CN113032953A - Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system - Google Patents

Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system Download PDF

Info

Publication number
CN113032953A
CN113032953A CN202110105180.XA CN202110105180A CN113032953A CN 113032953 A CN113032953 A CN 113032953A CN 202110105180 A CN202110105180 A CN 202110105180A CN 113032953 A CN113032953 A CN 113032953A
Authority
CN
China
Prior art keywords
production
injection
water
reservoir
development
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
CN202110105180.XA
Other languages
Chinese (zh)
Inventor
徐建春
曲尚浩
樊灵
李航宇
王晓璞
刘树阳
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 University of Petroleum East China
Original Assignee
China University of Petroleum East China
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 University of Petroleum East China filed Critical China University of Petroleum East China
Priority to CN202110105180.XA priority Critical patent/CN113032953A/en
Publication of CN113032953A publication Critical patent/CN113032953A/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
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0206—Price or cost determination based on market factors
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Software Systems (AREA)
  • Strategic Management (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Finance (AREA)
  • Development Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Accounting & Taxation (AREA)
  • Computational Linguistics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Biomedical Technology (AREA)
  • General Business, Economics & Management (AREA)
  • Molecular Biology (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Biophysics (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Agronomy & Crop Science (AREA)
  • Animal Husbandry (AREA)
  • Game Theory and Decision Science (AREA)
  • Mining & Mineral Resources (AREA)
  • Human Resources & Organizations (AREA)
  • Primary Health Care (AREA)
  • Tourism & Hospitality (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Injection Moulding Of Plastics Or The Like (AREA)

Abstract

The invention discloses an intelligent optimization method for injection and production parameters of a water-drive reservoir of a multi-well system, and relates to the field of intelligent oil field development. The method comprises the following steps; determining injection and production parameters to be optimized and a value range of a multi-well system water drive oil reservoir, and generating a development scheme of n groups of different injection and production parameter combinations; constructing a multi-well system injection-production development oil reservoir geological model, and carrying out numerical simulation on n groups of different injection-production development schemes to obtain corresponding oil reservoir production net current values; constructing a proxy model for water-flooding oil reservoir development and production based on a multivariate self-adaptive spline regression algorithm and the like to realize that the net current value of oil reservoir production is the maximum optimization target, and optimizing injection and production parameters by using a particle swarm optimization algorithm and the like to obtain the optimal injection and production parameter combination; and optimizing and adjusting the oil reservoir water drive development strategy according to the obtained optimal injection-production parameter combination. The method overcomes the defects that the traditional water drive reservoir injection-production parameter optimization and adjustment method is too dependent on experience, consumes time and labor and is difficult to find a real optimal solution.

