CN114638147A - Method for determining fracturing process parameters of oil and gas reservoir - Google Patents

Method for determining fracturing process parameters of oil and gas reservoir Download PDF

Info

Publication number
CN114638147A
CN114638147A CN202011492514.5A CN202011492514A CN114638147A CN 114638147 A CN114638147 A CN 114638147A CN 202011492514 A CN202011492514 A CN 202011492514A CN 114638147 A CN114638147 A CN 114638147A
Authority
CN
China
Prior art keywords
fracturing
well layer
parameters
factors
post
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
Application number
CN202011492514.5A
Other languages
Chinese (zh)
Inventor
李克智
姚昌宇
蒋艳芳
蔺研锋
赵治钢
王帆
魏琪
梁志彬
陆姣平
贺宇廷
邹佳玲
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.)
China Petroleum and Chemical Corp
Sinopec North China Oil and Gas Co
Original Assignee
China Petroleum and Chemical Corp
Sinopec North China Oil and Gas Co
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 China Petroleum and Chemical Corp, Sinopec North China Oil and Gas Co filed Critical China Petroleum and Chemical Corp
Priority to CN202011492514.5A priority Critical patent/CN114638147A/en
Publication of CN114638147A publication Critical patent/CN114638147A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • E—FIXED CONSTRUCTIONS
    • E21—EARTH OR ROCK DRILLING; MINING
    • E21B—EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS
    • E21B43/00—Methods or apparatus for obtaining oil, gas, water, soluble or meltable materials or a slurry of minerals from wells
    • E21B43/25—Methods for stimulating production
    • E21B43/26—Methods for stimulating production by forming crevices or fractures
    • 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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Mining & Mineral Resources (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Evolutionary Computation (AREA)
  • Tourism & Hospitality (AREA)
  • Geology (AREA)
  • General Business, Economics & Management (AREA)
  • Animal Husbandry (AREA)
  • Game Theory and Decision Science (AREA)
  • Agronomy & Crop Science (AREA)
  • Operations Research (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Quality & Reliability (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Environmental & Geological Engineering (AREA)
  • Fluid Mechanics (AREA)
  • Development Economics (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • Geochemistry & Mineralogy (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to a method for determining fracturing process parameters of an oil and gas reservoir, and belongs to the technical field of petroleum and natural gas reservoir transformation. The method comprises the steps of obtaining historical fracturing data, establishing a post-fracturing productivity prediction model according to the historical fracturing data, selecting a well layer closest to inherent parameters of a target well layer from the historical fracturing data as a similar well layer, and determining fracturing parameters of the target well layer by using the fracturing data of the similar well layer, the inherent parameters of the target well layer and the post-fracturing productivity prediction model. The invention fully considers the historical fracturing data, solves the problem that the independent fracturing design can not learn and reference the historical successful cases, improves the fracturing effect and realizes the purpose of increasing the production.

