Figure BDA0002598495350000091
Figure BDA0002598495350000091

S2:采用模糊聚类法对区块1至区块5进行储层划分。区块1至区块5为被分类的对象,而每个对象又由11个因素指标来决定,于是就可以得到原始数据矩阵。在11个因素指标中,不同的数据有不同的量纲,为了使具有不同量纲的数据能比较,同样将数据投影到区间(0,1)上。计算不同区块11个参数之间的相似程度,将每个区块的因素参数,看成空间中的一个点,并用某种度量,如欧式距离,测量点与点之间的距离,距离较近的归为一类,距离较远的点应属于同一类。S2: Use fuzzy clustering method to divide reservoirs from block 1 to block 5. Blocks 1 to 5 are classified objects, and each object is determined by 11 factor indicators, so the original data matrix can be obtained. Among the 11 factor indicators, different data have different dimensions. In order to make data with different dimensions can be compared, the data are also projected to the interval (0,1). Calculate the degree of similarity between 11 parameters of different blocks, regard the factor parameters of each block as a point in the space, and use some measure, such as Euclidean distance, to measure the distance between points, and the distance is relatively high. Points that are close are classified into one class, and points that are farther away should belong to the same class.

聚类结果如图1所示,从图1可以看出,在聚类联接树状图中,段与段之间的相似程度用相对距离(0~25)表示。距离越小,表明段与段相似程度越高;距离越大,差异越大。其中区块1和区块2的压裂地质参数的相对距离为0~1之间,两区块接近,归为第一类;区块3和区块5的压裂地质参数的相对距离为15~20之间,两区块接近,归为第二类;区块4的压裂地质参数的相对距离为20~25之间,与另外4个区块的差异较大,归为第三类。The clustering result is shown in Figure 1. It can be seen from Figure 1 that in the clustering connection tree diagram, the degree of similarity between segments is represented by relative distances (0-25). The smaller the distance, the higher the similarity between the segments; the larger the distance, the greater the difference. The relative distance between the fracturing geological parameters of block 1 and block 2 is between 0 and 1, and the two blocks are close to each other, which is classified as the first category; the relative distance of the fracturing geological parameters of block 3 and block 5 is Between 15 and 20, the two blocks are close, and are classified as the second category; the relative distance of the fracturing geological parameters of block 4 is between 20 and 25, which is quite different from the other four blocks, and is classified as the third kind.

S3:根据每个区块的地质特征,利用软件模拟水力裂缝形态,根据模拟结果确定划分后的每类储层的水力压裂目标。其中,区块1和区块2以复杂缝网为水力压裂目标,区块3和区块5以主裂缝+分支缝为水力压裂目标,区块4以主裂缝为水力压裂目标。S3: According to the geological characteristics of each block, use the software to simulate the hydraulic fracture shape, and determine the hydraulic fracturing target of each type of reservoir after division according to the simulation results. Among them, block 1 and block 2 take complex fracture network as the hydraulic fracturing target, block 3 and block 5 take the main fracture + branch fracture as the hydraulic fracturing target, and block 4 take the main fracture as the hydraulic fracturing target.

S4:根据所述水力压裂目标,采用数值模拟方法模拟不同施工参数的水力压裂结果,根据所述水力压裂结果优化施工参数。S4: According to the hydraulic fracturing target, use a numerical simulation method to simulate hydraulic fracturing results of different construction parameters, and optimize the construction parameters according to the hydraulic fracturing results.

其中,区块1和区块2的簇数数值模拟结果如图2-4所示,从图2-4可以看出,随着簇数增加水力裂缝总的改造体积呈现增加趋势,但裂缝的宽度、长度均减小;当簇数大于4簇以后,改造体积呈降低趋势,推荐簇数为3-4簇。采用相同方法获得区块3和区块5的推荐簇数为4-6簇,区块4的推荐簇数为6-8簇。Among them, the numerical simulation results of the number of clusters in Block 1 and Block 2 are shown in Figure 2-4. It can be seen from Figure 2-4 that with the increase of the number of clusters, the total reformed volume of hydraulic fractures shows an increasing trend, but the size of the fractures increases. The width and length are reduced; when the number of clusters is greater than 4 clusters, the transformation volume tends to decrease, and the recommended number of clusters is 3-4 clusters. Using the same method to obtain the recommended number of clusters for block 3 and block 5 is 4-6 clusters, and the recommended number of clusters for block 4 is 6-8 clusters.