Description

Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system
Technical Field
The invention relates to the field of intelligent oil field development, in particular to an intelligent optimization method for injection and production parameters of a water-drive oil reservoir of a multi-well system.
Background
Because the water injection process is relatively simple and the water source is easy to obtain, the water injection development becomes the main development mode of the oil field in China at present. According to statistics, the reserves of domestic water injection oil fields account for about 80 percent of the total reserves of the oil fields.
However, with the continuous and deep development of water flooding, most oil fields enter the development stage of old oil fields with high water content and high extraction degree. The heterogeneity of oil fields is more and more serious, the difficulty of production technology is more and more high, and the exploitation cost is higher and higher.
As is well known, whether the injection-production parameter setting is reasonable in the water flooding process is the key influencing the final water flooding development effect. The adjustment and optimization of the injection and production parameters of the domestic traditional water drive reservoir mainly have two modes. First, empirical methods (reservoir engineering methods). An oil reservoir engineer manually designs a plurality of groups of injection-production parameter schemes with different values based on personal experience, and compares and optimizes the schemes by using an oil reservoir numerical simulation method. But such methods are too dependent on the experience of the engineer. For multi-well system reservoirs, numerical simulations tend to be time consuming and it is difficult to find the true best solution. Second, the experimental method (orthogonal design of experiments). And selecting representative injection and production parameter combinations by an oil reservoir engineer to carry out field tests. Although the method is relatively in line with practical production, the method also has the disadvantage of consuming excessive manpower and material resources, and the better injection-production scheme is still basically maintained within the range defined by the original test.
Disclosure of Invention
The invention aims to solve the defects, and provides an intelligent optimization method for injection and production parameters of a multi-well system water-drive reservoir, which is characterized by firstly constructing a proxy model for reservoir water-drive development and production based on machine learning algorithms such as multivariate adaptive spline regression and the like, then optimizing injection and production parameters by using particle swarm optimization and the like to realize that the maximum Net Present Value (NPV) of reservoir production is the optimization target, and obtaining an optimal injection and production parameter combination scheme.
The invention specifically adopts the following technical scheme:
an intelligent optimization method for injection and production parameters of a water drive reservoir of a multi-well system comprises the following steps:
step 1, analyzing and determining injection and production parameters to be optimized of a multi-well system water drive reservoir, and generating n groups of development schemes with different injection and production parameter combinations by using a Latin hypercube sampling method based on upper and lower limits of values of the parameters to be optimized;
step 2, constructing a multi-well system injection-production development oil reservoir geological model including porosity, injection-production well position and permeability, respectively carrying out numerical simulation on n groups of injection-production parameter schemes based on the model, collecting corresponding oil reservoir water-drive development net current values, and constructing a water-drive oil field development and production big data set;
step 3, substituting n groups of different injection-production parameter schemes and corresponding oil reservoir net current values into a multivariate self-adaptive spline regression model machine learning method to train a proxy model for water drive oil reservoir development capable of replacing oil reservoir numerical simulation;
and 4, constructing an optimization model of injection and production parameters of the water-drive reservoir to realize the maximum optimization target of the net current value of reservoir production, optimizing the injection and production parameters by adopting a particle swarm optimization, solving an optimal injection and production parameter combination scheme, and realizing the optimization adjustment of the injection and production parameters of the water-drive reservoir.
Preferably, in the step 2, in the process of constructing the large data set for development and production of the water-drive oil field, the net current value of the water-drive development of the oil reservoir is calculated by adopting the formula (1):
Figure BDA0002917126790000021