Description

Method for determining fracturing process parameters of oil and gas reservoir
Technical Field
The invention relates to a method for determining fracturing process parameters of an oil and gas reservoir, and belongs to the technical field of petroleum and natural gas reservoir transformation.
Background
In the development of low-permeability, ultra-low permeability or middle-high permeability oil and gas fields polluted near a well bore, in order to obtain industrial oil and gas flow, proper fracturing stimulation measures must be taken on the reservoirs. In order to realize better fracturing yield increase effect as far as possible and obtain better fracture geometric parameters and post-fracturing yield, the fracturing construction parameters need to be optimally designed before hydraulic fracturing construction, and an economic and effective optimal scheme meeting geological, engineering and equipment conditions is sought.
However, the current crack extension model research still cannot reflect the real situation of crack extension after reservoir lamination more accurately. In addition, the extremely high solving difficulty of part of models reflects the great difficulty of simulating the actual situation by establishing a crack extension model from the side, and the conventional crack extension model research has the problems of too many model assumptions and the fact that the crack extension after the storage layer lamination cannot be reflected more accurately. In addition, the existing fracturing parameter design is used as an independent fracturing design process, and the successful cases existing in the past or in history are not learned and referred, so that the aim of maximizing the yield is not facilitated.
Disclosure of Invention
The invention aims to provide a method for determining fracturing process parameters of an oil and gas reservoir, which aims to solve the problems that historical data is not used for reference and yield increase is not utilized in the conventional fracturing process parameter determination.
The invention provides a method for determining fracturing process parameters of an oil and gas reservoir for solving the technical problems, which comprises the following steps:
1) obtaining historical fracturing data which comprises post-fracturing capacity and various influence factors influencing the post-fracturing capacity, determining the influence degree of the influence factors on the post-fracturing capacity, screening out main influence factors according to the influence degrees, and constructing a fracturing sample parameter database according to the screened main influence factors;
2) establishing a post-pressure energy production prediction model, and training the post-pressure energy production prediction model by using a fracturing sample parameter database;
3) acquiring stratum static factors in main influence factors of the target well layer, and finding out similar well layers from a fracturing sample parameter database;
4) selecting the fracturing parameters of similar well layers as constraint conditions of the fracturing parameters of the target well layer, adjusting the fracturing parameters of the target well layer under the constraint conditions, inputting the obtained static factors of the stratum of the target well layer and the fracturing parameters selected under the constraint conditions into a post-fracturing energy production model, and selecting the fracturing parameters which enable the post-fracturing energy production to be maximum as the finally determined fracturing parameters.
The method comprises the steps of obtaining historical fracturing data, establishing a post-fracturing productivity prediction model according to the historical fracturing data, selecting a well layer closest to inherent parameters of a target well layer from the historical fracturing data as a similar well layer, and determining fracturing parameters of the target well layer by using the fracturing data of the similar well layer, the inherent parameters of the target well layer and the post-fracturing productivity prediction model. The invention fully considers the historical fracturing data, solves the problem that the independent fracturing design can not learn and reference the historical successful cases, improves the fracturing effect and realizes the purpose of increasing the production.
Further, the post-compression performance prediction model in the step 2) is established by using a support vector machine.
Further, in order to accurately and quickly screen out the main influence factors, the screening process of the main influence factors in the step 1) is as follows:
a. obtaining historical fracturing data, forming original number series, and selecting one number series from the original number series as a reference number series X0The other series being comparison series XiCarrying out dimensionless treatment on the influence factors in the number series;
wherein the comparison sequence and the reference sequence are respectively expressed as:
Xi＝{Xi(1),Xi(2),Xi(3),…,Xi(n)},(i＝1,2,…,m)
X0＝{X0(1),X0(2),X0(3),…X0(n)}
m is the number of influencing factors, n is the number of samples, Xi(n) represents the ith influencing factor in the nth sample;
b. calculating a correlation coefficient between the comparison sequence and the reference sequence after the non-dimensionalization processing;
