A method, system, equipment and medium for optimizing injection and production parameters of water drive oil reservoir

Abstract

The invention discloses a method, system, equipment and medium for optimizing injection and production parameters of a water-drive oil reservoir, and specifically relates to the field of oil reservoir engineering technology. The technical key points are as follows: inputting basic sample points into at least one high-precision simulation model for calculation to obtain the net present value corresponding to the basic sample points, and using the basic sample points and the corresponding net present value to construct a high-fidelity training sample data set; inputting non-basic sample points into at least one low-precision simulation model for calculation to obtain the net present value corresponding to the non-basic sample points, and using the non-basic sample points and the corresponding net present value to construct a low-fidelity training sample data set; based on the high-fidelity training sample data set and the low-fidelity training sample data set, a variable credibility adaptive proxy model for the optimization problem of injection and production parameters of oil and water wells is constructed; using a differential evolution hybrid particle swarm algorithm in combination with the variable credibility adaptive proxy model to optimize design variables, and obtaining a maximum net present value and a corresponding injection and production plan.

Description

Technical Field

[0001] The invention relates to the technical field of oil reservoir engineering, and in particular to a method, system, equipment and medium for optimizing injection and production parameters of a water drive oil reservoir. Background Art

[0002] Due to the complex fracture network in fractured reservoirs, the injected water easily forms a rapid breakthrough channel along the fractures during water flooding, which fails to effectively displace the oil reservoir in the matrix, thereby reducing the recovery rate. By optimizing the injection and production parameters, the water flooding effect can be improved, the water breakthrough time can be delayed, and the water flooding swept volume can be increased, thereby improving the reservoir recovery rate. In addition, reasonable optimization of injection and production parameters can also reduce production costs, extend the economic life of the oil field, and adapt to the complex characteristics of the reservoir, making the development strategy more accurate and efficient. Therefore, the optimization of injection and production parameters is of key significance in the economic and efficient development of fractured reservoirs.

[0003] However, in the existing water flooding process of fractured oil reservoirs, the complex fracture network and reservoir heterogeneity will lead to the problem of low recovery rate. At the same time, the heterogeneity of the reservoir and the complexity of the fracture network in fractured oil reservoirs make it difficult for traditional water flooding optimization design to fully capture the real seepage behavior of the reservoir.

[0004] Therefore, the present invention aims to provide a method, system, equipment and medium for optimizing injection and production parameters of a water drive oil reservoir to solve the above-mentioned related problems. Summary of the invention

[0005] The technical problem to be solved by the present invention is that in the existing water flooding process of fractured oil reservoirs, due to the complex fracture network and reservoir heterogeneity, the related problem of low recovery rate is brought about. The purpose is to provide a method, system, equipment and medium for optimizing injection and production parameters of water-driven oil reservoirs. The kernel function idea of ​​the support vector regression model is adopted to realize multi-precision data fusion, and low-precision and high-precision samples are mapped into high-dimensional space, so as to construct a variable credibility adaptive proxy model suitable for complex fractured oil reservoirs; at the same time, when sampling multi-precision data, the base point sample points and non-base point sample points are sampled and processed, and the variable credibility adaptive proxy model is iteratively optimized; at the same time, the method combines the differential evolution hybrid particle swarm algorithm with the variable credibility adaptive proxy model to perform global optimization design on the injection and production parameters of the fractured oil reservoir; thereby, the advantages of multi-source data are fully utilized, and the calculation cost and time are significantly reduced while ensuring the accuracy, and the efficiency of the optimization process is improved; at the same time, by optimizing the injection and production parameters, the water flooding effect is also improved, the water breakthrough time is delayed, and the water flooding swept volume is increased, so as to significantly improve the oil reservoir recovery rate.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for optimizing injection and production parameters of a water drive oil reservoir, characterized in that the method comprises:

[0008] Determine the initial conditions of the oil and water well injection and production parameter optimization problem, where the initial conditions include design variables, design space, objective function and constraints;

[0009] Using a commercial simulator to establish fractured oil reservoir water flooding simulation models of different precisions, wherein the fractured oil reservoir water flooding simulation models of different precisions include at least one high-precision simulation model and at least one low-precision simulation model;

[0010] The design space is evenly divided according to the dimension of the design variables to obtain multiple small spaces, and the coordinates of the center points of each small space are obtained as basic sample points, and non-basic sample points are obtained by random sampling in the design space;

[0011] Inputting basic sample points into at least one high-precision simulation model for calculation to obtain the net present value corresponding to the basic sample points, and using the basic sample points and the corresponding net present value to construct a high-fidelity training sample data set; inputting non-basic sample points into at least one low-precision simulation model for calculation to obtain the net present value corresponding to the non-basic sample points, and using the non-basic sample points and the corresponding net present value to construct a low-fidelity training sample data set;

[0012] Based on high-fidelity training sample data sets and low-fidelity training sample data sets, a variable credibility adaptive proxy model for oil and water well injection and production parameter optimization problem is constructed;

[0013] The differential evolution hybrid particle swarm algorithm is combined with the variable credibility adaptive agent model to optimize the design variables and obtain the maximum net present value and the corresponding injection-production scheme.