其中,区块1和区块2在簇数为4的基础上,排量数值模拟结果如图5-7所示,综合考虑预防套变以及追求改造体积,4簇下推荐排量为12-14m3/min。采用相同方法获得区块3和区块5的推荐排量为14-16m3/min,区块4的推荐排量为16m3/min以上。Among them, block 1 and block 2 are based on the number of clusters of 4, and the numerical simulation results of displacement are shown in Figure 5-7. Taking into account the prevention of casing changes and the pursuit of transformation volume, the recommended displacement under 4 clusters is 12- 14m 3 /min. Using the same method to obtain the recommended displacement of block 3 and block 5 is 14-16m 3 /min, and the recommended displacement of block 4 is more than 16m 3 /min.

其中,区块1和区块2在簇数为4、排量为14m3/min的基础上,用液强度数值模拟结果如图8-10所示,从图8-10可以看出,簇数为4、排量为14m3/m情况下,推荐用液强度为24-26m3/m。采用相同方法获得区块3和区块5的推荐用液强度为24-28m3/m,区块4的推荐用液强度为30m3/m以上。Among them, block 1 and block 2 are based on the number of clusters of 4 and the displacement of 14m 3 /min, and the numerical simulation results of liquid strength are shown in Figure 8-10. When the number is 4 and the displacement is 14m 3 /m, the recommended liquid strength is 24-26m 3 /m. The recommended liquid strength of block 3 and block 5 obtained by the same method is 24-28 m 3 /m, and the recommended liquid strength of block 4 is more than 30 m 3 /m.

以上所述,仅是本发明的较佳实施例而已,并非对本发明作任何形式上的限制,虽然本发明已以较佳实施例揭露如上,然而并非用以限定本发明,任何熟悉本专业的技术人员,在不脱离本发明技术方案范围内,当可利用上述揭示的技术内容做出些许更动或修饰为等同变化的等效实施例,但凡是未脱离本发明技术方案的内容,依据本发明的技术实质对以上实施例所作的任何简单修改、等同变化与修饰,均仍属于本发明技术方案的范围内。The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention in any form. Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Technical personnel, within the scope of the technical solution of the present invention, can make some changes or modifications to equivalent examples of equivalent changes by using the technical content disclosed above, but any content that does not depart from the technical solution of the present invention, according to the present invention. The technical essence of the invention Any simple modifications, equivalent changes and modifications made to the above embodiments still fall within the scope of the technical solutions of the present invention.

Claims (5)