wherein NPV represents the net current value of reservoir water drive development; n represents the production life of the oil reservoir; np represents the total number of producing wells; n is a radical ofjRepresenting the total number of water injection wells; po tRepresenting the price of the crude oil in the t calculation period; pw tRepresents the price of treated water in the t-th calculation period; pw2 tRepresenting the price of the injected water in the t calculation period; qo，i tRepresents the cumulative oil production of the ith production well in the t calculation period, Qw，i tRepresenting the accumulated water production of the ith production well in the t calculation period; qw2，j tRepresenting the accumulated water injection quantity of the jth water injection well in the tth calculation period; e represents the discount rate of the enterprise.
Preferably, in step 3, the machine learning method of the multivariate self-adaptive spline regression model includes multivariate self-adaptive regression splines, a least square support vector machine and a back propagation neural network, and the regression fitting and prediction are sequentially performed on the data samples, and the root mean square error, the average absolute error and the coefficient index are introduced for evaluation.
Preferably, in step 4, the constructed optimization model is shown as formulas (2) and (3),
max NPV＝F(un)
s.t Umin≤un≤Umax (2)
Figure BDA0002917126790000022
wherein u isnRepresenting the oil reservoir injection-production parameters to be optimized in the nth regulation step;
Figure BDA0002917126790000023
representing the bottom hole pressure of the production well,
Figure BDA0002917126790000024
Representing the liquid extraction speed of the production well,
Figure BDA0002917126790000025
Representing the water injection pressure.
The invention has the following beneficial effects:
compared with the traditional method for adjusting and optimizing the injection and production parameters of the water-drive reservoir, the intelligent optimization method provided by the method is novel, simple and feasible, does not need to consume too much manpower and material resources, has an optimized result unrelated to experience, and can more easily achieve the optimal development effect of the reservoir.
Drawings
FIG. 1 is a structural diagram of a multi-well system water drive reservoir injection-production parameter intelligent optimization method;
FIG. 2 is a five-point method production well pattern layout;
FIG. 3 is a graph comparing MARS model expectation values with regression results;
FIG. 4 is a comparison graph of expected values of LS-SVM model and regression results;
FIG. 5 is a graph comparing expected values of the BP model with regression results;
FIG. 6 is a graph comparing predicted results and expected values of three regression models;
FIG. 7 is a graph of prediction error under three regression models;
FIG. 8 is a diagram of an iterative process of GA optimization;
FIG. 9 is a diagram of an SA optimization iterative process;
FIG. 10 is a diagram of an iterative process of PSO optimization.
Detailed Description
The following description of the embodiments of the present invention will be made with reference to the accompanying drawings:
with reference to fig. 1, the intelligent optimization method for injection and production parameters of the water drive reservoir of the multi-well system comprises the following steps:
step 1, analyzing and determining injection and production parameters to be optimized of a multi-well system water drive oil reservoir, and generating n groups of development schemes with different injection and production parameter combinations by using a Latin hypercube sampling method based on upper and lower limits of values of the parameters to be optimized.
Whether the setting of the oil reservoir injection and production parameters is reasonable or not is the key for influencing the final water drive development effect. Generally, all designed parameters of a production well and a water injection well in the water drive process can be used as injection and production parameters to be optimized of a water drive oil reservoir. According to the scheme, three parameters of the bottom pressure of the production well, the bottom pressure of the water injection well and the liquid extraction speed are selected as objects to be optimized, and the upper limit and the lower limit of the parameter value are determined. Based on the value range of the parameter to be optimized, the embodiment of the invention adopts a Latin Hypercube Sampling (LHS) method to obtain n groups of random injection-production parameter schemes. The LHS algorithm completely avoids sample collapse while retaining the advantages of the monte carlo simulation method, and can substantially improve sampling efficiency.
Step 2, constructing a multi-well system injection-production development oil reservoir geological model including porosity, injection-production well position and permeability, respectively carrying out numerical simulation on n groups of injection-production parameter schemes based on the model, collecting corresponding oil reservoir water-drive development net current values, and constructing a water-drive oil field development and production big data set;
in the process of constructing a large data set for developing and producing a water-flooding oil field, in order to simplify the problem, the embodiment of the invention only considers the income of crude oil sales of the oil field, the cost for processing water produced by a production well and the cost for pumping water into an injection well, and the net current value of the water-flooding development of the oil deposit is calculated by adopting an equation (1):
Figure BDA0002917126790000031