[0014] Furthermore, the initial conditions of the oil and water well injection and production parameter optimization problem are determined, wherein the initial conditions include design variables, design space, objective function and constraints, specifically:

[0015] The design variables are the injection and production parameters of oil and water wells, and the objective function is constructed with the maximum net present value of reservoir production as the goal, and the injection and production constraints are established. The specific objective function is: , in, is the net present value, which is the design variable oil and water well injection and production parameters Function of is the total time step for reservoir simulation; For the The length of a time step; is the discount rate; For the The cumulative time at the end of each time step; and denote the number of oil wells and water wells in the model, respectively; is the crude oil sales price; Cost of produced water treatment; injection costs for water injection wells; , and Respectively Oil producing wells in time steps Average oil production, water production and water injection wells The specific injection and production constraints are: ,in, , They represent the minimum and maximum oil-water well injection and production parameters, respectively.

[0016] Furthermore, a commercial simulator was used to establish a water flooding simulation model of fractured reservoirs with different precisions, specifically:

[0017] According to the actual physical parameters of the oil and water wells and the actual injection and production parameters, a three-dimensional reservoir model is constructed, and according to the model parameters of the three-dimensional reservoir model, a fractured reservoir water drive simulation model with different accuracies is established using one of the eclipse commercial simulators or the CMG commercial simulator, wherein the model parameters include the number of grids, the grid size and the reservoir physical properties.

[0018] Furthermore, the design space is evenly divided according to the dimension of the design variables to obtain multiple small spaces, specifically:

[0019] Determine the dimension of the hypercube and the corresponding number of partitions on each dimension of the hypercube according to the dimension of the design variables;

[0020] The hypercube space is evenly divided based on the number of partitions in each dimension of the hypercube to obtain multiple small spaces.

[0021] Furthermore, based on the high-fidelity training sample data set and the low-fidelity training sample data set, a variable credibility adaptive proxy model for the oil and water well injection and production parameter optimization problem is constructed, specifically:

[0022] The high-fidelity training sample data set and the low-fidelity training sample data set are combined to construct a variable credibility proxy model;

[0023] Using the kernel function idea of ​​the support vector regression model, based on the variable credibility proxy model kernel function, the high-fidelity training sample data set and the low-fidelity training sample data set are mapped to a high-dimensional space, and the kernel function matrix of the fusion of the high-fidelity training sample data set and the low-fidelity training sample data set is obtained. Based on the kernel function matrix, a variable credibility adaptive proxy model for the optimization problem of injection and production parameters of oil and water wells is constructed;

[0024] The particle swarm algorithm is used to adaptively search for the optimal hyperparameter combination in the variable credibility adaptive proxy model. The root mean square error between the net present value output by the variable credibility adaptive proxy model and the net present value in the high-fidelity training sample dataset is used as the evaluation index. The optimal variable credibility adaptive proxy model is obtained when the evaluation index is minimized.

[0025] Furthermore, the differential evolution hybrid particle swarm algorithm is used in combination with the variable credibility adaptive agent model to optimize the design variables and obtain the maximum net present value and the corresponding injection-production scheme, specifically:

[0026] Based on the variable credibility adaptive agent model, the differential evolution hybrid particle swarm algorithm is used to solve the optimization problem to optimize the design variables and obtain the maximum net present value and the corresponding injection and production scheme. The specific optimization problem is: ;

[0027] in, The variable credibility adaptive proxy model for net present value; is the injection and production parameters of oil and water wells, , They represent the minimum and maximum oil-water well injection and production parameters, respectively. It means finding the optimal injection and production parameters for oil and water wells.

[0028] The present invention also provides a water drive oil reservoir injection and production parameter optimization system, characterized in that the system is used in any one of the above-mentioned water drive oil reservoir injection and production parameter optimization methods, and the system comprises:

[0029] The first module is used to determine the initial conditions of the oil and water well injection and production parameter optimization problem, wherein the initial conditions include design variables, design space, objective function and constraint conditions;

[0030] The second module is used to establish fractured oil reservoir water flooding simulation models with different precisions using a commercial simulator, wherein the fractured oil reservoir water flooding simulation models with different precisions include at least one high-precision simulation model and at least one low-precision simulation model;

[0031] The third module is used to evenly divide the design space according to the dimension of the design variables to obtain multiple small spaces, obtain the coordinates of the center point of each small space as the basic sample point, and randomly sample in the design space to obtain non-basic sample points;

[0032] The fourth module is used to input the basic sample points into at least one high-precision simulation model for calculation, obtain the net present value corresponding to the basic sample points, and use the basic sample points and the corresponding net present value to construct a high-fidelity training sample data set; input the non-basic sample points into at least one low-precision simulation model for calculation, obtain the net present value corresponding to the non-basic sample points, and use the non-basic sample points and the corresponding net present value to construct a low-fidelity training sample data set;

[0033] The fifth module is used to construct a variable credibility adaptive proxy model for the oil and water well injection and production parameter optimization problem based on a high-fidelity training sample data set and a low-fidelity training sample data set;

[0034] The sixth module is used to optimize the design variables using the differential evolution hybrid particle swarm algorithm combined with the variable credibility adaptive agent model to obtain the maximum net present value and the corresponding injection and production plan.