c. and calculating the association degree of each influence factor according to the association coefficient, determining the weight of each influence factor according to the association degree, and selecting the influence factor with the weight larger than a set threshold value as a main influence factor.
Further, the correlation coefficient in the step b is calculated by using a dune correlation model, wherein a resolution coefficient selection principle is as follows:
when deltamax＞3ΔyWhen is e.gΔ≤ρ＜1.5∈Δ(ii) a When deltamax≤3ΔyWhen, 1.5 eΔ≤ρ≤2∈Δ；
Figure BDA0002841148020000031
Figure BDA0002841148020000032
Wherein Y isi(k) Is a non-dimensionalized comparison sequence, is a non-dimensionalized reference sequence, n is the number of samples, m is the number of influencing factors, DeltamaxAnd p is the absolute difference of the maximum sample data in all comparison arrays, and is a resolution coefficient.
Further, a decision function adopted for establishing the post-pressure performance prediction model by using the support vector machine is as follows:
Figure BDA0002841148020000033
wherein l is the number of samples, K (x)iU) as a support vector machine kernel function, b)*In order to support the constant term of the vector machine,
Figure BDA0002841148020000034
and alphaiIs a lagrange multiplier.
Further, in order to ensure the precision of the post-pressure production performance prediction model, the post-pressure production performance prediction model divides the fracturing sample parameter database into training samples and testing samples according to a set proportion during training, and ensures that the training samples and the testing samples simultaneously meet the precision requirement.
Further, in order to accurately and quickly determine the similar well layers, the determination process of the similar well layers in the step 3) is as follows:
A. calculating the grey correlation coefficient of the static factors of each stratum;
B. calculating the difference of any stratum static factor between the target well layer and each well layer in the database according to the grey correlation coefficient of each stratum static factor, each stratum static factor in the target well layer and the stratum static factor of each well layer in the fracturing sample parameter database;
C. and summing the differences of the static factors of the stratums, and selecting the well layer with the smallest difference with the target well layer from the well layers of the database, wherein the well layer is a similar well layer.
Drawings
FIG. 1 is a flow chart of a method of determining parameters of a fracturing process for a hydrocarbon reservoir in accordance with the present invention;
FIG. 2 is a schematic diagram of the training process of the post-compression performance prediction model of the present invention.
Detailed Description
The following further describes embodiments of the present invention with reference to the drawings.
The method comprises the steps of firstly, obtaining historical fracturing data, determining the influence degree of each influence factor on the magnitude of the post-fracturing energy, screening out main influence factors according to the influence degrees, and constructing a fracturing sample parameter database according to the screened main influence factors; then establishing a post-pressure productivity prediction model, and training the post-pressure productivity prediction model by utilizing a fracturing sample parameter database; acquiring stratum static factors in main influence factors of the target well layer, and finding out similar well layers from a fracturing sample parameter database; and selecting the fracturing parameters of the similar well layers as constraint conditions of the fracturing parameters of the target well layer, and selecting the fracturing parameter which enables the maximum post-fracturing yield as the finally determined fracturing parameter by using the trained post-fracturing yield prediction model. The implementation process of the method is shown in fig. 1, and the specific process is as follows.
1. Analyzing and weighting the influence factors of the historical fracturing data capacity to obtain the influence degrees of different influence factors on the post-fracturing capacity, screening out the main influence factors according to the influence degrees, and constructing a fracturing sample parameter database.
1) And acquiring historical fracturing data, including the post-fracturing productivity and various influencing factors influencing the post-fracturing productivity, and preprocessing the acquired historical fracturing data.
The factors influencing the post-fracturing productivity mainly include geological factors, well factors, fracturing parameter factors and the like, the factors are collectively called as influencing factors, n samples of the obtained historical fracturing data are assumed, and each fracturing data sample comprises m influencing factors which are called as an original number sequence.
Randomly selecting fracture data of one influence factor from each sample as a reference number sequence, wherein the reference number sequence selected in the embodiment can be represented as follows:
X0＝{X0(1),X0(2),X0(3),…X0(n)}
in this case, the other number sequence is called a comparison number sequence, and the comparison number sequence can be expressed as:
Xi＝{Xi(1),Xi(2),Xi(3),…,Xi(n)},(i＝1,2,…,m)
in the formula: i is the serial number of the comparison sequence; m is the number of independent variable factors; n is the number of data samples in the array.
