A method, system and device for automatically history matching a fractured reservoir model

Abstract

The application discloses a kind of fractured reservoir automatic history fitting model prediction method, system and equipment, establish fractured reservoir numerical simulation model, fracture network is divided into large-scale fracture and small-scale fracture;Parameterization is carried out to large-scale fracture and small-scale fracture respectively;Fracture network structure is established in combination with large-scale fracture and small-scale fracture parameters;Improved proxy assisted collaborative group optimization algorithm is established through fracture network structure, and radial basis neural network is changed into interpolation form radial basis function;Fractured reservoir is inverted and predicted using improved proxy assisted collaborative group optimization algorithm and data-driven evolutionary algorithm, and the optimal solution is the actual fracture distribution parameter.Can according to the production history of well, the distribution form of fracture that general earthquake is difficult to identify is predicted, and then the remaining oil distribution state of approximate real situation is obtained.

Description

Technical Field

[0001] This invention belongs to the field of oil and gas reservoir development and relates to a method, system and equipment for predicting fractured reservoirs using an automatic historical fitting model. Background Technology

[0002] Of the world's proven oil and gas reserves, carbonate reservoirs account for more than half, making them a crucial type of reservoir for oil and gas exploration and development. Carbonate oil and gas reservoirs are widely distributed in my country, possessing enormous exploration and development potential. However, carbonate reservoirs are highly heterogeneous and geologically complex. Furthermore, fractures often exist at multiple scales, significantly impacting reservoir development outcomes. Current geophysical exploration techniques cannot precisely describe the development of fracture-vuggy systems in carbonate reservoirs, primarily due to the extremely complex and uncertain distribution of fractures. Therefore, research into seismic fracture prediction techniques is needed to improve understanding of reservoir flow patterns and ultimately obtain accurate information on remaining oil distribution.

[0003] In reservoir fracture prediction methods, the history fitting method can be used to invert the geological model. The history fitting process involves continuously adjusting model parameters based on historical production data to determine the actual model parameters. However, manual history fitting is heavily influenced by human factors, time-consuming, and has many limitations. Automatic history fitting, which combines computer technology to solve the inverse problem, is faster and easier to operate than manual history fitting. Through automatic history fitting, the uncertainty of fracture network distribution can be reduced.

[0004] Historical simulation of fractured reservoirs first requires establishing a numerical simulation model to model fluid flow within the fracture network. Traditional methods for simulating fracture networks include dual-medium models, equivalent continuous-medium models, discrete fracture network models (DFM), and embedded discrete fracture network models (EDFM). In dual-medium models, fractures and matrix are represented by defining two types of permeability and porosity. In equivalent continuous-medium models, the matrix and fractures are characterized by separate continuous media, thus defining equivalent permeability and porosity. Dual-medium and equivalent continuous-medium models do not require separate meshes, making the numerical simulation process simple and stable; however, both methods struggle to explicitly describe the impact of fractures on seepage. Discrete fracture network models (DFM) typically use unstructured meshes to explicitly describe fracture locations, leading to computational costs. Embedded discrete fracture network models (EDFM) use structured meshes, reducing computational costs; however, their accuracy in simulating fractures is relatively lower than that of discrete fracture network models.

[0005] Set-based methods are widely used in reservoir history fitting. For example, the Markov Chain Monte Carlo (MCMC) method estimates the distribution characteristics of model parameters by sampling the posterior probability function of the historical fitting objective function. Other algorithms, such as Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), solve the objective function to find the most probable fracture distribution. Unfortunately, fractured reservoirs often have a large number of parameters, thus the resulting high-dimensional problems have not yet been effectively solved.

[0006] In summary, there is a lack of optimization algorithms that can adapt to high-dimensional problems in the process of solving the history fitting of fractured reservoirs. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide an automatic history fitting model prediction method, system and equipment for fractured reservoirs. It can predict the distribution pattern of fractures that are difficult to identify by general earthquakes based on the production history of wells, and thus obtain a near-realistic distribution of remaining oil.