[0035] The present invention also provides a computer device, comprising a system memory and a processor, wherein the system memory stores a computer program, and the processor implements the steps of any one of the above-mentioned methods when executing the computer program.

[0036] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any one of the methods described above are implemented.

[0037] The present invention also provides a computer program product comprising instructions, and when the instructions are executed by a computer device cluster, the computer device cluster executes any one of the methods described above.

[0038] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0039] In the present invention, multi-precision data fusion is realized by adopting the kernel function idea of ​​the support vector regression model, low-precision and high-precision samples are mapped into high-dimensional space, and a variable credibility adaptive proxy model suitable for complex fractured oil reservoirs is constructed; at the same time, when sampling multi-precision data, sampling and processing of base point sample points and non-base point sample points are performed, and the variable credibility adaptive proxy model is iteratively optimized; at the same time, the method combines the differential evolution hybrid particle swarm algorithm with the variable credibility adaptive proxy model to perform global optimization design on the injection and production parameters of the fractured oil reservoir; thereby making full use of the advantages of multi-source data, while ensuring accuracy, significantly reducing the calculation cost and time, and improving the efficiency of the optimization process; at the same time, by optimizing the injection and production parameters, the water drive effect is also improved, the water breakthrough time is delayed, and the water drive swept volume is increased, thereby significantly improving the oil reservoir recovery rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0041] Figure 1 This is a schematic diagram of a method flow chart of a method for optimizing injection and production parameters of a water drive oil reservoir in this embodiment;

[0042] Figure 2 This is a schematic diagram of a low-precision geological three-dimensional model of an oil reservoir in this embodiment;

[0043] Figure 3 This is a schematic diagram of a high-precision three-dimensional geological model of an oil reservoir in this embodiment;

[0044] Figure 4 is the NPV error index R of 20 runs of the comparative experiment in this embodiment. 2 Statistical diagram of the mean and standard deviation of ;

[0045] Figure 5 This is a schematic diagram of the distribution of injection-production schemes with the maximum NPV in this embodiment;

[0046] Figure 6 This is a module schematic diagram of a water drive reservoir injection and production parameter optimization system in this embodiment;

[0047] Figure 7 It is a structural schematic diagram of a computer device in this embodiment. DETAILED DESCRIPTION

[0048] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0049] In the present disclosure, unless otherwise specified, the use of the terms "first", "second", etc. to describe various elements is not intended to limit the positional relationship, timing relationship, or importance relationship of these elements, and such terms are only used to distinguish one element from another element. In some examples, the first element and the second element may refer to the same instance of the element, and in some cases, based on the description of the context, they may also refer to different instances.

[0050] The terms used in the description of various examples in this disclosure are only for the purpose of describing specific examples and are not intended to be limiting. Unless the context clearly indicates otherwise, if the number of elements is not specifically limited, the element can be one or more. In addition, the term "and / or" used in this disclosure covers any one of the listed items and all possible combinations.

[0051] Example 1

[0052] See also Figure 1 As shown, the present invention provides a method for optimizing injection and production parameters of a water drive oil reservoir, characterized in that the method comprises:

[0053] S1: Determine the initial conditions for the optimization problem of injection and production parameters of oil and water wells, where the initial conditions include design variables, design space, objective function and constraints.

[0054] Specifically, the injection and production parameters of oil and water wells are set as design variables, and the corresponding design space is determined. , for The bottom hole pressure or injection and production parameters set for the well; and denote the number of oil wells and water wells in the model, respectively. and are all positive integers;

[0055] The objective function is constructed with the goal of maximizing the net present value of reservoir production, and injection and production constraints are established;

[0056] In this embodiment, the objective function is specifically:

[0057] ;in, is the net present value, which is the design variable oil and water well injection and production parameters Function of is the total time step for reservoir simulation; For the The length of a time step, in units of ; is the discount rate, in ; For the The cumulative time at the end of the time step, in units of ; and denote the number of oil wells and water wells in the model, respectively; is the crude oil sales price, in ; is the production water treatment cost, in ; is the injection cost of the water injection well, in ; , and Respectively Oil producing wells in time steps Average oil production, water production and water injection wells The water injection volume, in units of ;

[0058] For the injection and production parameter constraints of oil and water wells, when producing at a certain bottom hole pressure, the liquid volume and pressure should be kept within a reasonable range. The injection and production constraints are: ,in, , They represent the minimum and maximum oil and water well injection and production parameters, respectively;

[0059] S2: Using a commercial simulator, establishing fractured oil reservoir water flooding simulation models with different accuracies, wherein the fractured oil reservoir water flooding simulation models with different accuracies include at least one high-precision simulation model and at least one low-precision simulation model.