1.一种深层页岩压裂工艺参数优化方法,其特征在于,包括以下步骤:1. a deep shale fracturing process parameter optimization method, is characterized in that, comprises the following steps: 收集多个不同区块的深层页岩的基本参数,所述基本参数包括脆性矿物、泊松比、弹性模量、抗张强度、抗剪强度、断裂韧性、天然裂缝、力学脆性指数、最大水平主应力、最小水平主应力、水平地应力差和垂向地应力差;Collect basic parameters of deep shale from multiple different blocks, including brittle minerals, Poisson's ratio, elastic modulus, tensile strength, shear strength, fracture toughness, natural fractures, mechanical brittleness index, maximum level Principal stress, minimum horizontal principal stress, horizontal in-situ stress difference and vertical in-situ stress difference; 采用模糊聚类法对所述深层页岩进行储层划分;所述模糊聚类法采用系统聚类法,具体包括以下步骤:The deep shale is divided by a fuzzy clustering method; the fuzzy clustering method adopts a systematic clustering method, which specifically includes the following steps: 设置所述基本参数的数据矩阵,并对矩阵中的数据进行数据标准化,将数据压缩到[0,1]区间上;所述数据矩阵建立方法如下:设论域U={x1,x2,……,xn}为被分类对象,每个对象有m个指标表示其性状,即:Set the data matrix of the basic parameters, standardize the data in the matrix, and compress the data to the [0,1] interval; the method for establishing the data matrix is as follows: set the universe of discourse U={x 1 , x 2 ,...,x n } is the object to be classified, and each object has m indicators to represent its characteristics, namely: xi={xi1,xi2,…,xim}且i=1,2,…,n (1)x i ={x i1 ,x i2 ,...,x im } and i=1,2,...,n (1) 于是,得到所述数据矩阵为:Then, the data matrix is obtained as:
Figure FDA0003535260480000011
Figure FDA0003535260480000011
其中xnm表示第n个分类对象的第m个指标的原始数据;where x nm represents the raw data of the mth index of the nth classification object; 建立模糊相似矩阵,计算每个区块的相似程度;Establish a fuzzy similarity matrix to calculate the similarity of each block; 根据所述相似程度的计算结果,对每个区块进行聚类;According to the calculation result of the similarity degree, cluster each block; 利用软件模拟水力裂缝形态,根据模拟结果确定划分后的每类储层的水力压裂目标;Use software to simulate the shape of hydraulic fractures, and determine the hydraulic fracturing targets for each type of reservoir after division according to the simulation results; 根据所述水力压裂目标,采用数值模拟方法模拟不同施工参数的水力压裂结果,根据所述水力压裂结果优化施工参数。According to the hydraulic fracturing target, a numerical simulation method is used to simulate the hydraulic fracturing results of different construction parameters, and the construction parameters are optimized according to the hydraulic fracturing results.
2.根据权利要求1所述的深层页岩压裂工艺参数优化方法,其特征在于,所述数据标准化采用平移、极差变换法实现,该方法具体如下:2. deep shale fracturing process parameter optimization method according to claim 1, is characterized in that, described data standardization adopts translation, range transformation method to realize, and this method is specifically as follows:
Figure FDA0003535260480000012
Figure FDA0003535260480000012
Figure FDA0003535260480000013
Figure FDA0003535260480000013
Figure FDA0003535260480000014
Figure FDA0003535260480000014
Figure FDA0003535260480000021
Figure FDA0003535260480000021
式中:0≤x″ik≤1,且消除了量纲的影响。In the formula: 0≤x″ ik ≤1, and the influence of dimension is eliminated.
3.根据权利要求1所述的深层页岩压裂工艺参数优化方法,其特征在于,所述相似程度通过相似系数法或距离法进行计算,所述相似系数法包括夹角余弦法、最大最小法、算数平均最小法、几何平均最小法、数量积法、相关系数法、指数相似系数法,所述距离法包括直接距离法、海明距离法、欧几里得距离法、切比雪夫距离法、倒数距离法、指数距离法。3. The deep shale fracturing process parameter optimization method according to claim 1, wherein the similarity degree is calculated by a similarity coefficient method or a distance method, and the similarity coefficient method includes an angle cosine method, a maximum and a minimum method, arithmetic mean minimum method, geometric mean minimum method, quantity product method, correlation coefficient method, exponential similarity coefficient method, the distance methods include direct distance method, Hamming distance method, Euclidean distance method, Chebyshev distance method method, reciprocal distance method, exponential distance method. 4.根据权利要求1所述的深层页岩压裂工艺参数优化方法,其特征在于,采用基于模糊等价矩阵聚类方法或直接聚类法进行聚类,聚类时需确定阈值,所述阈值通过F统计量进行确定。4. The deep shale fracturing process parameter optimization method according to claim 1, characterized in that, clustering is performed based on a fuzzy equivalent matrix clustering method or a direct clustering method, and a threshold value needs to be determined during clustering, and the The threshold is determined by the F statistic. 5.根据权利要求1所述的深层页岩压裂工艺参数优化方法,其特征在于,所述施工参数包括簇数、排量和用液强度。5 . The method for optimizing process parameters of deep shale fracturing according to claim 1 , wherein the construction parameters include the number of clusters, the displacement and the strength of the liquid used. 6 .
CN202010716805.1A 2020-07-23 2020-07-23 A method for optimizing process parameters of deep shale fracturing Active CN111797575B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010716805.1A CN111797575B (en) 2020-07-23 2020-07-23 A method for optimizing process parameters of deep shale fracturing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010716805.1A CN111797575B (en) 2020-07-23 2020-07-23 A method for optimizing process parameters of deep shale fracturing