To facilitate the subsequent calculation of the degree of correlation (or similarity) between the comparison sequence and the reference sequence, it is necessary to locate the two sequence curves in the same reference coordinate system, which requires that the data in the original sequence have the same order and dimension, otherwise, the two sequences are the irrelative sequences. Because the measurement units of all factors in the original data sequence are different, different data sequences have different orders and dimensions, the original data needs to be subjected to non-dimensionalization, and the non-dimensionalized data uses Yi(k) To represent; k is the sample number in the data sequence, k is 1,2, …, n.
2) Calculating a correlation coefficient between the comparison number series and the reference number series based on the original number series after the non-dimensionalization processing.
In this embodiment, a dune correlation model is used to calculate the correlation coefficient, and the specific calculation process is as follows:
the absolute difference between each sample value in the comparison array (ith) and the corresponding sample value in the reference array is represented by:
Δ0i(k)＝|Yi(k)-Y0(k)|
the absolute difference between the maximum and minimum sample data in all comparison series is
Figure BDA0002841148020000051
Figure BDA0002841148020000052
The Deng correlation coefficient is expressed as (i th comparison number series and reference number series, k th sample)
Figure BDA0002841148020000061
In the formula: ρ is a resolution coefficient, ρ ∈ (0, 1).
According to the resolution coefficient selection principle, the condition that the sequence abnormal value dominates the system correlation value is considered at the same time, and the mean value of all the difference absolute values is recorded as deltayExpressed by the following formula:
Figure BDA0002841148020000062
note the book
Figure BDA0002841148020000063
The value of the resolution coefficient is eΔ≤ρ≤2∈ΔAnd satisfies the relationship: when Δmax＞3ΔyWhen is e.gΔ≤ρ＜1.5∈Δ(ii) a When deltamax≤3ΔyWhen, 1.5 eΔ≤ρ≤2∈Δ。
3) And calculating the association degree of each influence factor according to the association coefficient.
The information reflected by the single correlation coefficient has dispersity, and if the correlation degree between the comparison array and the reference array is only expressed by the n correlation coefficients, the influence degree of each factor on the productivity cannot be reflected. Therefore, it is necessary to collectively process the correlation information and quantitatively reflect the degree of correlation between the arrays by using the average value thereof.
Figure BDA0002841148020000064
r0iThe correlation degree between the ith comparison number sequence and the reference number sequence is set, and the variation rule of the comparison number sequence and the reference number sequence is more similar when the correlation degree is higher.
4) Determining the weight of each influence factor, and screening the influence factors according to the weight.
Normalizing the obtained association degree to obtain the weight of the association degree of each comparison array (each influence factor), and using WiAnd (4) showing.
Figure BDA0002841148020000065
And arranging the association degrees or the weights according to the magnitude sequence to form an association sequence so as to reflect the primary and secondary relations among all the influence factors, wherein the influence of the influence factors closer to the front on the post-pressure production performance is larger, so that the influence factors of the set number closer to the front (or the weights are larger than a set threshold) are selected as main influence factors, and the main influence factors in the original sequence are screened out to construct a fracturing sample parameter database.
The implementation analyzes the influence weight of each influence factor of the post-pressure production energy through a big data method to obtain the main influence factor, finds that the static factor of the stratum is an important factor influencing the fracturing effect, but the fracturing construction parameters are closely related to the post-pressure production energy. The main influence factors screened out are 24, and the significance ranks of the 24 influence factors are as follows: the sand adding strength per meter, the prepad fluid ratio, the effective thickness of a reservoir, the Young modulus, the thickness of large and medium conglomerates, the thickness of fine sandstones, the brittleness index, the thickness of small and medium conglomerates, the horizontal stress difference, the thickness of coarse sandstones, the permeability, the neutron porosity, the construction displacement, the oil saturation, the minimum horizontal principal stress, the pore structure index, the vertical stress, the acoustic time difference, the Poisson ratio, the argillaceous content, the resistivity, the average sand ratio, the porosity and the density.
2. And establishing a post-pressure energy production prediction model, and training the model by using a fracturing sample parameter database.