[0060] Specifically, a three-dimensional reservoir model is constructed according to the actual physical parameters of the oil and water wells and the actual injection and production parameters, and a fractured reservoir water flooding simulation model with different precision is established using one of the eclipse commercial simulators or the CMG commercial simulator according to the model parameters of the three-dimensional reservoir model, wherein the model parameters include the number of grids, the grid size and the reservoir physical properties;

[0061] It should be noted that, in this embodiment, the number of high-precision simulation models and low-precision simulation models is determined according to actual conditions, and can be one, two, three or other numbers, which will not be described in detail here;

[0062] At the same time, the high-precision simulation model accurately measures and defines the fracture width, porosity, permeability and fluid mechanics characteristics to ensure the physical reality of each parameter in the model, and uses a high-precision numerical solver to capture complex fracture flow characteristics; the high-precision simulation model uses average values ​​or empirical formulas to determine fracture-related physical parameters to reduce complexity and computational costs, and uses fast solution methods to provide preliminary flow trends and production forecasts;

[0063] At the same time, it should be noted that using commercial simulators to establish water drive simulation models of fractured reservoirs with different accuracies is a standard technical means in the field of reservoir development. For detailed technical content, please refer to [Han Dakuang. Fundamentals of Numerical Simulation of Reservoirs [M]. Petroleum Industry Press, 1993], and no further elaboration will be given here.

[0064] S3: Divide the design space evenly according to the dimension of the design variables to obtain multiple small spaces, and obtain the coordinates of the center points of each small space as basic sample points. Suppose the set of all basic sample points is , and randomly sample non-basic sample points in the design space. Let the set of all non-basic sample points be .

[0065] Specifically, the hypercube dimension and the corresponding number of partitions in each dimension of the hypercube are determined according to the dimension of the design variable; the hypercube space is evenly divided based on the number of partitions in each dimension of the hypercube to obtain multiple small spaces;

[0066] Specifically, in this embodiment, the Latin cube sampling method is used to randomly sample in the design space to obtain non-basic sample points. In other embodiments, other sampling methods may also be used, which will not be described in detail here.

[0067] S4: inputting the basic sample points into at least one high-precision simulation model for calculation, obtaining the net present value corresponding to the basic sample points, and constructing a high-fidelity training sample data set using the basic sample points and the corresponding net present value; inputting the non-basic sample points into at least one low-precision simulation model for calculation, obtaining the net present value corresponding to the non-basic sample points, and constructing a low-fidelity training sample data set using the non-basic sample points and the corresponding net present value;

[0068] S5: Based on high-fidelity training sample data sets and low-fidelity training sample data sets, a variable credibility adaptive proxy model for oil and water well injection and production parameter optimization problem is constructed.

[0069] Specifically: The high-fidelity training sample dataset and the low-fidelity training sample dataset are fused to construct a variable credibility proxy (MF) model ,in, is the design variable vector; It is a variable credibility agent model; is the vector used to fit the difference between low-fidelity training sample data and high-fidelity training sample data; is the function response of the low-precision simulation model; is the function response of the high-precision simulation model;

[0070] Using the kernel function idea of ​​the support vector regression model (SVR), based on the variable credibility proxy model kernel function, the high-fidelity training sample data set and the low-fidelity training sample data set are mapped to a high-dimensional space, and the kernel function matrix of the fusion of the high-fidelity training sample data set and the low-fidelity training sample data set is obtained. Based on the kernel function matrix, a variable credibility adaptive proxy model (MFSVR) for the optimization problem of injection and production parameters of oil and water wells is constructed. The kernel function matrix is ​​specifically: ;

[0071] in, is the kernel function matrix of the variable credibility adaptive agent model; It is a high-fidelity training sample dataset; It is a low-fidelity training sample dataset; , , , are the elements in the kernel function matrix, where Represents the kernel function matrix between low-fidelity training sample data, Represents the kernel function matrix between the low-fidelity training sample data and the high-fidelity training sample data, Represents the kernel function matrix between the high-fidelity training sample data and the low-fidelity training sample data from the perspective of high-fidelity training sample data, Represents the kernel function matrix between high-fidelity training sample data; is the number of low-fidelity training sample data, is a positive integer; is the number of high-fidelity training sample data, is a positive integer; in this embodiment, a radial basis kernel function is used, and the mathematical forms of the four kernel functions are defined as follows: ; ;

[0072] ; ;in, Represents the first and Basic sample points; Represents the first and Basic sample points; represents the design variable, , is the number of design variables; and is the kernel function parameter; It is a high-fidelity training sample dataset; It is a low-fidelity training sample dataset;

[0073] The particle swarm algorithm (PSO algorithm) is used to adaptively search for the optimal hyperparameter combination in the variable credibility adaptive proxy model. The root mean square error between the net present value output by the variable credibility adaptive proxy model and the net present value in the high-fidelity training sample data set is used as the evaluation index. When the evaluation index is minimized, the optimal variable credibility adaptive proxy model is obtained.