wherein NPV represents the net current value of reservoir water drive development; n represents the production life of the oil reservoir; np represents the total number of producing wells; n is a radical ofjRepresenting the total number of water injection wells; po tRepresenting the price of the crude oil in the t calculation period; pw tRepresents the price of treated water in the t-th calculation period; pw2 tRepresenting the price of the injected water in the t calculation period; qo，i tRepresents the cumulative oil production of the ith production well in the t calculation period, Qw，i tRepresenting the accumulated water production of the ith production well in the t calculation period; qw2，j tRepresenting the accumulated water injection quantity of the jth water injection well in the tth calculation period; e represents the discount rate of the enterprise.
Step 3, substituting n groups of different injection-production parameter schemes and corresponding oil reservoir net current values into a multivariate self-adaptive spline regression model machine learning method to train a proxy model for water drive oil reservoir development capable of replacing oil reservoir numerical simulation; the adopted machine learning method of the multivariate self-adaptive spline regression model comprises multivariate self-adaptive regression splines (MARS), a least square support vector machine (LS-SVM) and a back propagation neural network (BP), wherein regression fitting and prediction are sequentially carried out on data samples, and four indexes such as Root Mean Square Error (RMSE) and determination coefficient index (R-square) are introduced for evaluation.
And 4, constructing an optimization model of injection and production parameters of the water-drive reservoir to realize the maximum optimization target of the net current value of reservoir production, optimizing the injection and production parameters by adopting a particle swarm optimization, solving an optimal injection and production parameter combination scheme, and realizing the optimization adjustment of the injection and production parameters of the water-drive reservoir. The constructed optimization model is shown as formulas (2) and (3),
max NPV＝F(un)
s.t Umin≤un≤Umax (2)
Figure BDA0002917126790000041
wherein u isnRepresenting the oil reservoir injection-production parameters to be optimized in the nth regulation step;
Figure BDA0002917126790000043
representing the bottom hole pressure of the production well,
Figure BDA0002917126790000044
Representing the liquid extraction speed of the production well,
Figure BDA0002917126790000045
Representing the water injection pressure.
Aiming at the optimization model of the water-drive reservoir injection and production parameters, three global intelligent optimization algorithms of a Genetic Algorithm (GA), a Particle Swarm Optimization (PSO) and Simulated Annealing (SA) are adopted to respectively solve.
Example one
(1) The number of the grids of the established reservoir model is 21 multiplied by 5, and the grid step length in the X, Y, Z direction is 10 m. The reservoir was developed using a five-point well pattern, as shown in fig. 2. The mean values for the permeability distribution of reservoir X, Y, Z in the direction were 20mD, 30mD, and 1mD, respectively, and the mean value for the porosity distribution was 20%. Reservoir permeability and porosity are generated by a stochastic modeling method. The basic reservoir formation parameters are shown in table 1.
TABLE 1
Figure BDA0002917126790000042
(2) Training sample generation
In order to obtain the optimal injection and production parameter combination of the oil reservoir, three parameters of bottom pressure of a production well, bottom pressure of a water injection well and liquid extraction speed are selected as research objects, and the upper limit and the lower limit of parameter values are determined. Adopt Latin Hypercube Sampling (LHS) method to obtain 100 groups of random simulation data samples. The LHS algorithm can completely avoid sampling collapse while retaining the advantages of the monte carlo simulation method, and can improve sampling efficiency to a large extent essentially, with upper and lower limits for three parameters as in table 2. The net present value is then calculated using equation (1).
TABLE 2
Figure BDA0002917126790000051
(3) Regression model research based on machine learning
In order to construct the mathematical correspondence between the oil reservoir development injection-production parameters and the net current production value and realize the conversion from the engineering problem to the mathematical problem, a supervised machine learning algorithm is adopted. In supervised learning, it is first necessary to input a set of labels or target variables to the algorithm, which correspond to a given set of predicted variables, and then train the model to generate a function that can map the inputs to the target variables. The invention selects three methods of a multivariate self-adaptive regression spline, a least square support vector machine and a back propagation neural network for machine learning. Firstly, the real training values (expected values) of 100 groups of injection-production parameter schemes in the training set are compared with the regression results under different regression models, and the comparison results are shown in fig. 3, fig. 4 and fig. 5.