The post-compaction productivity prediction model can be established by adopting a machine learning algorithm, a support vector machine is used as the post-compaction productivity prediction model in the embodiment, and when the support vector machine is applied to productivity prediction, the productivity of different well layers needs to be divided into different grades through a support vector classifier, so that the purpose of primary classification is achieved; and then predicting the post-compression performance by a support vector regression machine. Therefore, before the model is trained, the productivity data in the fracturing sample parameter database (database for short) needs to be graded, the invention adopts the rice fluid production index to represent the productivity, and the grading standard shown in the table 1 is adopted to grade the productivity according to the actual situation of the productivity data in the database.
TABLE 1
Figure BDA0002841148020000071
The determination of the capacity level is a classification problem. The support vector classifier cannot handle the classification problem of six classes simultaneously, so the pairwise classifier is chosen to classify. The method derives from a two-class problem, the decision function is not unique. When there are n classes, the pairwise classification method needs to construct n (n-1)/2 decision functions, and has the same number of requirements for the support vector machine.
The rank of the division takes on the value {1,2,3,4,5,6}, and (i, j) ∈ { (i, j) | i < j, i, j { (i, j) | 1,2,3,4,5,6 }. And extracting all data points of y-i and y-j, and constructing different support vector classification models. And inputting the independent variable of the test sample into different decision functions to obtain different classification results. And scoring the classified categories of the test samples, then summarizing the results of all decision functions, and taking the category with the highest score as a final classification label of the test samples. In the present embodiment, the yield rating evaluation problem has six categories, and 15 decision functions must be constructed to solve the yield classification problem through calculation. And (3) training the support vector machine model by using the database established in the step 1, wherein the training process is as shown in fig. 2.
Assuming that the fracture data samples in the database in this embodiment include the known main influential factor data of 53 well layers in the target block, the database is now divided into two parts, i.e., a training sample and a test sample, according to a specific ratio before training. A typical division is that the training set accounts for 70% of the total samples and the test set accounts for 30%, so the number of samples is determined to be 37 training samples and 16 test samples.
Firstly, assigning the post-pressure performance influence factors of the training samples to independent variables of a prediction model, assigning the post-pressure performance influence factors of the training samples to dependent variables of the prediction model, training a support vector regression machine, and obtaining an expression of a decision function through training; then fitting the historical data of the productivity of the training sample according to the obtained decision function expression; on the basis of historical data fitting, denoising processing is carried out on the data of the fitting result, and a large amount of information loss is avoided while smooth denoising is carried out as much as possible; finally, judging whether the error of the denoised analysis sample meets the requirement, selecting a relative error (namely an absolute value of a ratio of a difference value between a predicted value and a true value to the true value) when the error is analyzed, and if the error meets the requirement, determining a capacity prediction model, and predicting the post-compression capacity on the basis of the model; if the requirements are not met, the support vector regression machine is retrained by adjusting the parameters until the requirements are met.
It is noted that the final capacity prediction model can only be determined if the training samples and the test samples meet the accuracy requirements at the same time.
The training of the productivity prediction model can be realized through the training process.
3. And acquiring geological factors and well factors in main influence factors of the target well layer, and accordingly finding similar well layers from the database.
And the determination of similar reservoirs has important reference significance for fracturing construction parameter optimization of the target well layer. The similar reservoir judgment is based on reservoir conditions, wherein 20 stratum static factors including reservoir geological factors and well factors are selected from 24 main factors to serve as judgment factors, and a determination method of the similar reservoir is provided based on a grey correlation analysis method in consideration of differences of different factors.
The 20 formation static factors can be obtained through geological exploration and well logging data, therefore, for a target well layer, the geological factors and well factors of the main influencing factors are known and can be obtained through the existing geological exploration means and well logging means, and the 20 main factors are as follows: reservoir effective thickness, young's modulus, large and medium conglomerate thickness, fine sandstone thickness, brittleness index, small and medium conglomerate thickness, horizontal stress difference, coarse sandstone thickness, permeability, neutron porosity, oil saturation, minimum horizontal principal stress, pore structure index, vertical stress, acoustic moveout, poisson's ratio, shale content, resistivity, porosity, and density.