[0074] Specifically, in this embodiment, the particle swarm algorithm is used to adaptively search for the best combination of hyperparameters in the variable credibility adaptive proxy model. During the training process of the variable credibility adaptive proxy model, the particle swarm algorithm is used to adaptively search for the optimal hyperparameters to improve the prediction performance of the variable credibility adaptive proxy model.

[0075] At the same time, for the goal of model hyperparameter optimization, the root mean square error (RMSE) between the objective function response value (net present value) in the high-fidelity training sample data set calculated by the variable credibility adaptive proxy model and the high-precision simulation model is used. RMSE ) as an indicator, adaptively search for the model hyperparameter combination that minimizes its value, and apply it in the variable credibility adaptive proxy model RMSE The formula is:

[0076] ,in, and It is the first one evaluated by high-precision simulation model and variable credibility proxy model The target value of high-fidelity training sample data; represents the number of high-fidelity training sample data, is a positive integer;

[0077] The optimal variable credibility adaptive proxy model obtained is as follows: ,in, for The objective function response value (net present value); is the bias term; All are Lagrange multipliers; Represented as the number of low-fidelity training sample data, is a positive integer; Represented as elements in the sample data vector.

[0078] S6: The differential evolution hybrid particle swarm algorithm (DE-PSO) is used in combination with the variable credibility adaptive agent model to optimize the design variables and obtain the maximum net present value and the corresponding injection-production scheme.

[0079] Specifically, based on the variable credibility adaptive agent model, the differential evolution hybrid particle swarm algorithm is used to solve the optimization problem to optimize the design variables and obtain the maximum net present value and the corresponding injection and production scheme. The optimization problem is specifically: ;

[0080] in, The variable credibility adaptive proxy model for net present value; is the injection and production parameters of oil and water wells, It means finding the optimal injection and production parameters for oil and water wells.

[0081] Specifically, in this embodiment, the steps of differential evolution hybrid particle swarm algorithm (DE-PSO) calculation are as follows:

[0082] (1) Initialization

[0083] If the injection and production parameters are continuous variables, they can be initialized directly using the following formula: .in, and are the lower and upper bounds of each dimension of the design variable, respectively; Indicates that in the interval Random numbers on .

[0084] (2) Mixed operation

[0085] ①DE mutation and crossover operation

[0086] Differential evolution (DE) achieves individual mutation through differential strategy. The common differential strategy is to randomly select two or more individuals in the population, calculate the vector difference, and then scale and synthesize the vector difference and further sum it with the individual to be mutated. Depending on the mutation operator, the differential evolution algorithm has a variety of mutation modes. DE / rand algorithms have strong global search capabilities, but slow convergence speed, and DE / best have strong local search capabilities and fast convergence speed, but are prone to fall into local optimality.

[0087] Randomly select half of the individuals in the population and use the DE / rand-to-best operator that combines DE / rand and DE / best to balance the global and local search capabilities of the algorithm: .in, Represents an individual The current position vector; It is called the scaling factor and is a certain constant; Represents the position vector of the best individual in the current population; represents the position vectors of two different individuals randomly selected from the population, and are two different individuals randomly selected. The purpose of the crossover operation is to randomly select individuals, because differential evolution is also a random algorithm. The crossover operator adopted by the differential evolution algorithm is: .in, Indicates that the individual The position of the step; Indicates that the individual The speed of the steps; Indicates that the individual g The position of the step; is the crossover probability; Indicates Iteration steps.

[0088] ②PSO velocity and position update

[0089] For another part of individuals in the population, the PSO speed and position update formula is used to generate new individuals. n In the dimensional space, there are Particles, particles exist Moment The position in dimension is ,particle exist Moment The speed in dimension is ;exist Moment, Particle In the The speed and position in each dimension are updated as follows: ;

[0090] ;

[0091] ;

[0092] in, represents the inertia weight; Represents particles At the moment At that time, Speed ​​in dimension; Represents particles At the moment At that time, Speed ​​in dimension; represents the acceleration constant; Random numbers, usually generated in the range [0,1]; Represents particles At the moment , No. The best historical position in terms of dimension; Represents particles At the moment At that time, Position in dimension; Represents particles At the moment At that time, Position in dimension; The global best position is The value of the dimension; Indicates the maximum speed limit.

[0093] (3) Select an operation

[0094] According to the greedy selection strategy in DE, that is, selecting the better individual as the new individual, for the maximization problem, the selection operation operator is: ,in, Represents particles In the The value after the iteration step screening; Represents particles In the The value of the iteration step; Represents particles In the The value of the iteration step.