[0008] To achieve the above objectives, the present invention employs the following technical solution:

[0009] An automatic history fitting model prediction method for fractured reservoirs includes the following steps:

[0010] Step 1: Establish a numerical simulation model of fractured reservoirs, and divide the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0011] Step 2: Parameterize large-scale cracks and small-scale cracks respectively;

[0012] Step 3: Establish a crack network structure by combining parameters of large-scale and small-scale cracks;

[0013] Step 4: An improved agent-assisted cooperative group optimization algorithm is established through the crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function.

[0014] Step 5: The improved agent-assisted cooperative group optimization algorithm and the data-driven evolutionary algorithm are used to perform inversion prediction on fractured reservoirs. The optimal solution obtained is the actual fracture distribution parameters.

[0015] Preferably, the fracture characterization parameters in the numerical simulation model of fractured reservoirs are fracture length, orientation, and midpoint coordinates.

[0016] Preferably, the process of parameterizing large-scale cracks is as follows: large-scale crack parameters are randomly generated by taking into account the range of crack length, orientation, and midpoint coordinates; the crack aperture is calculated based on the cube law; and then the large-scale crack parameters are combined.

[0017] Preferably, the parameterization process for small-scale cracks is as follows: setting the maximum and minimum crack lengths and fractal dimension, calculating the number of all cracks, generating a small-scale crack network based on the relationship between crack length and number, dividing the small-scale cracks into a dataset, and then combining the small-scale crack parameters.

[0018] Preferably, in step three, a crack network structure is formed by combining the parameters of large-scale cracks and small-scale cracks, and numerical simulation is performed using an embedded discrete crack network model based on the parameters of the crack network structure.

[0019] Preferably, the radial basis function in interpolation form is expressed as:

[0020]

[0021] In the formula, f(x) is the radial basis function in interpolation form; N is the number of hidden layer nodes; ω k φ is the weighting coefficient; φ is the basis function; x k Center point.

[0022] Preferably, the specific process of step five is as follows: using an improved surrogate-assisted cooperative group optimization algorithm to perform history fitting on fractured reservoirs, setting an objective function for history fitting, and using the improved surrogate-assisted cooperative group optimization algorithm to solve this objective function, the optimal solution obtained is the actual fracture distribution parameters.

[0023] An automatic history fitting model prediction system for fractured reservoirs includes:

[0024] The simulation model building module is used to build a numerical simulation model of fractured reservoirs, and divides the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0025] The parameterization module is used to parameterize large-scale cracks and small-scale cracks respectively.

[0026] The crack network structure establishment module is used to establish a crack network structure by combining parameters of large-scale cracks and small-scale cracks.

[0027] An optimization algorithm establishment module is used to establish an improved agent-assisted cooperative group optimization algorithm through a crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated form of radial basis function.

[0028] The module for obtaining actual fracture distribution parameters is used to perform inversion prediction on fractured reservoirs using an improved surrogate-assisted cooperative group optimization algorithm and a data-driven evolutionary algorithm. The optimal solution obtained is the actual fracture distribution parameters.

[0029] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the automatic history fitting model prediction method for fractured reservoirs as described in any of the preceding claims.

[0030] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the automatic history fitting model prediction method for fractured reservoirs as described in any of the preceding claims.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] This invention divides the fracture network into large-scale fractures and small-scale fractures, which greatly reduces the scale of parameters used to characterize fracture distribution and lowers the fitting difficulty. It adopts an improved surrogate-assisted cooperative group optimization algorithm and replaces the radial basis neural network with an interpolated radial basis function, which can accelerate the iteration speed of the evolutionary algorithm and improve the evolution speed of the evolutionary algorithm in the optimization process. This greatly reduces the computational time cost required for automatic history fitting. The overall method can improve the accuracy of predicting the distribution pattern of fractures and thus obtain a near-realistic distribution of remaining oil. Attached Figure Description

[0033] Figure 1 This is a reference model diagram of the original crack distribution in this invention;

[0034] Figure 2 This is the initial crack distribution feature map of the present invention;

[0035] Figure 3 This is a schematic diagram of the optimized crack network model of the present invention;

[0036] Figure 4 This is a schematic diagram of the residual oil saturation in the optimized model of the present invention;

[0037] Figure 5 This is a schematic diagram of the residual oil saturation of the reference model of the present invention. Detailed Implementation