Publications (2)

Publication Number Publication Date
CN111797575A CN111797575A (en) 2020-10-20
CN111797575B true CN111797575B (en) 2022-05-20

Family

ID=72828318

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010716805.1A Active CN111797575B (en) 2020-07-23 2020-07-23 A method for optimizing process parameters of deep shale fracturing

Country Status (1)

Country Link
CN (1) CN111797575B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115099164B (en) * 2022-08-26 2022-12-30 中国科学技术大学 Machine learning-based tin-based perovskite thin film transistor optimization method and device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9201164B2 (en) * 2011-11-15 2015-12-01 Chevron U.S.A. Inc. System and method of using spatially independent subsets of data to calculate property distribution uncertainty of spatially correlated reservoir data
CN105156103B (en) * 2015-09-29 2018-06-05 西南石油大学 A kind of multiple dimensioned shale reservoir three-dimensional compressibility evaluation method of landwaste-rock core-wellbore-reservoir
CN105201479B (en) * 2015-10-09 2017-10-24 西南石油大学 A kind of horizontal well on shale reservoir stratum staged fracturing perforation cluster method for optimally designing parameters
CN107387051B (en) * 2017-09-05 2020-03-27 西南石油大学 Repeated fracturing well selection method for multi-stage fractured horizontal well with low-permeability heterogeneous oil reservoir

Also Published As

Publication number Publication date
CN111797575A (en) 2020-10-20

Similar Documents

Publication Publication Date Title
CN108009716B (en) A multi-level evaluation method for influencing factors of volume fracturing effect in horizontal wells
CN108960651B (en) A comprehensive evaluation method for multi-stage fracturing horizontal well completion efficiency in tight oil and gas reservoirs
CN111046341B (en) A Method of Unconventional Natural Gas Fracturing Effect Evaluation and Productivity Prediction Based on Principal Component Analysis
CN108446797B (en) Method for predicting capacity of tight oil reservoir horizontal well at initial volume fracturing stage
Dongqi et al. Fine classification of ultra-low permeability reservoirs around the Placanticline of Daqing oilfield (PR of China)
CN105447762B (en) A kind of calculation method of the low-permeability oil deposit water logging information of fluid replacement
CN112576238B (en) A system, method and application for determining the position and content of remaining oil in a low-permeability reservoir
CN109882163A (en) A productivity prediction method for fracturing horizontal wells in tight oil reservoirs
CN108229811A (en) A kind of method for evaluating fracture and vug carbonate reservoir flood effectiveness
CN108643875A (en) Water injection development adjustment method, device and system for low-permeability clastic rock oil reservoir
CN111797575B (en) A method for optimizing process parameters of deep shale fracturing
CN114460639A (en) Method and device for predicting permeability of shale oil reservoir
CN111242490A (en) Low-permeability reservoir characteristic master control factor extraction method and device and readable storage medium
Li et al. Productivity forecast for multi-stage fracturing in shale gas wells based on a random forest algorithm
CN114943015A (en) Oil reservoir analogy method and device based on similarity calculation model
Li et al. Flow field characterization and evaluation method based on unsupervised machine learning
CN115455628A (en) Classified evaluation method for residual oil in ultra-high water-cut period
CN106968668B (en) Score type well logging lithology optimization interpretation method
Yang et al. Multifactor Coupled Evaluation Model of Fracability Based on a Combinatorial Weighting Method: A Case Study of the Low-Permeability Sandstone Reservoirs of Yanchang Formation, Southwestern Ordos Basin, China
CN115270959A (en) Shale Lithology Identification Method and Device Based on Recursive Feature Elimination and Fusion Random Forest
Chen et al. Evaluation of coalbed methane resources in Xinjing Baoan block based on PCA, TOPSIS, & MLFM
Xu et al. A lightweight model hyperparameters searching method for fast, accurate and on-site lithology identification
Lu et al. Main control factors affecting mechanical oil recovery efficiency in complex blocks identified using the improved k-means algorithm
Guo et al. Study on gradational optimization of oil reservoir streamline field based on an artificial intelligence algorithm
Zhao et al. Prediction of Collapsibility of Loess of Construction Sites in Xining Based on Machine Learning Methods

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