[0095] (4) Iteration: Repeat the mixing operation and selection operation until the termination condition is reached.

[0096] Specifically, in this embodiment, an exemplary method is also provided, based on the actual geological three-dimensional model of a certain oil field work area (such as Figure 2 and Figure 3 As shown in the figure, the bottom hole pressure of the injection well is set to 25-30MPa, and the bottom hole pressure of the production well is set to 10-25MPa. The Latin hypercube sampling method is used to generate base point sample points and non-base point sample points. The low-precision simulation model numerical simulator is used to calculate the non-base point sample points (such as Figure 2 As shown in the figure, the basic sample points are calculated using a high-precision simulation model numerical simulator (as shown in the figure). Figure 3 The NPV of different injection-production schemes is calculated.

[0097] Based on the multi-source data samples with different precision obtained above, the MFSVR model is constructed according to the method in step S5, and the low-precision and high-precision data are transformed using the kernel function Mapping to high-dimensional space can achieve effective fusion of data with different precisions.

[0098] Run the MFSVR model with multiple sources and different precision data samples for another twenty times to verify the accuracy R 2 Indicator characterization: ;

[0099] In the above formula, For the The true value of the data; For the The predicted value of data; is the average value of the real data; is the number of samples. Count twenty times R 2 The mean and variance of the values, such as Figure 4 As shown, from Figure 4 It can be seen that when the number of high-precision base point sample points is 300, The best indicator reached 0.9832, which is the most suitable base point sample number for establishing the MFSVR model.

[0100] According to the established multi-precision data fusion variable credibility adaptive agent model, the DE-PSO hybrid algorithm in step S6 is used to optimize the injection and production parameters of the water drive reservoir, and then the scheme with the largest NPV is selected as the optimal scheme for optimizing the injection and production parameters of the water drive reservoir. The detailed injection and production parameters are as follows: Figure 5 shown.

[0101] Specifically, in this embodiment, multi-precision data fusion is realized by adopting the kernel function idea of ​​the support vector regression model, low-precision and high-precision samples are mapped to high-dimensional space, and a variable credibility adaptive proxy model suitable for complex fractured oil reservoirs is constructed; at the same time, when sampling multi-precision data, sampling and processing of base point sample points and non-base point sample points are performed, and the variable credibility adaptive proxy model is iteratively optimized; at the same time, this method combines the differential evolution hybrid particle swarm algorithm with the variable credibility adaptive proxy model to perform global optimization design of the injection and production parameters of the fractured oil reservoir; thereby making full use of the advantages of multi-source data, while ensuring accuracy, significantly reducing the calculation cost and time, and improving the efficiency of the optimization process; at the same time, by optimizing the injection and production parameters, the water drive effect is also improved, the water breakthrough time is delayed, and the water drive swept volume is increased, thereby significantly improving the oil reservoir recovery rate.

[0102] Example 2

[0103] See also Figure 6 As shown, the present invention also provides a water drive oil reservoir injection and production parameter optimization system, characterized in that the system is used in any one of the above-mentioned water drive oil reservoir injection and production parameter optimization methods, and the system comprises:

[0104] The first module 100 is used to determine the initial conditions of the oil and water well injection and production parameter optimization problem, wherein the initial conditions include design variables, design space, objective function and constraint conditions;

[0105] The second module 200 is used to establish fractured oil reservoir water flooding simulation models of different precisions using a commercial simulator, wherein the fractured oil reservoir water flooding simulation models of different precisions include at least one high-precision simulation model and at least one low-precision simulation model;

[0106] The third module 300 is used to evenly divide the design space according to the dimension of the design variables to obtain a plurality of small spaces, obtain the coordinates of the center point of each small space as a basic sample point, and randomly sample in the design space to obtain non-basic sample points;

[0107] The fourth module 400 is used to input the basic sample points into at least one high-precision simulation model for calculation, obtain the net present value corresponding to the basic sample points, and construct a high-fidelity training sample data set using the basic sample points and the corresponding net present value; input the non-basic sample points into at least one low-precision simulation model for calculation, obtain the net present value corresponding to the non-basic sample points, and construct a low-fidelity training sample data set using the non-basic sample points and the corresponding net present value;

[0108] The fifth module 500 is used to construct a variable credibility adaptive proxy model for the oil and water well injection and production parameter optimization problem based on the high-fidelity training sample data set and the low-fidelity training sample data set;

[0109] The sixth module 600 is used to optimize the design variables by using the differential evolution hybrid particle swarm algorithm combined with the variable credibility adaptive agent model to obtain the maximum net present value and the corresponding injection and production plan

[0110] It should be noted that the modules in the system of Example 2 correspond to the steps in the method of Example 1. The steps in the method of Example 1 have been described in detail in Example 1. In this Example 2, the contents of the modules in the system will not be described in detail.

[0111] Example 3

[0112] See also Figure 7 As shown, this embodiment further provides a computer device, including a system memory 1005 and a processor 1001, wherein the system memory 1005 stores a computer program, and the processor 1001 implements the steps of any of the above methods when executing the computer program.