Setting the set of grey correlation coefficients of static factors of each stratum as r, and setting stratum parameters of two different well layers as X1，X2Then they can be expressed as:
r＝{r1,r2,…,rs}
X1＝{X1(1),X1(2),…,X1(s)}
X2＝{X2(1),X2(2),…,X2(s)}
in the formula: s is the number of static factors of the formation, and s is equal to 20 in the embodiment.
Subtracting the parameters of the target well layer from the corresponding parameters of each well layer in the database, taking an absolute value, and multiplying the grey correlation coefficient of the factor to obtain the difference delta of any stratum factor between the target well layer and each well layer in the databaseiExpressed as:
Δi＝ri·|X1(i)-X2(i)|
the differences between the different formation factors are summed and expressed as:
Figure BDA0002841148020000091
delta, which is the sum of the differences between different stratigraphic factors, describes the difference in reservoir conditions between two well layers, with the smaller the delta, the more similar the reservoir conditions. Comparing and analyzing the stratum static factors of the target well layer with each well layer in the fracturing database, wherein the well layer with the smallest delta is the well layer with the reservoir conditions most similar to the reservoir conditions of the target well layer, namely the similar well layer, and acquiring the fracturing parameters corresponding to the similar well layer.
4. Selecting the fracturing parameters of similar well layers to determine the constraint conditions of the fracturing parameters of the target well layer, adjusting the fracturing parameters of the target well layer under the constraint conditions, inputting the obtained static factors of the stratum of the target well layer and the fracturing parameters selected under the constraint conditions into a post-pressure energy production model, and selecting the fracturing parameters which enable the post-pressure energy production to be maximum as the optimized fracturing parameters.
The research on the productivity prediction model of the support vector machine determines that the decision function is the following formula, and the kernel function is selected as the following formula. Order to
Figure BDA0002841148020000092
At this point, the decision function is expressed as:
Figure BDA0002841148020000101
the fracturing process is optimized, essentially, the purpose of maximizing the post-fracturing productivity is achieved by continuously changing the values of a plurality of construction parameters, namely, the maximum value is reached under the constraint condition, and the construction parameter corresponding to the maximum predicted productivity is the optimal construction parameter.
In the process of establishing the optimization problem, the inherent parameters (stratum static factors) of the reservoir are separated from the fracturing construction parameters, and then if all the productivity influence factors are expressed as:
xi＝(xi1,xi2,…,xi24)
then xi1～xi20Represents 20 formation static factors, in miRepresents this set, and xi21～xi24Representing fracturing construction parameters by niTo represent this set. Namely:
mi＝(xi1,xi2,…,xi20)
ni＝(xi21,xi18,…,xi24)
in the same way, order
p＝(u1,u2,…,u20)
q＝(u21,u22,u23,u24)＝(q1,q2,q3,q4)
In the formula: p represents an intrinsic parameter; q represents a construction parameter, independent variable; q. q of1、q2、q3、q4Respectively representing the pre-posed liquid ratio, the construction displacement, the sand adding strength per meter and the average sand ratio.
And writing the target function into the following form after substituting the kernel function Gaussian radial basis function:
Figure BDA0002841148020000102
let F (q) be f (u) and let F (q) be f (u) at the same time
Figure BDA0002841148020000103
The objective function is further expressed as:
Figure BDA0002841148020000104
the optimization problem with the fracture construction parameters q is written in the form:
Figure BDA0002841148020000111
Figure BDA0002841148020000112
the constraint condition is that the fracturing parameter of the target well layer is greater than or equal to the minimum value of the same parameter of the similar well layer and less than or equal to the maximum value of the same parameter of the similar well layer. The method comprises the following steps: firstly, finding out a similar well layer of a target well layer, counting the maximum value and the minimum value of fracturing parameters of the similar well layer, and taking the maximum value and the minimum value as constraint conditions.
Solving for the optimal solution
Figure BDA0002841148020000113
Namely the optimal fracturing construction parameters.
In order to prove the feasibility of the method, the method is applied to an example well layer AH2-1, and verified by predicting the unobstructed flow and analyzing simulated cracks, and the test well layer is found to be based on optimized fracturing construction parameters (specifically, the pad fluid ratio is 32.7%, and the construction displacement is 3.5 m)3Min, sand strength per meter 11.3m3And the average sand ratio is 19.8%), better fracture size, proppant laying and flow conductivity distribution can be realized, the fracturing effect is obviously improved, and the effectiveness of the fracturing construction parameter optimization method is demonstrated. The average relative error of the post-compression performance prediction model is 9.50%, and good prediction accuracy can be achieved.