[0038] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0039] It should be noted that the terms “front,” “back,” “left,” “right,” “up,” and “down” used in the following description refer to the directions shown in the attached diagram, while the terms “inside” and “outside” refer to the directions toward or away from the geometric center of a specific component, respectively.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0041] The automatic history fitting model prediction method for fractured reservoirs according to the present invention includes the following steps:

[0042] Step 1: Establish a numerical simulation model of fractured reservoirs, and divide the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0043] In the numerical simulation model of fractured reservoirs, the characteristic parameters of fractures are fracture length, orientation, and midpoint coordinates.

[0044] Step 2: Parameterize the large-scale cracks and the small-scale cracks respectively.

[0045] The process of parameterizing large-scale cracks is as follows: large-scale crack parameters are randomly generated by taking into account the range of crack length, orientation, and midpoint coordinates; the crack aperture is calculated based on the cubic law; and then the large-scale crack parameters are combined.

[0046] The parameterization process for small-scale cracks is as follows: set the maximum and minimum crack lengths and fractal dimension, calculate the total number of cracks, generate a small-scale crack network based on the relationship between crack length and number, divide the small-scale cracks into datasets, and then combine the small-scale crack parameters.

[0047] Step 3: Establish a crack network structure by combining parameters of large-scale and small-scale cracks.

[0048] By combining the parameters of large-scale and small-scale cracks, a crack network structure is formed. Based on the parameters of the crack network structure, an embedded discrete crack network model is used for numerical simulation.

[0049] Step 4: Establish an improved agent-assisted cooperative group optimization algorithm through the crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function.

[0050] The radial basis function in interpolation form is expressed as:

[0051]

[0052] In the formula, f(x) is the radial basis function in interpolation form; N is the number of hidden layer nodes; ω k φ is the weighting coefficient; φ is the basis function; x k Center point.

[0053] Step 5: The improved agent-assisted cooperative group optimization algorithm and the data-driven evolutionary algorithm are used to perform inversion prediction on fractured reservoirs. The optimal solution obtained is the actual fracture distribution parameters.

[0054] An improved surrogate-assisted cooperative group optimization algorithm was used to perform history fitting on fractured reservoirs. An objective function for history fitting was set, and the improved surrogate-assisted cooperative group optimization algorithm was used to solve this objective function. The optimal solution obtained is the actual fracture distribution parameters.

[0055] The specific process of the above scheme is as follows:

[0056] Step 1: Establish a numerical simulation model for fractured reservoirs.

[0057] A numerical simulation model of fractured reservoirs was established based on an embedded discrete fracture network (EDFM) implemented using the MRST toolkit in MATLAB software. The fracture characterization parameters are fracture length, azimuth, and midpoint coordinates. The fracture network can be divided into two parts: large-scale fractures and small-scale fractures. Depending on the parameter settings, small-scale fractures can also be transformed into medium-scale and large-scale fractures.

[0058] Step 2: Parameterize large-scale cracks

[0059] For the large-scale crack parameterization method, large-scale crack parameters can be randomly generated by considering the range of crack length, orientation, and midpoint coordinates. The crack aperture can then be calculated based on the cubic law, as shown in the following formula:

[0060] e = βl

[0061] In the formula, e is the crack opening, β is a constant, and l is the crack length.

[0062] Combining the parameters of large-scale cracks, the parameters of large-scale cracks can be expressed as follows:

[0063]

[0064] In the formula, m l For a large-scale crack parameter set, l i Let x be the length of the i-th crack. i y i Let θ be the coordinate of the midpoint of the i-th crack. i Let N be the orientation of the i-th crack.ma This refers to the number of large cracks.

[0065] Step 3: Parameterize small-scale cracks based on fractal theory.

[0066] First, the maximum and minimum crack lengths and fractal dimension are set, and the total number of cracks is calculated. The parameterization method for small-scale cracks is based on fractal theory, generating a small-scale crack network through the relationship between crack length and number. The relationship between crack length and number in fractal theory is as follows:

[0067]

[0068] In the formula, N to l represents the total number of cracks. max The length of the longest crack, l min D is the length of the shortest crack. l It is the fractal dimension.