[0113] It should be noted that the processor 1001 is used to execute the steps in the above method embodiments according to the instructions in the program code. Alternatively, the processor 1001 implements the functions of each module / unit in the above system / device embodiments when executing the computer program.

[0114] Specifically, in this embodiment, the computer program may be divided into one or more modules / units, one or more modules / units are stored in the system memory 1005, and are executed by the processor 1001 to complete the present application. One or more modules / units may be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.

[0115] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor 1001 and a system memory 1005. Those skilled in the art will appreciate that this does not constitute a limitation on the terminal device, and may include more or less components than shown in the figure, or combine certain components, or different components. For example, the terminal device may also include an input / output device 1003, a network access device 1002, a bus 1006, etc.

[0116] The processor 1001 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor 1001 may be a microprocessor or any conventional processor, etc.

[0117] The system memory 1005 may be an internal storage unit of the terminal device, such as a hard disk or memory of the terminal device. The system memory 1005 may also be a storage device 1004 of the terminal device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the terminal device. Further, the system memory 1005 may also include both the internal storage unit of the terminal device and the storage device 1004. The system memory 1005 is used to store computer programs and other programs and data required by the terminal device. The system memory 1005 may also be used to temporarily store data that has been output or is to be output.

[0118] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0119] Example 4

[0120] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above methods are implemented.

[0121] Among them, the computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, system or device, or any combination of the above. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk. Random Access Memory (RAM), Read-Only Memory (ROM), Erasable Programmable Read Only Memory (EPROM), registers, hard disks, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above, or any other form of computer-readable storage medium known in the art.

[0122] An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application-specific integrated circuit (ASIC). In an embodiment of the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system, or device.

[0123] Example 5

[0124] This embodiment also provides a computer program product including instructions. When the instructions are executed by a computer device cluster, the computer device cluster executes the method described in any one of the above-mentioned embodiments 1.

[0125] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for optimizing injection and production parameters of a water drive oil reservoir, characterized in that: Methods include: Determine the initial conditions of the oil and water well injection and production parameter optimization problem, where the initial conditions include design variables, design space, objective function and constraints, specifically: The design variables are the injection and production parameters of oil and water wells, and the objective function is constructed with the maximum net present value of reservoir production as the goal, and the injection and production constraints are established. The specific objective function is: Where, NPV(u) is the net present value, which is a function of the design variable oil and water well injection and production parameter u; N t is the total time step of reservoir simulation; Δt n is the length of the nth time step; b is the discount rate; t n is the cumulative time at the end of the nth time step; N o and N w Respectively represent the number of oil wells and water wells in the model; r o is the crude oil sales price; c pw is the production water treatment cost; c iw injection costs for water injection wells; and are the average oil production and water production of oil producing well i and the water injection of water injection well j in the nth time step respectively; the specific injection and production constraints are: min ≤u i ≤u max ,i=1,2,…,N o ,N o +1,N o +N w , where u min 、u max They represent the minimum and maximum oil and water well injection and production parameters, respectively; Using a commercial simulator to establish fractured oil reservoir water flooding simulation models of different precisions, wherein the fractured oil reservoir water flooding simulation models of different precisions include at least one high-precision simulation model and at least one low-precision simulation model; The design space is evenly divided according to the dimension of the design variables to obtain multiple small spaces, and the coordinates of the center points of each small space are obtained as basic sample points, and non-basic sample points are obtained by random sampling in the design space; Inputting basic sample points into at least one high-precision simulation model for calculation to obtain the net present value corresponding to the basic sample points, and using the basic sample points and the corresponding net present value to construct a high-fidelity training sample data set; inputting non-basic sample points into at least one low-precision simulation model for calculation to obtain the net present value corresponding to the non-basic sample points, and using the non-basic sample points and the corresponding net present value to construct a low-fidelity training sample data set; Based on the high-fidelity training sample data set and the low-fidelity training sample data set, a variable credibility adaptive proxy model for the oil and water well injection and production parameter optimization problem is constructed. Specifically, the high-fidelity training sample data set and the low-fidelity training sample data set are fused to construct the variable credibility proxy model; Using the kernel function idea of ​​the support vector regression model, based on the variable credibility proxy model kernel function, the high-fidelity training sample data set and the low-fidelity training sample data set are mapped to a high-dimensional space, and the kernel function matrix of the fusion of the high-fidelity training sample data set and the low-fidelity training sample data set is obtained. Based on the kernel function matrix, a variable credibility adaptive proxy model for the optimization problem of injection and production parameters of oil and water wells is constructed. The kernel function matrix is ​​specifically: Among them, K MF is the kernel function matrix of the variable credibility adaptive agent model; X H is a high-fidelity training sample dataset; X L is a low-fidelity training sample dataset; K L-L , K L-H , K H-L , K H-H are the elements in the kernel function matrix, where K L-L (X L , X L ) represents the kernel function matrix between low-fidelity training sample data, K L-H (X L , X H ) represents the kernel function matrix between the low-fidelity training sample data and the high-fidelity training sample data, K H-L (X H , X L ) represents the kernel function matrix between the high-fidelity training sample data and the low-fidelity training sample data, K H-H (X H , X H ) represents the kernel function matrix between high-fidelity training sample data; N L is the number of low-fidelity training sample data, NL is a positive integer; N H is the number of high-fidelity training sample data, N H is a positive integer; The particle swarm algorithm is used to adaptively search for the optimal hyperparameter combination in the variable credibility adaptive proxy model. The root mean square error between the net present value output by the variable credibility adaptive proxy model and the net present value in the high-fidelity training sample data set is used as the evaluation index. When the evaluation index is minimized, the optimal variable credibility adaptive proxy model is obtained. The optimal variable credibility adaptive proxy model is specifically: Among them, y n For x n The objective function response value of; b is the bias term; α i are all Lagrange multipliers; N L Represented as the number of low-fidelity training sample data, N L is a positive integer; x i 、x n Represented as elements in the sample data vector; The differential evolution hybrid particle swarm algorithm is combined with the variable credibility adaptive agent model to optimize the design variables and obtain the maximum net present value and the corresponding injection-production scheme.