Claims (7)

1. A method for determining fracturing process parameters of an oil and gas reservoir is characterized by comprising the following steps:
1) obtaining historical fracturing data, including post-fracturing productivity and various influence factors influencing the post-fracturing productivity, determining the influence degree of the influence factors on the post-fracturing productivity, screening out main influence factors according to the influence degrees, and constructing a fracturing sample parameter database according to the screened main influence factors;
2) establishing a post-pressure energy production prediction model, and training the post-pressure energy production prediction model by using a fracturing sample parameter database;
3) acquiring stratum static factors in main influence factors of the target well layer, and finding out similar well layers from a fracturing sample parameter database;
4) selecting the fracturing parameters of similar well layers as constraint conditions of the fracturing parameters of the target well layer, adjusting the fracturing parameters of the target well layer under the constraint conditions, inputting the obtained static factors of the stratum of the target well layer and the fracturing parameters selected under the constraint conditions into a post-fracturing energy production model, and selecting the fracturing parameters which enable the post-fracturing energy production to be maximum as the finally determined fracturing parameters.
2. The method for determining the fracturing process parameters of the oil and gas reservoir according to claim 1, wherein the predicted after-pressure production model in the step 2) is established by using a support vector machine.
3. The method for determining the parameters of the fracturing process of the oil and gas reservoir as claimed in claim 1 or 2, wherein the screening process of the main influencing factors in the step 1) is as follows:
a. obtaining historical fracturing data, forming original number series, and selecting one number series from the original number series as a reference number series X0The other series being comparison series XiCarrying out dimensionless treatment on the influence factors in the number series;
wherein the comparison sequence and the reference sequence are respectively expressed as:
Xi＝{Xi(1),Xi(2),Xi(3),…,Xi(n)},(i＝1,2,…,m)
X0＝{X0(1),X0(2),X0(3),…X0(n)}
m is the number of influencing factors, n is the number of samples, Xi(n) represents the ith influencing factor in the nth sample;
b. calculating a correlation coefficient between the comparison series after the non-dimensionalization processing and the reference series;
c. and calculating the association degree of each influence factor according to the association coefficient, determining the weight of each influence factor according to the association degree, and selecting the influence factor with the weight larger than a set threshold value as a main influence factor.
4. The method for determining the parameters of the hydrocarbon reservoir fracturing process according to claim 3, wherein the correlation coefficient in the step b is calculated by using a Duncus correlation model, and the resolution coefficient is selected according to the following principle:
when deltamax＞3ΔyWhen is e.gΔ≤ρ＜1.5∈Δ(ii) a When deltamax≤3ΔyAt 1.5 epsilonΔ≤ρ≤2∈Δ；
Figure FDA0002841148010000021
Figure FDA0002841148010000022
Wherein Y isi(k) Is a non-dimensionalized comparison sequence, is a non-dimensionalized reference sequence, n is the number of samples, m is the number of influencing factors, DeltamaxAnd p is the absolute difference of the maximum sample data in all comparison arrays, and is a resolution coefficient.
5. The method of claim 2, wherein the decision function used to build the post-pressure production prediction model using a support vector machine is:
Figure FDA0002841148010000023
wherein l is the number of samples, K (x)iU) is a support vector machine kernel function, b*In order to support the constant term of the vector machine,
Figure FDA0002841148010000024
Figure FDA0002841148010000025
and alphaiIs a lagrange multiplier.
6. The method for determining the parameters of the hydrocarbon reservoir fracturing process of claim 2, wherein the post-fracturing energy production prediction model divides the fracturing sample parameter database into training samples and testing samples according to a set proportion during training, and ensures that the training samples and the testing samples simultaneously meet the precision requirement.
7. The method for determining the parameters of the fracturing process of the oil and gas reservoir as claimed in claim 1 or 2, wherein the determination process of the similar well layers in the step 3) is as follows:
A. calculating the grey correlation coefficient of the static factors of each stratum;
B. calculating the difference of any stratum static factor between the target well layer and each well layer in the database according to the grey correlation coefficient of each stratum static factor, each stratum static factor in the target well layer and the stratum static factor of each well layer in the fracturing sample parameter database;
C. and summing the differences of the static factors of the stratums, and selecting the well layer with the smallest difference with the target well layer from the well layers of the database, wherein the well layer is a similar well layer.
CN202011492514.5A 2020-12-16 2020-12-16 Method for determining fracturing process parameters of oil and gas reservoir Pending CN114638147A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011492514.5A CN114638147A (en) 2020-12-16 2020-12-16 Method for determining fracturing process parameters of oil and gas reservoir