[0069] Then, by setting a random number R, the length of each crack can be obtained. The method for generating the length of each specific crack is as follows:

[0070]

[0071] In the formula, R is a random number between 0 and 1.

[0072] Furthermore, the small-scale cracks are divided into datasets. Different datasets have different crack generation parameters and properties, thus characterizing the multi-scale crack network. The crack intensity D in each dataset can be calculated based on the number of cracks in each dataset. s The calculation formula is as follows:

[0073]

[0074] in N represents the number of cracks in crack set i. mi N represents the number of all small cracks. se The number of data sets to be divided.

[0075] Finally, the angle of each small crack is generated by a Gaussian distribution function with a mean of . The variance is σ i The parameters of small-scale cracks can be combined according to the actual custom settings.

[0076] Step 4: Establish a crack network structure by combining large-scale cracks and small-scale cracks.

[0077] By combining the parameters of large-scale and small-scale cracks, a crack network structure is constructed, and the parameters of the crack network structure are expressed as follows:

[0078] m = {m1, m2}

[0079] Numerical simulations were performed using the Embedded Discrete Crack Network Model (EDFM) based on the parameters of the crack network.

[0080] Step 5: Develop an improved agent-assisted collaborative group optimization algorithm.

[0081] Proxy-Assisted Cooperative Population Optimization (SACOSO) is a data-driven evolutionary algorithm. Data-driven algorithms can accelerate convergence and reduce computational cost. SACOSO is suitable for optimization problems of 50-100 dimensions. The traditional SACOSO algorithm solves the objective function through the synergistic effect of two particle swarm optimization algorithms: one based on an adaptive evaluation strategy and the other based on radial basis functions. The synergy of these two algorithms balances the search performance of the evolutionary algorithm.

[0082] The specific process is as follows:

[0083] 1) Combine particle swarm optimization algorithm with fitness value evaluation, where fitness value evaluation needs to be combined with the numerical simulation in step 4, and iteratively updated to find the optimal solution;

[0084] 2) Use the data from the first step of the algorithm search process to form a sample set, and train a surrogate model for the radial basis function;

[0085] 3) Combine particle swarm optimization algorithm with radial basis function surrogate model evaluation, update and iterate to find the optimal solution, and this optimal solution will guide the search process in the first step;

[0086] 4) Repeat steps 1-3 above until the algorithm converges.

[0087] The improved agent-assisted cooperative group optimization algorithm (moSACOSO) is based on the SACOSO algorithm, replacing the radial basis function neural network (RBF-NN) with an interpolated form of radial basis function (RBF-IN). The interpolated form of RBF-IN allows for faster model building and yields more accurate results. The interpolated form of RBF-IN can be expressed as:

[0088]

[0089] In the formula, f(x) is the radial basis function in interpolation form; N is the number of hidden layer nodes; ω k φ is the weighting coefficient; φ is the basis function; x k Let be the center point. The basis function φ can be chosen as:

[0090] φ(r)=r 3

[0091] Weight vector ω=(ω1,...,ω N ) T It can be represented as:

[0092] ω=(Φ T Φ) -1 Φ T y

[0093]

[0094] Where y = (y1,...,y) N ) T .

[0095] Step 6: Use a data-driven evolutionary algorithm to perform inversion prediction on fractured reservoirs.

[0096] Automatic history fitting was performed on fractured reservoir models using an improved agent-assisted cooperative group optimization algorithm (moSACOSO).

[0097] First, define the historical fitting objective function. Based on the relationship between the model parameters and historical production data:

[0098] d=g(m)+ε

[0099] In the formula, d represents the observed data, m represents the model parameters, g(m) represents the numerical simulation data, and ε represents the error.

[0100] According to Bayesian theory, the posterior probability can be represented by the prior probability and the likelihood function. In the problem of automatic history fitting in reservoirs, the posterior probability density function p(d|m) can be expressed as:

[0101]

[0102] In the formula, C d Let C be the covariance of the observed data error, and C be a constant.

[0103] Therefore, the objective function for history fitting is set as:

[0104]

[0105] Then, the improved agent-assisted cooperative group optimization algorithm (moSACOSO) is used to solve this objective function, and the optimal solution obtained is the actual crack distribution parameters.