The invention adopts Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and determination coefficient (R-square) as the mathematical index for evaluating the regression model. When the Root Mean Square Error (RMSE) and the Mean Absolute Error (MAE) are taken as evaluation indexes, the closer to 0, the smaller the error between the fitting data and the real data is, and the better the fitting effect is. When the coefficient (R-square) is determined as the evaluation index, the closer to 1, the stronger the explanatory ability of the variables of the regression equation to y is shown, and the better the model fits to the data. The results are shown in Table 3, and the evaluation formula is as follows:
Figure BDA0002917126790000052
Figure BDA0002917126790000053
Figure BDA0002917126790000054
wherein, yiRepresenting the true value;
Figure BDA0002917126790000055
mean values representing the true values;
Figure BDA0002917126790000056
indicating the predicted value.
TABLE 3
Figure BDA0002917126790000061
In the case, from the root mean square error data obtained by calculation, the MARS model is reduced compared with the LS-SVM model and the BP model; from the calculated average absolute error data, the MARS model is reduced by 35.21 compared with the LS-SVM model and 57.62 compared with the BP model; from the calculated determination coefficient data, the determination coefficient of the MARS model is improved from 82% of the BP model and 92% of the LS-SVM model to 99%. Therefore, when the engineering problem of reservoir water flooding development is converted into a mathematical problem, the fitting precision of a Multivariate Adaptive Regression Spline (MARS) model is the highest, and the fitting effect of a least square support vector machine (LS-SVM) model is inferior to that of the Multivariate Adaptive Regression Spline (MARS) model and slightly superior to that of a back propagation neural network (BP) model.
In order to further verify the accuracy of the Multivariate Adaptive Regression Spline (MARS) model, 10 sets of injection-production parameter combination schemes are generated by using a latin hypercube sampling (LSH) method, a comparison graph (fig. 6) and an error graph (fig. 7) of a software simulation result and a model prediction result are drawn by respectively using a numerical simulation method and regression model prediction, and relative error indexes are introduced to evaluate the prediction effects of different machine learning models. Equation 7, where RE represents the relative error, ytThe actual value is represented by the value of,
Figure BDA0002917126790000064
and representing the model predicted value. The evaluation of the predicted effect is compared in table 4.
Figure BDA0002917126790000062
TABLE 4
Figure BDA0002917126790000063
It can be seen that the average relative errors predicted by using Multivariate Adaptive Regression Splines (MARS), least squares support vector machines (LS-SVM) and back propagation neural networks (BP) are 0.16%, 0.66% and 0.49%, respectively. The prediction value and the expected value of the MARS model are the closest, the error is the smallest, and the prediction effect is the best in the three methods.
(4) Optimization model research based on multi-well system
The average net present value predicted by 100 groups of injection and production parameter MARS regression models is used as a reference point, three global intelligent optimization algorithms including a Genetic Algorithm (GA), a Particle Swarm Optimization (PSO) and Simulated Annealing (SA) are respectively used for solving, and table 5 shows the optimization results of different optimization models, so that the three optimization algorithms can achieve convergence to achieve the purpose of optimizing, and the final Net Present Value (NPV) is improved by about 7%. The accuracy of the optimal solution obtained by PSO (particle swarm optimization) is slightly better than the results obtained by SA (simulated annealing) algorithm and GA (genetic algorithm), and the optimization result of GA (genetic algorithm) is the worst.
TABLE 5
Figure BDA0002917126790000071
Fig. 8, 9 and 10 show the iterative process of different optimization algorithms respectively. It can be seen that the Genetic Algorithm (GA) almost reaches convergence after 10 iterations. The particle swarm optimization algorithm (PSO) can achieve convergence after 40 iterations. The simulated annealing algorithm (SA) can reach the optimal state after being iterated for about 50 times, and the local optimization at the later stage is started.
Table 6 shows the evaluation results of the different optimization models. The invention is based on the principle that the convergence speed is considered again on the premise of meeting the convergence precision, and the following conclusion is obtained: compared with the optimization process of a Genetic Algorithm (GA) and a simulated annealing algorithm (SA), the Particle Swarm Optimization (PSO) has excellent convergence precision performance on the optimization problem of the injection and production parameters of the water-drive oil reservoir, has good performance in the convergence speed aspect, and can better meet the requirement of intelligent optimization of the injection and production parameters in the development process of the water-drive oil field.
TABLE 6
Figure BDA0002917126790000072
It is to be understood that the above description is not intended to limit the present invention, and the present invention is not limited to the above examples, and those skilled in the art may make modifications, alterations, additions or substitutions within the spirit and scope of the present invention.