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011492514.5A CN114638147A (en) 2020-12-16 2020-12-16 Method for determining fracturing process parameters of oil and gas reservoir

Publications (1)

Publication Number Publication Date
CN114638147A true CN114638147A (en) 2022-06-17

Family

ID=81945515

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011492514.5A Pending CN114638147A (en) 2020-12-16 2020-12-16 Method for determining fracturing process parameters of oil and gas reservoir

Country Status (1)

Country Link
CN (1) CN114638147A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115310357A (en) * 2022-08-09 2022-11-08 大庆正方软件科技股份有限公司 Fracturing analysis method based on data-driven decision
CN115577645A (en) * 2022-12-08 2023-01-06 中国石油大学(华东) Construction method and prediction method of combustion and explosion fracturing fracture range prediction model

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115310357A (en) * 2022-08-09 2022-11-08 大庆正方软件科技股份有限公司 Fracturing analysis method based on data-driven decision
CN115577645A (en) * 2022-12-08 2023-01-06 中国石油大学(华东) Construction method and prediction method of combustion and explosion fracturing fracture range prediction model
CN115577645B (en) * 2022-12-08 2023-04-18 中国石油大学(华东) Construction method and prediction method of combustion and explosion fracturing fracture range prediction model

Similar Documents

Publication Publication Date Title
CN107703560B (en) shale lithofacies fine identification method based on triple information
CN112989708B (en) Well logging lithology identification method and system based on LSTM neural network
CN105760673B (en) A kind of fluvial depositional reservoir seismic-sensitive parameterized template analysis method
CN111489034B (en) Construction method and application of oil and gas reservoir permeability prediction model
CN109800954B (en) Reservoir evaluation method based on logging data
US20090276157A1 (en) System and method for interpretation of well data
CN114638147A (en) Method for determining fracturing process parameters of oil and gas reservoir
Wang et al. Improved pore structure prediction based on MICP with a data mining and machine learning system approach in Mesozoic strata of Gaoqing field, Jiyang depression
CN110412662A (en) Method for prediction of reservoirs of thin interbeded based on seismic multi-attribute deep learning
CN107165621B (en) Method for identifying multi-well sedimentary microfacies by using clustering method
CN114818076A (en) Machine learning-based fault closed hydrocarbon column height evaluation method
CN113807021B (en) Gas well productivity grade prediction method based on index analysis and multi-model fusion
CN112576238B (en) System, method and application for determining position and content of residual oil in low-permeability reservoir
CN107133879B (en) Method for screening similar oil fields
CN116168224A (en) Machine learning lithology automatic identification method based on imaging gravel content
Liu et al. Identification of tight sandstone reservoir lithofacies based on CNN image recognition technology: A case study of Fuyu reservoir of Sanzhao Sag in Songliao Basin
Lu et al. Identifying flow units by FA-assisted SSOM—An example from the Eocene basin-floor-fan turbidite reservoirs in the Daluhu Oilfield, Dongying Depression, Bohai Bay Basin, China
CN117093922A (en) Improved SVM-based complex fluid identification method for unbalanced sample oil reservoir
CN116796231A (en) Method and device for automatically dividing lithology and comparing lithology based on logging curve
Abedini et al. Porosity classification from thin sections using image analysis and neural networks including shallow and deep learning in Jahrum formation
Anifowose et al. A functional networks-type-2 fuzzy logic hybrid model for the prediction of porosity and permeability of oil and gas reservoirs
Fan et al. Semi-Supervised Learning–Based Petrophysical Facies Division and “Sweet Spot” Identification of Low-Permeability Sandstone Reservoir
Palmer Using AI and Machine Learning to Indicate Shale Anisotropy and Assist in Completions Design
CN114755744A (en) Total organic carbon well logging interpretation method and system based on mud shale heterogeneity characteristics
CN104699977B (en) Using the method for weight quantitative analysis of mineral material resource

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