[0106] The original crack distribution is as follows Figure 1 As shown, it contains 6 large cracks and many small cracks.

[0107] The results of the improved method are as follows Figure 2 and Figure 3 As shown, Figure 2Based on the initial crack distribution characteristics, the characteristic parameters of the cracks were optimized using an evolutionary algorithm, and the final optimal crack inversion result is as follows. Figure 3 As shown, the distribution characteristics and locations of the large cracks can be determined by... Figure 1 The actual crack locations are relatively consistent, demonstrating high accuracy.

[0108] Based on the crack model, the remaining oil saturation was also predicted, such as... Figure 4 and Figure 5 As shown, Figure 4 For the predicted results, Figure 5 The figures show the saturation distribution results from the actual model. A comparison of the two figures shows that the predicted remaining oil distribution is basically consistent with the actual reservoir condition.

[0109] Scope of application and application prospects.

[0110] Scope of application: This method is applicable to fractured reservoirs and can approximately fit the distribution of underground fractures. It is particularly suitable for well group models with a small scope.

[0111] Application Prospects: Fractures in carbonate reservoirs are the primary oil transport channels; therefore, subsurface fracture prediction is a crucial area of ​​research for current development. This method can approximate the distribution of subsurface fractures, improving engineers' understanding of reservoir permeability characteristics and remaining oil distribution, which is significant for subsequent well placement and development policy adjustments. Furthermore, this method is adaptable to various fractured reservoirs, offering broad application potential.

[0112] The following are embodiments of the apparatus of the present invention, which can be used to execute embodiments of the method of the present invention. For details not omitted in the apparatus embodiments, please refer to the embodiments of the method of the present invention.

[0113] In another embodiment of the present invention, an automatic history fitting model prediction system for fractured reservoirs is provided. This automatic history fitting model prediction system for fractured reservoirs can be used to implement the above-mentioned automatic history fitting model prediction method for fractured reservoirs. Specifically, the automatic history fitting model prediction system for fractured reservoirs includes a simulation model establishment module, a parameterization module, a fracture network structure establishment module, an optimization algorithm establishment module, and an actual fracture distribution parameter acquisition module.

[0114] The simulation model building module is used to build a numerical simulation model of fractured reservoirs, and divides the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0115] The parameterization module is used to parameterize large-scale cracks and small-scale cracks respectively.

[0116] The crack network structure creation module is used to create a crack network structure by combining parameters of large-scale cracks and small-scale cracks.

[0117] An optimization algorithm establishment module is used to establish an improved agent-assisted cooperative group optimization algorithm through a crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function.

[0118] The module for obtaining actual fracture distribution parameters is used to perform inversion prediction on fractured reservoirs using an improved surrogate-assisted cooperative group optimization algorithm and a data-driven evolutionary algorithm. The optimal solution obtained is the actual fracture distribution parameters.

[0119] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment can be used for automatic historical fitting model prediction of fractured reservoirs. The operation of the method includes:

[0120] Step 1: Establish a numerical simulation model of fractured reservoirs, and divide the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0121] Step 2: Parameterize the large-scale cracks and the small-scale cracks respectively.

[0122] Step 3: Establish a crack network structure by combining parameters of large-scale and small-scale cracks.

[0123] Step 4: Establish an improved agent-assisted cooperative group optimization algorithm through the crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function.

[0124] Step 5: The improved agent-assisted cooperative group optimization algorithm and the data-driven evolutionary algorithm are used to perform inversion prediction on fractured reservoirs. The optimal solution obtained is the actual fracture distribution parameters.

[0125] In another embodiment, the present invention also provides a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It is understood that the computer-readable storage medium here may include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which may be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here may be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0126] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the automatic history fitting model prediction method for fractured reservoirs in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor in the following steps:

[0127] Step 1: Establish a numerical simulation model of fractured reservoirs, and divide the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures.

[0128] Step 2: Parameterize the large-scale cracks and the small-scale cracks respectively.

[0129] Step 3: Establish a crack network structure by combining parameters of large-scale and small-scale cracks.