2. A method for optimizing injection and production parameters of a water drive reservoir according to claim 1, characterized in that: A commercial simulator was used to establish water flooding simulation models of fractured reservoirs with different precisions, specifically: According to the actual physical parameters of the oil and water wells and the actual injection and production parameters, a three-dimensional reservoir model is constructed, and according to the model parameters of the three-dimensional reservoir model, a fractured reservoir water drive simulation model with different accuracies is established using one of the eclipse commercial simulators or the CMG commercial simulator, wherein the model parameters include the number of grids, the grid size and the reservoir physical properties.

3. A method for optimizing injection and production parameters of a water drive reservoir according to claim 1, characterized in that: The design space is evenly divided according to the dimension of the design variables to obtain multiple small spaces, specifically: Determine the dimension of the hypercube and the corresponding number of partitions on each dimension of the hypercube according to the dimension of the design variables; The hypercube space is evenly divided based on the number of partitions in each dimension of the hypercube to obtain multiple small spaces.

4. A method for optimizing injection and production parameters of a water drive reservoir according to claim 1, characterized in that: The differential evolution hybrid particle swarm algorithm is used in combination with the variable credibility adaptive agent model to optimize the design variables and obtain the maximum net present value and the corresponding injection-production scheme, which are as follows: Based on the variable credibility adaptive agent model, the differential evolution hybrid particle swarm algorithm is used to solve the optimization problem to optimize the design variables and obtain the maximum net present value and the corresponding injection and production scheme. The specific optimization problem is: in, is the variable credibility adaptive proxy model of net present value; u is the injection and production parameter of oil and water wells; u min 、u max They represent the minimum injection and production parameters of oil and water wells and the maximum injection and production parameters of oil and water wells respectively; Find u means finding the optimal injection and production parameters of oil and water wells.

5. A water drive reservoir injection and production parameter optimization system, characterized in that: The system is used in a method for optimizing injection and production parameters of a water drive oil reservoir as claimed in any one of claims 1 to 4, and the system comprises: The first module is used to determine the initial conditions of the oil and water well injection and production parameter optimization problem, wherein the initial conditions include design variables, design space, objective function and constraint conditions; The second module is used to establish fractured oil reservoir water flooding simulation models with different precisions using a commercial simulator, wherein the fractured oil reservoir water flooding simulation models with different precisions include at least one high-precision simulation model and at least one low-precision simulation model; The third module is used to evenly divide the design space according to the dimension of the design variables to obtain multiple small spaces, obtain the coordinates of the center point of each small space as the basic sample point, and randomly sample in the design space to obtain non-basic sample points; The fourth module is used to input the basic sample points into at least one high-precision simulation model for calculation, obtain the net present value corresponding to the basic sample points, and use the basic sample points and the corresponding net present value to construct a high-fidelity training sample data set; input the non-basic sample points into at least one low-precision simulation model for calculation, obtain the net present value corresponding to the non-basic sample points, and use the non-basic sample points and the corresponding net present value to construct a low-fidelity training sample data set; The fifth module is used to construct a variable credibility adaptive proxy model for the oil and water well injection and production parameter optimization problem based on a high-fidelity training sample data set and a low-fidelity training sample data set; The sixth module is used to optimize the design variables using the differential evolution hybrid particle swarm algorithm combined with the variable credibility adaptive agent model to obtain the maximum net present value and the corresponding injection and production plan.

6. A computer device comprising a system memory and a processor, wherein the system memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

8. A computer program product comprising instructions, characterized in that When the instructions are executed by a computer device cluster, the computer device cluster is caused to perform the method according to any one of claims 1 to 4.

Images