Claims (4)

1. An intelligent optimization method for injection and production parameters of a multi-well system water-drive reservoir is characterized by comprising the following steps:
step 1, analyzing and determining injection and production parameters to be optimized of a multi-well system water drive reservoir, and generating n groups of development schemes with different injection and production parameter combinations by using a Latin hypercube sampling method based on upper and lower limits of values of the parameters to be optimized;
step 2, constructing a multi-well system injection-production development oil reservoir geological model including porosity, injection-production well position and permeability, respectively carrying out numerical simulation on n groups of injection-production parameter schemes based on the model, collecting corresponding oil reservoir water-drive development net current values, and constructing a water-drive oil field development and production big data set;
step 3, substituting n groups of different injection-production parameter schemes and corresponding oil reservoir net current values into a multivariate self-adaptive spline regression model machine learning method to train a proxy model for water drive oil reservoir development capable of replacing oil reservoir numerical simulation;
and 4, constructing an optimization model of injection and production parameters of the water-drive reservoir to realize the maximum optimization target of the net current value of reservoir production, optimizing the injection and production parameters by adopting a particle swarm optimization, solving an optimal injection and production parameter combination scheme, and realizing the optimization adjustment of the injection and production parameters of the water-drive reservoir.
2. The intelligent optimization method for the injection-production parameters of the water-drive reservoir of the multi-well system according to claim 1, wherein in the step 2, in the process of constructing the large data set for development and production of the water-drive oil field, the net current value of the water-drive development of the reservoir is calculated by adopting the formula (1):
Figure FDA0002917126780000011
wherein NPV represents the net current value of reservoir water drive development; n represents the production life of the oil reservoir; np represents the total number of producing wells; n is a radical ofjRepresenting the total number of water injection wells; po tRepresenting the price of the crude oil in the t calculation period; pw tRepresents the price of treated water in the t-th calculation period; pw2 tRepresenting the price of the injected water in the t calculation period; qo，i tRepresenting the accumulated oil production of the ith production well in the t calculation period; qw，i tRepresenting the accumulated water production of the ith production well in the t calculation period; qw2，j tRepresenting the accumulated water injection quantity of the jth water injection well in the tth calculation period; e represents the discount rate of the enterprise.
3. The method for intelligently optimizing the injection-production parameters of the water-drive reservoir of the multi-well system according to claim 1, wherein in the step 3, the adopted multivariate self-adaptive spline regression model machine learning method comprises multivariate self-adaptive regression splines, a least square support vector machine and a back propagation neural network, the regression fitting and the prediction are sequentially carried out on the data samples, and the root mean square error, the average absolute error and the determined coefficient index are introduced for evaluation.
4. The intelligent optimization method for the injection-production parameters of the water-drive reservoir of the multi-well system according to claim 1, wherein in the step 4, the constructed optimization model is shown as the formulas (2) and (3),
max NPV＝F(un)
s.t Umin≤un≤Umax (2)
Figure FDA0002917126780000012
wherein u isnRepresenting the oil reservoir injection-production parameters to be optimized in the nth regulation step;
Figure FDA0002917126780000021
representing the bottom hole pressure of the production well,
Figure FDA0002917126780000022
Representing the liquid extraction speed of the production well,
Figure FDA0002917126780000023
Representing the water injection pressure.
CN202110105180.XA 2021-01-26 2021-01-26 Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system Pending CN113032953A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110105180.XA CN113032953A (en) 2021-01-26 2021-01-26 Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110105180.XA CN113032953A (en) 2021-01-26 2021-01-26 Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system