[0130] Step 4: Establish an improved agent-assisted cooperative group optimization algorithm through the crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function.

[0131] Step 5: The improved agent-assisted cooperative group optimization algorithm and the data-driven evolutionary algorithm are used to perform inversion prediction on fractured reservoirs. The optimal solution obtained is the actual fracture distribution parameters.

[0132] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0133] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0134] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0135] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0136] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0137] It should be understood that the above description is for illustrative purposes and not for limitation. Many embodiments and applications beyond the provided examples will be apparent to those skilled in the art upon reading the above description. Therefore, the scope of this teaching should not be determined by reference to the above description, but rather by reference to the foregoing claims and the full scope of their equivalents. For purposes of completeness, all articles and references, including patent applications and publications, are incorporated herein by reference. The omission of any aspect of the subject matter disclosed herein in the foregoing claims is not intended as a waiver of that subject matter, nor should it be construed as an indication that the applicant has not considered that subject matter as part of the disclosed inventive subject matter.

Claims

1. A method for predicting fractured reservoirs using an automatic history fitting model, characterized in that, Includes the following steps: Step 1: Establish a numerical simulation model of fractured reservoirs, and divide the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures. In the numerical simulation model of fractured reservoirs, the characteristic parameters of fractures are fracture length, orientation, and midpoint coordinates. Step 2: Parameterize large-scale cracks and small-scale cracks respectively; The process of parameterizing large-scale cracks is as follows: large-scale crack parameters are randomly generated by taking into account the range of crack length, orientation and midpoint coordinates, the crack aperture is calculated based on the cube law, and then the large-scale crack parameters are combined. The parameterization process for small-scale cracks is as follows: set the maximum and minimum crack lengths and fractal dimension, calculate the number of all cracks, generate a small-scale crack network based on the relationship between crack length and number, divide the small-scale cracks into datasets, and then combine the small-scale crack parameters. Step 3: Establish a crack network structure by combining the parameters of large-scale and small-scale cracks, and perform numerical simulation using an embedded discrete crack network model based on the parameters of the crack network structure. Step 4: An improved agent-assisted cooperative group optimization algorithm is established through the crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated radial basis function. The radial basis function in interpolation form is expressed as: ； In the formula, These are radial basis functions in interpolation form; N This represents the number of hidden layer nodes. These are the weighting coefficients; are basis functions; Center point; Step 5: The improved agent-assisted cooperative group optimization algorithm and the data-driven evolutionary algorithm are used to perform inversion prediction on fractured reservoirs. The optimal solution obtained is the actual fracture distribution parameters.

2. The automatic history fitting model prediction method for fractured reservoirs according to claim 1, characterized in that, The specific process of step five is as follows: use the improved surrogate-assisted cooperative group optimization algorithm to perform history fitting on fractured reservoirs, set the objective function for history fitting, and use the improved surrogate-assisted cooperative group optimization algorithm to solve this objective function. The optimal solution obtained is the actual fracture distribution parameters.

3. An automatic history fitting model prediction system for fractured reservoirs based on the method of claim 1, characterized in that, include: The simulation model building module is used to build a numerical simulation model of fractured reservoirs, and divides the fracture network in the numerical simulation model of fractured reservoirs into large-scale fractures and small-scale fractures. The parameterization module is used to parameterize large-scale cracks and small-scale cracks respectively. The crack network structure establishment module is used to establish a crack network structure by combining parameters of large-scale cracks and small-scale cracks. An optimization algorithm establishment module is used to establish an improved agent-assisted cooperative group optimization algorithm through a crack network structure. In the improved agent-assisted cooperative group optimization algorithm, the radial basis neural network is replaced with an interpolated form of radial basis function. The module for obtaining actual fracture distribution parameters is used to perform inversion prediction on fractured reservoirs using an improved surrogate-assisted cooperative group optimization algorithm and a data-driven evolutionary algorithm. The optimal solution obtained is the actual fracture distribution parameters.

4. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the automatic history fitting model prediction method for fractured reservoirs as described in any one of claims 1 to 2.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the automatic history fitting model prediction method for fractured reservoirs as described in any one of claims 1 to 2.

Images