Publications (1)

Publication Number Publication Date
CN113032953A true CN113032953A (en) 2021-06-25

Family

ID=76460375

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110105180.XA Pending CN113032953A (en) 2021-01-26 2021-01-26 Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system

Country Status (1)

Country Link
CN (1) CN113032953A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115577562A (en) * 2022-11-09 2023-01-06 中国石油大学(华东) Fractured reservoir well position optimization method
CN116205164A (en) * 2023-04-27 2023-06-02 中国石油大学(华东) Multi-agent injection and production optimization method based on self-adaptive basis function selection

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109872007A (en) * 2019-03-12 2019-06-11 中国地质大学（北京） Oil reservoir injection based on support vector machines agent model adopts parameter Multipurpose Optimal Method
CN111625922A (en) * 2020-04-15 2020-09-04 中国石油大学(华东) Large-scale oil reservoir injection-production optimization method based on machine learning agent model

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109872007A (en) * 2019-03-12 2019-06-11 中国地质大学（北京） Oil reservoir injection based on support vector machines agent model adopts parameter Multipurpose Optimal Method
CN111625922A (en) * 2020-04-15 2020-09-04 中国石油大学(华东) Large-scale oil reservoir injection-production optimization method based on machine learning agent model

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
卢昊: "《基于矩方法的相关失效模式机械结构系统可靠性稳健设计》", 30 September 2014 *
夏位荣等: "《油气田开发地质学》", 31 March 1999 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115577562A (en) * 2022-11-09 2023-01-06 中国石油大学(华东) Fractured reservoir well position optimization method
CN116205164A (en) * 2023-04-27 2023-06-02 中国石油大学(华东) Multi-agent injection and production optimization method based on self-adaptive basis function selection

Similar Documents

Publication Publication Date Title
CN111861774B (en) Oil reservoir production machine learning method based on parallel agent model
Chen et al. Global and local surrogate-model-assisted differential evolution for waterflooding production optimization
CN111625922B (en) Large-scale oil reservoir injection-production optimization method based on machine learning agent model
Guyaguler et al. Optimization of well placement in a Gulf of Mexico waterflooding project
Zhao et al. A surrogate-assisted multi-objective evolutionary algorithm with dimension-reduction for production optimization
CN113032953A (en) Intelligent optimization method for injection and production parameters of water-drive oil reservoir of multi-well system
CN112282714B (en) Full well network water injection development optimization method based on deep learning and graph theory
CN114595608B (en) Fracturing construction parameter and working system parameter optimization method and system
CN109583588A (en) A kind of short-term wind speed forecasting method and system
Han et al. Well placement and control optimization of horizontal steamflooding wells using derivative-free algorithms
Ogbeiwi et al. Robust optimisation of water flooding using an experimental design-based surrogate model: A case study of a Niger-Delta oil reservoir
Xiao et al. Surrogate-assisted hydraulic fracture optimization workflow with applications for shale gas reservoir development: A comparative study of machine learning models
CN112800673A (en) Method for predicting blasting block degree based on SA-GA-BP algorithm
CN116579095A (en) CO based on multi-target interaction 2 Reinjection strategy optimization evaluation method
CN108133286B (en) Underground water multi-target calculation method based on ground settlement substitution model
Shirangi et al. Retrospective optimization of well controls under uncertainty using kernel clustering
CN112796747B (en) Oil and gas drilling strategy prediction method and device based on multiple targets
CN109117517A (en) A kind of base pit dewatering optimum design method based on genetic algorithm and groundwater model
Li et al. Ensembles‐based and GA‐based optimization for landfill gas production
Al-Aghbari et al. Stochastic multi-objective optimization approaches in a real-world oil field waterflood management
Wang et al. An intelligent multi-fidelity surrogate-assisted multi-objective reservoir production optimization method based on transfer stacking
CN111222271B (en) Matrix-fracture-unsteady state fluid channeling oil reservoir fracture numerical simulation method and system
Madadgar et al. Forced Water main design; mixed ant colony optimization
Chen et al. Development and application of a coupling method for well pattern and production optimization in unconventional reservoirs
CN118036477A (en) Well position and well control parameter optimization method based on space-time diagram neural network

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